4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
25 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
26 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
27 and require returning octet ptr parameters from providers that
28 would like to support them which complicates provider implementations.
32 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
33 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
34 RAND_set_rand_method(). Provider based random number generators should
35 be used instead via EVP_RAND(3).
38 * The SRP APIs have been deprecated. The old APIs do not work via providers,
39 and there is no EVP interface to them. Unfortunately there is no replacement
40 for these APIs at this time.
44 * Add a compile time option to prevent the caching of provider fetched
45 algorithms. This is enabled by including the no-cached-fetch option
46 at configuration time.
50 * The openssl speed command does not use low-level API calls anymore. This
51 implies some of the performance numbers might not be fully comparable
52 with the previous releases due to higher overhead. This applies
53 particularly to measuring performance on smaller data chunks.
57 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
58 Typically if OpenSSL has no EC or DH algorithms then it cannot support
59 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
60 through providers. Therefore third party providers may supply group
61 implementations even where there are no built-in ones. Attempting to create
62 TLS connections in such a build without also disabling TLSv1.3 at run time or
63 using third party provider groups may result in handshake failures. TLSv1.3
64 can be disabled at compile time using the "no-tls1_3" Configure option.
68 * The undocumented function X509_certificate_type() has been deprecated;
69 applications can use X509_get0_pubkey() and X509_get0_signature() to
70 get the same information.
74 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
75 functions. They are identical to BN_rand() and BN_rand_range()
80 * Deprecated the obsolete X9.31 RSA key generation related functions
81 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
82 BN_X931_generate_prime_ex().
86 * The default key generation method for the regular 2-prime RSA keys was
87 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
88 Conditions Based on Auxiliary Probable Primes). This method is slower
89 than the original method.
93 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
94 They are replaced with the BN_check_prime() function that avoids possible
95 misuse and always uses at least 64 rounds of the Miller-Rabin
96 primality test. At least 64 rounds of the Miller-Rabin test are now also
97 used for all prime generation, including RSA key generation.
98 This increases key generation time, especially for larger keys.
102 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
103 as they are not useful with non-deprecated functions.
107 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
108 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
109 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
110 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
111 were used to collect all necessary data to form a HTTP request, and to
112 perform the HTTP transfer with that request. With OpenSSL 3.0, the
113 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
114 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
115 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
116 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
117 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
118 OSSL_HTTP_REQ_CTX_set_max_response_length().
120 *Rich Salz and Richard Levitte*
122 * Validation of SM2 keys has been separated from the validation of regular EC
123 keys, allowing to improve the SM2 validation process to reject loaded private
124 keys that are not conforming to the SM2 ISO standard.
125 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
130 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
131 switches: a validation failure triggers an early exit, returning a failure
132 exit status to the parent process.
136 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
137 to ignore unknown ciphers.
141 * The `-cipher-commands` and `-digest-commands` options
142 of the command line utility `list` have been deprecated.
143 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
147 * All of the low level EC_KEY functions have been deprecated including:
149 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
150 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
151 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
152 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
153 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
154 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
155 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
156 EC_KEY_METHOD_get_verify,
157 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
158 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
159 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
160 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
161 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
162 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
163 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
164 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
165 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
166 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
167 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
168 Applications that need to implement an EC_KEY_METHOD need to consider
169 implementation of the functionality in a special provider.
170 For replacement of the functions manipulating the EC_KEY objects
171 see the EVP_PKEY-EC(7) manual page.
173 Additionally functions that read and write EC_KEY objects such as
174 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
175 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
176 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
177 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
178 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
179 have also been deprecated. Applications should instead use the
180 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
182 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
183 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
184 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
185 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
186 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
188 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
190 * Deprecated all the libcrypto and libssl error string loading
191 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
192 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
193 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
194 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
195 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
196 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
197 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
198 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
199 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
200 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
202 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
203 now loads error strings automatically.
207 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
208 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
209 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
210 are to be used by servers requiring ephemeral DH keys. Instead applications
211 should consider using the built-in DH parameters that are available by
212 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
213 necessary then applications can use the alternative functions
214 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
215 replacement for the "callback" functions. The callback was originally useful
216 in order to have different parameters for export and non-export ciphersuites.
217 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
218 functions should be replaced by one of the other methods described above.
222 * The `-crypt` option to the `passwd` command line tool has been removed.
226 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
231 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
233 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
234 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
235 The inverse ciphers use AES decryption for wrapping, and
236 AES encryption for unwrapping.
240 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
241 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
242 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
243 OpenSSL 3.0 these are replaced by the more generic functions
244 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
245 The old versions have been converted to deprecated macros that just call the
250 * The security callback, which can be customised by application code, supports
251 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
252 in the "other" parameter. In most places this is what is passed. All these
253 places occur server side. However there was one client side call of this
254 security operation and it passed a DH object instead. This is incorrect
255 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
256 of the other locations. Therefore this client side call has been changed to
257 pass an EVP_PKEY instead.
261 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
262 interface. Their functionality remains unchanged.
266 * Added new option for 'openssl list', '-providers', which will display the
267 list of loaded providers, their names, version and status. It optionally
268 displays their gettable parameters.
272 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
273 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
274 type is internally recognised so the workaround is no longer needed.
276 Functionality is still retained as it is, but will only work with
277 EVP_PKEYs with a legacy internal key.
281 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
282 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
286 * Changed all "STACK" functions to be macros instead of inline functions. Macro
287 parameters are still checked for type safety at compile time via helper
292 * Remove the RAND_DRBG API
294 The RAND_DRBG API did not fit well into the new provider concept as
295 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
296 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
297 and some of its API calls are rather low-level. This holds in particular
298 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
300 Adding a compatibility layer to continue supporting the RAND_DRBG API as
301 a legacy API for a regular deprecation period turned out to come at the
302 price of complicating the new provider API unnecessarily. Since the
303 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
306 *Paul Dale and Matthias St. Pierre*
308 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
309 as well as actual hostnames.
313 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
314 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
315 conversely, silently ignore DTLS protocol version bounds when configuring
316 TLS-based contexts. The commands can be repeated to set bounds of both
317 types. The same applies with the corresponding "min_protocol" and
318 "max_protocol" command-line switches, in case some application uses both TLS
321 SSL_CTX instances that are created for a fixed protocol version (e.g.
322 `TLSv1_server_method()`) also silently ignore version bounds. Previously
323 attempts to apply bounds to these protocol versions would result in an
324 error. Now only the "version-flexible" SSL_CTX instances are subject to
325 limits in configuration files in command-line options.
329 * Deprecated the `ENGINE` API. Engines should be replaced with providers
334 * Reworked the recorded ERR codes to make better space for system errors.
335 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
336 given code is a system error (true) or an OpenSSL error (false).
340 * Reworked the test perl framework to better allow parallel testing.
342 *Nicola Tuveri and David von Oheimb*
344 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
345 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
349 * 'Configure' has been changed to figure out the configuration target if
350 none is given on the command line. Consequently, the 'config' script is
351 now only a mere wrapper. All documentation is changed to only mention
354 *Rich Salz and Richard Levitte*
356 * Added a library context `OSSL_LIB_CTX` that applications as well as
357 other libraries can use to form a separate context within which
358 libcrypto operations are performed.
360 There are two ways this can be used:
362 - Directly, by passing a library context to functions that take
363 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
365 - Indirectly, by creating a new library context and then assigning
366 it as the new default, with `OSSL_LIB_CTX_set0_default`.
368 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
369 apart from the functions directly related to `OSSL_LIB_CTX`, accept
370 NULL to indicate that the default library context should be used.
372 Library code that changes the default library context using
373 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
374 second call before returning to the caller.
376 _(Note: the library context was initially called `OPENSSL_CTX` and
377 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
381 * Handshake now fails if Extended Master Secret extension is dropped
386 * Dropped interactive mode from the `openssl` program. From now on,
387 running it without arguments is equivalent to `openssl help`.
391 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
392 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
393 While the old function names have been retained for backward compatibility
394 they should not be used in new developments
395 because their return values are confusing: Unlike other `_cmp()` functions
396 they do not return 0 in case their arguments are equal.
400 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
401 `EC_GROUP_get_field_type()`.
405 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
406 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
407 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
408 Applications should rely on the library automatically assigning a suitable
409 EC_METHOD internally upon EC_GROUP construction.
413 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
414 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
415 assigned internally without application intervention.
416 Users of EC_GROUP_new() should switch to a different suitable constructor.
420 * Add CAdES-BES signature verification support, mostly derived
421 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
423 *Filipe Raimundo da Silva*
425 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
429 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
430 functions are not widely used and now OpenSSL automatically perform this
431 conversion when needed.
435 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
436 EC_KEY_precompute_mult(). These functions are not widely used and
437 applications should instead switch to named curves which OpenSSL has
438 hardcoded lookup tables for.
442 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
443 should instead use the L<EC_POINT_mul(3)> function.
447 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
448 that are not applicable to the new provider model. Applications should
449 instead use EVP_default_properties_is_fips_enabled() and
450 EVP_default_properties_enable_fips().
454 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
455 is set, an unexpected EOF is ignored, it pretends a close notify was received
456 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
460 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
461 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
462 used and applications should instead use the
463 L<EC_POINT_set_affine_coordinates(3)> and
464 L<EC_POINT_get_affine_coordinates(3)> functions.
468 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
469 arrays to be more easily constructed via a series of utility functions.
470 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
471 the various push functions and finally convert to a passable OSSL_PARAM
472 array using OSSL_PARAM_BLD_to_param().
476 * The security strength of SHA1 and MD5 based signatures in TLS has been
477 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
478 working at the default security level of 1 and instead requires security
479 level 0. The security level can be changed either using the cipher string
480 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
484 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
485 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
486 internal keys, if they correspond to one of those built in types.
490 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
491 contain a provider side internal key.
495 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
496 They are old functions that we don't use, and that you could disable with
497 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
501 * Project text documents not yet having a proper file name extension
502 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
503 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
507 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
508 have been converted to Markdown with the goal to produce documents
509 which not only look pretty when viewed online in the browser, but
510 remain well readable inside a plain text editor.
512 To achieve this goal, a 'minimalistic' Markdown style has been applied
513 which avoids formatting elements that interfere too much with the
514 reading flow in the text file. For example, it
516 * avoids [ATX headings][] and uses [setext headings][] instead
517 (which works for `<h1>` and `<h2>` headings only).
518 * avoids [inline links][] and uses [reference links][] instead.
519 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
521 [ATX headings]: https://github.github.com/gfm/#atx-headings
522 [setext headings]: https://github.github.com/gfm/#setext-headings
523 [inline links]: https://github.github.com/gfm/#inline-link
524 [reference links]: https://github.github.com/gfm/#reference-link
525 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
526 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
528 *Matthias St. Pierre*
530 * The test suite is changed to preserve results of each test recipe.
531 A new directory test-runs/ with subdirectories named like the
532 test recipes are created in the build tree for this purpose.
536 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
537 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
538 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
540 *David von Oheimb, Martin Peylo*
542 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
543 The legacy OCSP-focused and only partly documented API is retained for
544 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
548 * Added `util/check-format.pl`, a tool for checking adherence to the
549 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
550 The checks performed are incomplete and yield some false positives.
551 Still the tool should be useful for detecting most typical glitches.
555 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
556 If domain name resolution yields multiple IP addresses all of them are tried
557 after `connect()` failures.
561 * All of the low level RSA functions have been deprecated including:
563 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
564 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
565 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
566 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
567 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
568 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
569 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
570 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
571 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
572 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
573 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
574 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
575 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
576 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
577 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
578 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
579 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
580 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
581 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
582 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
583 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
584 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
585 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
586 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
587 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
588 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
589 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
590 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
592 Use of these low level functions has been informally discouraged for a long
593 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
594 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
595 L<EVP_PKEY_decrypt(3)>.
599 * X509 certificates signed using SHA1 are no longer allowed at security
601 In TLS/SSL the default security level is 1. It can be set either
602 using the cipher string with `@SECLEVEL`, or calling
603 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
604 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
606 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
607 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
608 options of the commands.
612 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
613 modified to use PKEY APIs. These commands are now in maintenance mode
614 and no new features will be added to them.
618 * The command line utility rsautl has been deprecated.
619 Instead use the pkeyutl program.
623 * The command line utilities genrsa and rsa have been modified to use PKEY
624 APIs. They now write PKCS#8 keys by default. These commands are now in
625 maintenance mode and no new features will be added to them.
629 * All of the low level DH functions have been deprecated including:
631 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
632 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
633 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
634 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
635 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
636 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
637 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
638 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
639 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
640 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
641 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
642 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
643 DH_meth_get_generate_params and DH_meth_set_generate_params.
645 Use of these low level functions has been informally discouraged for a long
646 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
647 and L<EVP_PKEY_derive(3)>.
649 Additionally functions that read and write DH objects such as d2i_DHparams,
650 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
651 functions have also been deprecated. Applications should instead use the
652 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
654 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
655 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
656 `EVP_PKEY_set1_DH()` are also deprecated.
657 Applications should instead either read or write an
658 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
659 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
661 *Paul Dale and Matt Caswell*
663 * All of the low level DSA functions have been deprecated including:
665 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
666 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
667 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
668 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
669 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
670 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
671 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
672 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
673 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
674 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
675 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
676 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
677 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
678 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
679 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
680 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
681 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
683 Use of these low level functions has been informally discouraged for a long
684 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
685 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
689 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
690 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
691 This means that applications don't have to look at the curve NID and
692 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
693 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
694 a no-op when the EVP_PKEY is already of the given type.
696 Parameter and key generation is also reworked to make it possible
697 to generate EVP_PKEY_SM2 parameters and keys without having to go
698 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
699 However, code that does the latter will still work as before.
703 * Deprecated low level ECDH and ECDSA functions. These include:
705 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
706 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
709 Use of these low level functions has been informally discouraged for a long
710 time. Instead applications should use the EVP_PKEY_derive(3),
711 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
715 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
716 and EVP_PKEY_decrypt() instead.
717 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
718 and EVP_PKEY_encrypt() instead.
722 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
723 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
724 a new formulation to include all the things it can be used for,
725 as well as words of caution.
729 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
730 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
734 * All of the low level HMAC functions have been deprecated including:
736 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
737 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
740 Use of these low level functions has been informally discouraged for a long
741 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
742 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
743 and L<EVP_MAC_final(3)>.
747 * Over two thousand fixes were made to the documentation, including:
748 - Common options (such as -rand/-writerand, TLS version control, etc)
749 were refactored and point to newly-enhanced descriptions in openssl.pod.
750 - Added style conformance for all options (with help from Richard Levitte),
751 documented all reported missing options, added a CI build to check
752 that all options are documented and that no unimplemented options
754 - Documented some internals, such as all use of environment variables.
755 - Addressed all internal broken L<> references.
759 * All of the low level CMAC functions have been deprecated including:
761 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
762 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
764 Use of these low level functions has been informally discouraged for a long
765 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
766 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
767 and L<EVP_MAC_final(3)>.
771 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
772 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
775 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
776 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
777 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
778 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
779 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
780 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
781 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
782 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
783 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
784 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
786 Use of these low level functions has been informally discouraged
787 for a long time. Applications should use the EVP_DigestInit_ex(3),
788 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
792 * Corrected the documentation of the return values from the `EVP_DigestSign*`
793 set of functions. The documentation mentioned negative values for some
794 errors, but this was never the case, so the mention of negative values
797 Code that followed the documentation and thereby check with something
798 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
802 * All of the low level cipher functions have been deprecated including:
804 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
805 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
806 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
807 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
808 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
809 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
810 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
811 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
812 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
813 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
814 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
815 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
816 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
817 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
818 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
819 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
820 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
821 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
822 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
823 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
824 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
825 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
826 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
827 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
828 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
829 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
830 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
831 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
832 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
834 Use of these low level functions has been informally discouraged for
835 a long time. Applications should use the high level EVP APIs, e.g.
836 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
837 equivalently named decrypt functions instead.
839 *Matt Caswell and Paul Dale*
841 * Removed include/openssl/opensslconf.h.in and replaced it with
842 include/openssl/configuration.h.in, which differs in not including
843 <openssl/macros.h>. A short header include/openssl/opensslconf.h
844 was added to include both.
846 This allows internal hacks where one might need to modify the set
847 of configured macros, for example this if deprecated symbols are
848 still supposed to be available internally:
850 #include <openssl/configuration.h>
852 #undef OPENSSL_NO_DEPRECATED
853 #define OPENSSL_SUPPRESS_DEPRECATED
855 #include <openssl/macros.h>
857 This should not be used by applications that use the exported
858 symbols, as that will lead to linking errors.
862 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
863 used in exponentiation with 512-bit moduli. No EC algorithms are
864 affected. Analysis suggests that attacks against 2-prime RSA1024,
865 3-prime RSA1536, and DSA1024 as a result of this defect would be very
866 difficult to perform and are not believed likely. Attacks against DH512
867 are considered just feasible. However, for an attack the target would
868 have to re-use the DH512 private key, which is not recommended anyway.
869 Also applications directly using the low level API BN_mod_exp may be
870 affected if they use BN_FLG_CONSTTIME.
875 * Most memory-debug features have been deprecated, and the functionality
876 replaced with no-ops.
880 * Added documentation for the STACK API.
884 * Introduced a new method type and API, OSSL_ENCODER, to
885 represent generic encoders. An implementation is expected to
886 be able to encode an object associated with a given name (such
887 as an algorithm name for an asymmetric key) into forms given by
888 implementation properties.
890 Encoders are primarily used from inside libcrypto, through
891 calls to functions like EVP_PKEY_print_private(),
892 PEM_write_bio_PrivateKey() and similar.
894 Encoders are specified in such a way that they can be made to
895 directly handle the provider side portion of an object, if this
896 provider side part comes from the same provider as the encoder
897 itself, but can also be made to handle objects in parametrized
898 form (as an OSSL_PARAM array of data). This allows a provider to
899 offer generic encoders as a service for any other provider.
903 * Added a .pragma directive to the syntax of configuration files, to
904 allow varying behavior in a supported and predictable manner.
905 Currently added pragma:
909 This allows dollar signs to be a keyword character unless it's
910 followed by a opening brace or parenthesis. This is useful for
911 platforms where dollar signs are commonly used in names, such as
912 volume names and system directory names on VMS.
916 * Added functionality to create an EVP_PKEY from user data. This
917 is effectively the same as creating a RSA, DH or DSA object and
918 then assigning them to an EVP_PKEY, but directly using algorithm
919 agnostic EVP functions. A benefit is that this should be future
920 proof for public key algorithms to come.
924 * Change the interpretation of the '--api' configuration option to
925 mean that this is a desired API compatibility level with no
926 further meaning. The previous interpretation, that this would
927 also mean to remove all deprecated symbols up to and including
928 the given version, no requires that 'no-deprecated' is also used
929 in the configuration.
931 When building applications, the desired API compatibility level
932 can be set with the OPENSSL_API_COMPAT macro like before. For
933 API compatibility version below 3.0, the old style numerical
934 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
935 For version 3.0 and on, the value is expected to be the decimal
936 value calculated from the major and minor version like this:
938 MAJOR * 10000 + MINOR * 100
942 -DOPENSSL_API_COMPAT=30000 For 3.0
943 -DOPENSSL_API_COMPAT=30200 For 3.2
945 To hide declarations that are deprecated up to and including the
946 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
947 given when building the application as well.
951 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
952 access to certificate and CRL stores via URIs and OSSL_STORE
955 This adds the following functions:
957 - X509_LOOKUP_store()
958 - X509_STORE_load_file()
959 - X509_STORE_load_path()
960 - X509_STORE_load_store()
961 - SSL_add_store_cert_subjects_to_stack()
962 - SSL_CTX_set_default_verify_store()
963 - SSL_CTX_load_verify_file()
964 - SSL_CTX_load_verify_dir()
965 - SSL_CTX_load_verify_store()
969 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
970 The presence of this system service is determined at run-time.
974 * Added functionality to create an EVP_PKEY context based on data
975 for methods from providers. This takes an algorithm name and a
976 property query string and simply stores them, with the intent
977 that any operation that uses this context will use those strings
978 to fetch the needed methods implicitly, thereby making the port
979 of application written for pre-3.0 OpenSSL easier.
983 * The undocumented function NCONF_WIN32() has been deprecated; for
984 conversion details see the HISTORY section of doc/man5/config.pod
988 * Introduced the new functions EVP_DigestSignInit_ex() and
989 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
990 EVP_DigestVerifyUpdate() have been converted to functions. See the man
991 pages for further details.
995 * Over two thousand fixes were made to the documentation, including:
996 adding missing command flags, better style conformance, documentation
999 *Rich Salz, Richard Levitte*
1001 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1002 X25519, X448, Ed25519 and Ed448.
1006 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1011 * Deprecated the public definition of `ERR_STATE` as well as the function
1012 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1017 * Added ERR functionality to give callers access to the stored function
1018 names that have replaced the older function code based functions.
1020 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1021 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1022 ERR_peek_error_all() and ERR_peek_last_error_all().
1024 These functions have become deprecated: ERR_get_error_line(),
1025 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1026 ERR_peek_last_error_line_data() and ERR_func_error_string().
1028 Users are recommended to use ERR_get_error_all(), or to pick information
1029 with ERR_peek functions and finish off with getting the error code by using
1034 * Extended testing to be verbose for failing tests only. The make variables
1035 VERBOSE_FAILURE or VF can be used to enable this:
1037 $ make VF=1 test # Unix
1038 $ mms /macro=(VF=1) test ! OpenVMS
1039 $ nmake VF=1 test # Windows
1043 * Added the `-copy_extensions` option to the `x509` command for use with
1044 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1045 all extensions in the request are copied to the certificate or vice versa.
1047 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1049 * Added the `-copy_extensions` option to the `req` command for use with
1050 `-x509`. When given with the `copy` or `copyall` argument,
1051 all extensions in the certification request are copied to the certificate.
1055 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1056 they generate are by default RFC 5280 compliant in the following sense:
1057 There is a subjectKeyIdentifier extension with a hash value of the public key
1058 and for not self-signed certs there is an authorityKeyIdentifier extension
1059 with a keyIdentifier field or issuer information identifying the signing key.
1060 This is done unless some configuration overrides the new default behavior,
1061 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1065 * Added several checks to `X509_verify_cert()` according to requirements in
1066 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1067 (which may be done by using the CLI option `-x509_strict`):
1068 * The basicConstraints of CA certificates must be marked critical.
1069 * CA certificates must explicitly include the keyUsage extension.
1070 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1071 * The issuer name of any certificate must not be empty.
1072 * The subject name of CA certs, certs with keyUsage crlSign,
1073 and certs without subjectAlternativeName must not be empty.
1074 * If a subjectAlternativeName extension is given it must not be empty.
1075 * The signatureAlgorithm field and the cert signature must be consistent.
1076 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1077 must not be marked critical.
1078 * The authorityKeyIdentifier must be given for X.509v3 certs
1079 unless they are self-signed.
1080 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1084 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1085 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1089 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1090 used even when parsing explicit parameters, when loading a encoded key
1091 or calling `EC_GROUP_new_from_ecpkparameters()`/
1092 `EC_GROUP_new_from_ecparameters()`.
1093 This prevents bypass of security hardening and performance gains,
1094 especially for curves with specialized EC_METHODs.
1095 By default, if a key encoded with explicit parameters is loaded and later
1096 encoded, the output is still encoded with explicit parameters, even if
1097 internally a "named" EC_GROUP is used for computation.
1101 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1102 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1103 NULL. After this change, only the cofactor parameter can be NULL. It also
1104 does some minimal sanity checks on the passed order.
1109 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1110 An attack is simple, if the first CMS_recipientInfo is valid but the
1111 second CMS_recipientInfo is chosen ciphertext. If the second
1112 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1113 encryption key will be replaced by garbage, and the message cannot be
1114 decoded, but if the RSA decryption fails, the correct encryption key is
1115 used and the recipient will not notice the attack.
1116 As a work around for this potential attack the length of the decrypted
1117 key must be equal to the cipher default key length, in case the
1118 certifiate is not given and all recipientInfo are tried out.
1119 The old behaviour can be re-enabled in the CMS code by setting the
1120 CMS_DEBUG_DECRYPT flag.
1124 * Early start up entropy quality from the DEVRANDOM seed source has been
1125 improved for older Linux systems. The RAND subsystem will wait for
1126 /dev/random to be producing output before seeding from /dev/urandom.
1127 The seeded state is stored for future library initialisations using
1128 a system global shared memory segment. The shared memory identifier
1129 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1130 the desired value. The default identifier is 114.
1134 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1135 when primes for RSA keys are computed.
1136 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1137 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1138 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1139 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1140 This avoids possible fingerprinting of newly generated RSA modules.
1144 * Correct the extended master secret constant on EBCDIC systems. Without this
1145 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1146 negotiate EMS will fail. Unfortunately this also means that TLS connections
1147 between EBCDIC systems with this fix, and EBCDIC systems without this
1148 fix will fail if they negotiate EMS.
1152 * Changed the library initialisation so that the config file is now loaded
1153 by default. This was already the case for libssl. It now occurs for both
1154 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1155 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1159 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1160 where the former acts as a replacement for `ERR_put_error()`, and the
1161 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1162 `ERR_raise_data()` adds more flexibility by taking a format string and
1163 an arbitrary number of arguments following it, to be processed with
1168 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1169 to check if a named provider is loaded and available. When called, it
1170 will also activate all fallback providers if such are still present.
1174 * Enforce a minimum DH modulus size of 512 bits.
1178 * Changed DH parameters to generate the order q subgroup instead of 2q.
1179 Previously generated DH parameters are still accepted by DH_check
1180 but DH_generate_key works around that by clearing bit 0 of the
1181 private key for those. This avoids leaking bit 0 of the private key.
1185 * Significantly reduce secure memory usage by the randomness pools.
1189 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1194 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1195 algorithms. An implementation of a key exchange algorithm can be obtained
1196 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1197 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1198 the older EVP_PKEY_derive_init() function. See the man pages for the new
1199 functions for further details.
1203 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1207 * Removed the function names from error messages and deprecated the
1210 * Removed NextStep support and the macro OPENSSL_UNISTD
1214 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1215 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1216 Also removed "export var as function" capability; we do not export
1217 variables, only functions.
1221 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1222 an error and 1 indicating success. In previous versions of OpenSSL this
1223 was a void type. If a key was set longer than the maximum possible this
1228 * Support SM2 signing and verification schemes with X509 certificate.
1232 * Use SHA256 as the default digest for TS query in the `ts` app.
1236 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1237 This checks that the salt length is at least 128 bits, the derived key
1238 length is at least 112 bits, and that the iteration count is at least 1000.
1239 For backwards compatibility these checks are disabled by default in the
1240 default provider, but are enabled by default in the fips provider.
1241 To enable or disable these checks use the control
1242 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1246 * Default cipher lists/suites are now available via a function, the
1247 #defines are deprecated.
1251 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1252 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1253 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1257 * Join the directories crypto/x509 and crypto/x509v3
1261 * Added command 'openssl kdf' that uses the EVP_KDF API.
1265 * Added command 'openssl mac' that uses the EVP_MAC API.
1269 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1270 as default directories. Also added the command 'openssl info'
1271 for scripting purposes.
1275 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1276 deprecated. These undocumented functions were never integrated into the EVP
1277 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1278 Bi-directional IGE mode. These modes were never formally standardised and
1279 usage of these functions is believed to be very small. In particular
1280 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1281 is ever used. The security implications are believed to be minimal, but
1282 this issue was never fixed for backwards compatibility reasons. New code
1283 should not use these modes.
1287 * Add prediction resistance to the DRBG reseeding process.
1291 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1292 mandated by IEEE Std 1619-2018.
1296 * Added newline escaping functionality to a filename when using openssl dgst.
1297 This output format is to replicate the output format found in the `*sum`
1298 checksum programs. This aims to preserve backward compatibility.
1300 *Matt Eaton, Richard Levitte, and Paul Dale*
1302 * Removed the heartbeat message in DTLS feature, as it has very
1303 little usage and doesn't seem to fulfill a valuable purpose.
1304 The configuration option is now deprecated.
1308 * Changed the output of 'openssl {digestname} < file' to display the
1309 digest name in its output.
1313 * Added a new generic trace API which provides support for enabling
1314 instrumentation through trace output. This feature is mainly intended
1315 as an aid for developers and is disabled by default. To utilize it,
1316 OpenSSL needs to be configured with the `enable-trace` option.
1318 If the tracing API is enabled, the application can activate trace output
1319 by registering BIOs as trace channels for a number of tracing and debugging
1322 The `openssl` program has been expanded to enable any of the types
1323 available via environment variables defined by the user, and serves as
1324 one possible example on how to use this functionality.
1326 *Richard Levitte & Matthias St. Pierre*
1328 * Added build tests for C++. These are generated files that only do one
1329 thing, to include one public OpenSSL head file each. This tests that
1330 the public header files can be usefully included in a C++ application.
1332 This test isn't enabled by default. It can be enabled with the option
1333 'enable-buildtest-c++'.
1337 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1341 * Add KMAC to EVP_MAC.
1345 * Added property based algorithm implementation selection framework to
1350 * Added SCA hardening for modular field inversion in EC_GROUP through
1351 a new dedicated field_inv() pointer in EC_METHOD.
1352 This also addresses a leakage affecting conversions from projective
1353 to affine coordinates.
1355 *Billy Bob Brumley, Nicola Tuveri*
1357 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1358 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1359 those algorithms that were already supported through the EVP_PKEY API
1360 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1361 and scrypt are now wrappers that call EVP_KDF.
1365 * Build devcrypto engine as a dynamic engine.
1367 *Eneas U de Queiroz*
1369 * Add keyed BLAKE2 to EVP_MAC.
1373 * Fix a bug in the computation of the endpoint-pair shared secret used
1374 by DTLS over SCTP. This breaks interoperability with older versions
1375 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1376 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1377 interoperability with such broken implementations. However, enabling
1378 this switch breaks interoperability with correct implementations.
1380 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1381 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1385 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1389 * Change the license to the Apache License v2.0.
1393 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1395 - Major releases (indicated by incrementing the MAJOR release number)
1396 may introduce incompatible API/ABI changes.
1397 - Minor releases (indicated by incrementing the MINOR release number)
1398 may introduce new features but retain API/ABI compatibility.
1399 - Patch releases (indicated by incrementing the PATCH number)
1400 are intended for bug fixes and other improvements of existing
1401 features only (like improving performance or adding documentation)
1402 and retain API/ABI compatibility.
1406 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1410 * Remove the 'dist' target and add a tarball building script. The
1411 'dist' target has fallen out of use, and it shouldn't be
1412 necessary to configure just to create a source distribution.
1416 * Recreate the OS390-Unix config target. It no longer relies on a
1417 special script like it did for OpenSSL pre-1.1.0.
1421 * Instead of having the source directories listed in Configure, add
1422 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1427 * Add GMAC to EVP_MAC.
1431 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1435 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1436 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1437 to facilitate the continued use of MACs through raw private keys in
1438 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1442 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1443 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1447 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1448 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1449 are retained for backwards compatibility.
1453 * AES-XTS mode now enforces that its two keys are different to mitigate
1454 the attacked described in "Efficient Instantiations of Tweakable
1455 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1456 Details of this attack can be obtained from:
1457 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1461 * Rename the object files, i.e. give them other names than in previous
1462 versions. Their names now include the name of the final product, as
1463 well as its type mnemonic (bin, lib, shlib).
1467 * Added new option for 'openssl list', '-objects', which will display the
1468 list of built in objects, i.e. OIDs with names.
1472 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1473 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1478 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1479 improves application performance by removing data copies and providing
1480 applications with zero-copy system calls such as sendfile and splice.
1484 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1485 option is set, openssl cleanses (zeroize) plaintext bytes from
1486 internal buffers after delivering them to the application. Note,
1487 the application is still responsible for cleansing other copies
1488 (e.g.: data received by SSL_read(3)).
1492 * `PKCS12_parse` now maintains the order of the parsed certificates
1493 when outputting them via `*ca` (rather than reversing it).
1497 * Deprecated pthread fork support methods. These were unused so no
1498 replacement is required.
1500 - OPENSSL_fork_prepare()
1501 - OPENSSL_fork_parent()
1502 - OPENSSL_fork_child()
1509 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1511 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1513 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1514 create a unique hash value based on the issuer and serial number data
1515 contained within an X509 certificate. However it was failing to correctly
1516 handle any errors that may occur while parsing the issuer field (which might
1517 occur if the issuer field is maliciously constructed). This may subsequently
1518 result in a NULL pointer deref and a crash leading to a potential denial of
1524 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1525 padding mode to correctly check for rollback attacks. This is considered a
1526 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1531 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1532 functions. Previously they could overflow the output length argument in some
1533 cases where the input length is close to the maximum permissable length for
1534 an integer on the platform. In such cases the return value from the function
1535 call would be 1 (indicating success), but the output length value would be
1536 negative. This could cause applications to behave incorrectly or crash.
1541 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1542 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1543 could be exploited in a side channel attack to recover the password. Since
1544 the attack is local host only this is outside of the current OpenSSL
1545 threat model and therefore no CVE is assigned.
1547 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1552 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1554 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1555 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1556 If an attacker can control both items being compared then this could lead
1557 to a possible denial of service attack. OpenSSL itself uses the
1558 GENERAL_NAME_cmp function for two purposes:
1559 1) Comparing CRL distribution point names between an available CRL and a
1560 CRL distribution point embedded in an X509 certificate
1561 2) When verifying that a timestamp response token signer matches the
1562 timestamp authority name (exposed via the API functions
1563 TS_RESP_verify_response and TS_RESP_verify_token)
1568 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1570 * Certificates with explicit curve parameters are now disallowed in
1571 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1575 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1576 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1577 conversely, silently ignore DTLS protocol version bounds when configuring
1578 TLS-based contexts. The commands can be repeated to set bounds of both
1579 types. The same applies with the corresponding "min_protocol" and
1580 "max_protocol" command-line switches, in case some application uses both TLS
1583 SSL_CTX instances that are created for a fixed protocol version (e.g.
1584 TLSv1_server_method()) also silently ignore version bounds. Previously
1585 attempts to apply bounds to these protocol versions would result in an
1586 error. Now only the "version-flexible" SSL_CTX instances are subject to
1587 limits in configuration files in command-line options.
1591 * Handshake now fails if Extended Master Secret extension is dropped
1596 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1598 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1600 * Fixed segmentation fault in SSL_check_chain()
1601 Server or client applications that call the SSL_check_chain() function
1602 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1603 dereference as a result of incorrect handling of the
1604 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1605 or unrecognised signature algorithm is received from the peer. This could
1606 be exploited by a malicious peer in a Denial of Service attack.
1611 * Added AES consttime code for no-asm configurations
1612 an optional constant time support for AES was added
1613 when building openssl for no-asm.
1614 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1615 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1616 At this time this feature is by default disabled.
1617 It will be enabled by default in 3.0.
1621 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1623 * Revert the change of EOF detection while reading in libssl to avoid
1624 regressions in applications depending on the current way of reporting
1625 the EOF. As the existing method is not fully accurate the change to
1626 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1627 branch and will be present in the 3.0 release.
1631 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1632 when primes for RSA keys are computed.
1633 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1634 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1635 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1636 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1637 This avoids possible fingerprinting of newly generated RSA modules.
1641 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1643 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1644 while reading in libssl then we would report an error back to the
1645 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1646 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1647 therefore give a hint as to what went wrong.
1651 * Check that ed25519 and ed448 are allowed by the security level. Previously
1652 signature algorithms not using an MD were not being checked that they were
1653 allowed by the security level.
1657 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1658 was not quite right. The behaviour was not consistent between resumption
1659 and normal handshakes, and also not quite consistent with historical
1660 behaviour. The behaviour in various scenarios has been clarified and
1661 it has been updated to make it match historical behaviour as closely as
1666 * *[VMS only]* The header files that the VMS compilers include automatically,
1667 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1668 that the C++ compiler doesn't understand. This is a shortcoming in the
1669 compiler, but can be worked around with `__cplusplus` guards.
1671 C++ applications that use OpenSSL libraries must be compiled using the
1672 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1673 functions. Otherwise, only functions with symbols of less than 31
1674 characters can be used, as the linker will not be able to successfully
1675 resolve symbols with longer names.
1679 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1680 The presence of this system service is determined at run-time.
1684 * Added newline escaping functionality to a filename when using openssl dgst.
1685 This output format is to replicate the output format found in the `*sum`
1686 checksum programs. This aims to preserve backward compatibility.
1688 *Matt Eaton, Richard Levitte, and Paul Dale*
1690 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1695 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1697 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1698 number generator (RNG). This was intended to include protection in the
1699 event of a fork() system call in order to ensure that the parent and child
1700 processes did not share the same RNG state. However this protection was not
1701 being used in the default case.
1703 A partial mitigation for this issue is that the output from a high
1704 precision timer is mixed into the RNG state so the likelihood of a parent
1705 and child process sharing state is significantly reduced.
1707 If an application already calls OPENSSL_init_crypto() explicitly using
1708 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1711 *Matthias St. Pierre*
1713 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1714 used even when parsing explicit parameters, when loading a encoded key
1715 or calling `EC_GROUP_new_from_ecpkparameters()`/
1716 `EC_GROUP_new_from_ecparameters()`.
1717 This prevents bypass of security hardening and performance gains,
1718 especially for curves with specialized EC_METHODs.
1719 By default, if a key encoded with explicit parameters is loaded and later
1720 encoded, the output is still encoded with explicit parameters, even if
1721 internally a "named" EC_GROUP is used for computation.
1725 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1726 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1727 NULL. After this change, only the cofactor parameter can be NULL. It also
1728 does some minimal sanity checks on the passed order.
1733 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1734 An attack is simple, if the first CMS_recipientInfo is valid but the
1735 second CMS_recipientInfo is chosen ciphertext. If the second
1736 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1737 encryption key will be replaced by garbage, and the message cannot be
1738 decoded, but if the RSA decryption fails, the correct encryption key is
1739 used and the recipient will not notice the attack.
1740 As a work around for this potential attack the length of the decrypted
1741 key must be equal to the cipher default key length, in case the
1742 certifiate is not given and all recipientInfo are tried out.
1743 The old behaviour can be re-enabled in the CMS code by setting the
1744 CMS_DEBUG_DECRYPT flag.
1749 * Early start up entropy quality from the DEVRANDOM seed source has been
1750 improved for older Linux systems. The RAND subsystem will wait for
1751 /dev/random to be producing output before seeding from /dev/urandom.
1752 The seeded state is stored for future library initialisations using
1753 a system global shared memory segment. The shared memory identifier
1754 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1755 the desired value. The default identifier is 114.
1759 * Correct the extended master secret constant on EBCDIC systems. Without this
1760 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1761 negotiate EMS will fail. Unfortunately this also means that TLS connections
1762 between EBCDIC systems with this fix, and EBCDIC systems without this
1763 fix will fail if they negotiate EMS.
1767 * Use Windows installation paths in the mingw builds
1769 Mingw isn't a POSIX environment per se, which means that Windows
1770 paths should be used for installation.
1775 * Changed DH_check to accept parameters with order q and 2q subgroups.
1776 With order 2q subgroups the bit 0 of the private key is not secret
1777 but DH_generate_key works around that by clearing bit 0 of the
1778 private key for those. This avoids leaking bit 0 of the private key.
1782 * Significantly reduce secure memory usage by the randomness pools.
1786 * Revert the DEVRANDOM_WAIT feature for Linux systems
1788 The DEVRANDOM_WAIT feature added a select() call to wait for the
1789 /dev/random device to become readable before reading from the
1790 /dev/urandom device.
1792 It turned out that this change had negative side effects on
1793 performance which were not acceptable. After some discussion it
1794 was decided to revert this feature and leave it up to the OS
1795 resp. the platform maintainer to ensure a proper initialization
1796 during early boot time.
1798 *Matthias St. Pierre*
1800 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1802 * Add build tests for C++. These are generated files that only do one
1803 thing, to include one public OpenSSL head file each. This tests that
1804 the public header files can be usefully included in a C++ application.
1806 This test isn't enabled by default. It can be enabled with the option
1807 'enable-buildtest-c++'.
1811 * Enable SHA3 pre-hashing for ECDSA and DSA.
1815 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1816 This changes the size when using the `genpkey` command when no size is given.
1817 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1818 generation commands to use 2048 bits by default.
1822 * Reorganize the manual pages to consistently have RETURN VALUES,
1823 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1824 util/fix-doc-nits accordingly.
1826 *Paul Yang, Joshua Lock*
1828 * Add the missing accessor EVP_PKEY_get0_engine()
1832 * Have commands like `s_client` and `s_server` output the signature scheme
1833 along with other cipher suite parameters when debugging.
1837 * Make OPENSSL_config() error agnostic again.
1841 * Do the error handling in RSA decryption constant time.
1845 * Prevent over long nonces in ChaCha20-Poly1305.
1847 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1848 for every encryption operation. RFC 7539 specifies that the nonce value
1849 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1850 and front pads the nonce with 0 bytes if it is less than 12
1851 bytes. However it also incorrectly allows a nonce to be set of up to 16
1852 bytes. In this case only the last 12 bytes are significant and any
1853 additional leading bytes are ignored.
1855 It is a requirement of using this cipher that nonce values are
1856 unique. Messages encrypted using a reused nonce value are susceptible to
1857 serious confidentiality and integrity attacks. If an application changes
1858 the default nonce length to be longer than 12 bytes and then makes a
1859 change to the leading bytes of the nonce expecting the new value to be a
1860 new unique nonce then such an application could inadvertently encrypt
1861 messages with a reused nonce.
1863 Additionally the ignored bytes in a long nonce are not covered by the
1864 integrity guarantee of this cipher. Any application that relies on the
1865 integrity of these ignored leading bytes of a long nonce may be further
1866 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1867 is safe because no such use sets such a long nonce value. However user
1868 applications that use this cipher directly and set a non-default nonce
1869 length to be longer than 12 bytes may be vulnerable.
1871 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1877 * Add DEVRANDOM_WAIT feature for Linux systems
1879 On older Linux systems where the getrandom() system call is not available,
1880 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1881 Contrary to getrandom(), the /dev/urandom device will not block during
1882 early boot when the kernel CSPRNG has not been seeded yet.
1884 To mitigate this known weakness, use select() to wait for /dev/random to
1885 become readable before reading from /dev/urandom.
1887 * Ensure that SM2 only uses SM3 as digest algorithm
1891 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1893 * Change the info callback signals for the start and end of a post-handshake
1894 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1895 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1896 confused by this and assume that a TLSv1.2 renegotiation has started. This
1897 can break KeyUpdate handling. Instead we no longer signal the start and end
1898 of a post handshake message exchange (although the messages themselves are
1899 still signalled). This could break some applications that were expecting
1900 the old signals. However without this KeyUpdate is not usable for many
1905 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1907 * Timing vulnerability in DSA signature generation
1909 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1910 timing side channel attack. An attacker could use variations in the signing
1911 algorithm to recover the private key.
1913 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1918 * Timing vulnerability in ECDSA signature generation
1920 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1921 timing side channel attack. An attacker could use variations in the signing
1922 algorithm to recover the private key.
1924 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1929 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1930 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1931 of two gigabytes and the error handling improved.
1933 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1934 categorized as a normal bug, not a security issue, because the DRBG reseeds
1935 automatically and is fully functional even without additional randomness
1936 provided by the application.
1938 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1940 * Add a new ClientHello callback. Provides a callback interface that gives
1941 the application the ability to adjust the nascent SSL object at the
1942 earliest stage of ClientHello processing, immediately after extensions have
1943 been collected but before they have been processed. In particular, this
1944 callback can adjust the supported TLS versions in response to the contents
1949 * Add SM2 base algorithm support.
1953 * s390x assembly pack: add (improved) hardware-support for the following
1954 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1955 aes-cfb/cfb8, aes-ecb.
1959 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1960 parameter is no longer accepted, as it leads to a corrupt table. NULL
1961 pem_str is reserved for alias entries only.
1965 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1966 step for prime curves. The new implementation is based on formulae from
1967 differential addition-and-doubling in homogeneous projective coordinates
1968 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1969 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1970 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1971 to work in projective coordinates.
1973 *Billy Bob Brumley, Nicola Tuveri*
1975 * Change generating and checking of primes so that the error rate of not
1976 being prime depends on the intended use based on the size of the input.
1977 For larger primes this will result in more rounds of Miller-Rabin.
1978 The maximal error rate for primes with more than 1080 bits is lowered
1981 *Kurt Roeckx, Annie Yousar*
1983 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1987 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1988 moving between systems, and to avoid confusion when a Windows build is
1989 done with mingw vs with MSVC. For POSIX installs, there's still a
1990 symlink or copy named 'tsget' to avoid that confusion as well.
1994 * Revert blinding in ECDSA sign and instead make problematic addition
1995 length-invariant. Switch even to fixed-length Montgomery multiplication.
1999 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2000 step for binary curves. The new implementation is based on formulae from
2001 differential addition-and-doubling in mixed Lopez-Dahab projective
2002 coordinates, modified to independently blind the operands.
2004 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2006 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2007 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2008 EC_METHODs to implement their own specialized "ladder step", to take
2009 advantage of more favorable coordinate systems or more efficient
2010 differential addition-and-doubling algorithms.
2012 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2014 * Modified the random device based seed sources to keep the relevant
2015 file descriptors open rather than reopening them on each access.
2016 This allows such sources to operate in a chroot() jail without
2017 the associated device nodes being available. This behaviour can be
2018 controlled using RAND_keep_random_devices_open().
2022 * Numerous side-channel attack mitigations have been applied. This may have
2023 performance impacts for some algorithms for the benefit of improved
2024 security. Specific changes are noted in this change log by their respective
2029 * AIX shared library support overhaul. Switch to AIX "natural" way of
2030 handling shared libraries, which means collecting shared objects of
2031 different versions and bitnesses in one common archive. This allows to
2032 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2033 doesn't affect the way 3rd party applications are linked, only how
2034 multi-version installation is managed.
2038 * Make ec_group_do_inverse_ord() more robust and available to other
2039 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2040 mitigations are applied to the fallback BN_mod_inverse().
2041 When using this function rather than BN_mod_inverse() directly, new
2042 EC cryptosystem implementations are then safer-by-default.
2046 * Add coordinate blinding for EC_POINT and implement projective
2047 coordinate blinding for generic prime curves as a countermeasure to
2048 chosen point SCA attacks.
2050 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2052 * Add blinding to ECDSA and DSA signatures to protect against side channel
2053 attacks discovered by Keegan Ryan (NCC Group).
2057 * Enforce checking in the `pkeyutl` command to ensure that the input
2058 length does not exceed the maximum supported digest length when performing
2059 a sign, verify or verifyrecover operation.
2063 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2064 I/O in combination with something like select() or poll() will hang. This
2065 can be turned off again using SSL_CTX_clear_mode().
2066 Many applications do not properly handle non-application data records, and
2067 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2068 around the problems in those applications, but can also break some.
2069 It's recommended to read the manpages about SSL_read(), SSL_write(),
2070 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2071 SSL_CTX_set_read_ahead() again.
2075 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2076 now allow empty (zero character) pass phrases.
2080 * Apply blinding to binary field modular inversion and remove patent
2081 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2085 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2086 binary and prime elliptic curves.
2090 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2091 constant time fixed point multiplication.
2095 * Revise elliptic curve scalar multiplication with timing attack
2096 defenses: ec_wNAF_mul redirects to a constant time implementation
2097 when computing fixed point and variable point multiplication (which
2098 in OpenSSL are mostly used with secret scalars in keygen, sign,
2099 ECDH derive operations).
2100 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2103 * Updated CONTRIBUTING
2107 * Updated DRBG / RAND to request nonce and additional low entropy
2108 randomness from the system.
2110 *Matthias St. Pierre*
2112 * Updated 'openssl rehash' to use OpenSSL consistent default.
2116 * Moved the load of the ssl_conf module to libcrypto, which helps
2117 loading engines that libssl uses before libssl is initialised.
2121 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2125 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2127 *Ingo Schwarze, Rich Salz*
2129 * Added output of accepting IP address and port for 'openssl s_server'
2133 * Added a new API for TLSv1.3 ciphersuites:
2134 SSL_CTX_set_ciphersuites()
2135 SSL_set_ciphersuites()
2139 * Memory allocation failures consistently add an error to the error
2144 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2145 in libcrypto when run as setuid/setgid.
2149 * Load any config file by default when libssl is used.
2153 * Added new public header file <openssl/rand_drbg.h> and documentation
2154 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2156 *Matthias St. Pierre*
2158 * QNX support removed (cannot find contributors to get their approval
2159 for the license change).
2163 * TLSv1.3 replay protection for early data has been implemented. See the
2164 SSL_read_early_data() man page for further details.
2168 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2169 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2170 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2171 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2172 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2173 configuration has been separated out. See the ciphers man page or the
2174 SSL_CTX_set_ciphersuites() man page for more information.
2178 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2179 in responder mode now supports the new "-multi" option, which
2180 spawns the specified number of child processes to handle OCSP
2181 requests. The "-timeout" option now also limits the OCSP
2182 responder's patience to wait to receive the full client request
2183 on a newly accepted connection. Child processes are respawned
2184 as needed, and the CA index file is automatically reloaded
2185 when changed. This makes it possible to run the "ocsp" responder
2186 as a long-running service, making the OpenSSL CA somewhat more
2187 feature-complete. In this mode, most diagnostic messages logged
2188 after entering the event loop are logged via syslog(3) rather than
2193 * Added support for X448 and Ed448. Heavily based on original work by
2198 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2199 objects loaded. This adds the functions OSSL_STORE_expect() and
2200 OSSL_STORE_find() as well as needed tools to construct searches and
2201 get the search data out of them.
2205 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2206 version of OpenSSL should review their configuration settings to ensure
2207 that they are still appropriate for TLSv1.3. For further information see:
2208 <https://wiki.openssl.org/index.php/TLS1.3>
2212 * Grand redesign of the OpenSSL random generator
2214 The default RAND method now utilizes an AES-CTR DRBG according to
2215 NIST standard SP 800-90Ar1. The new random generator is essentially
2216 a port of the default random generator from the OpenSSL FIPS 2.0
2217 object module. It is a hybrid deterministic random bit generator
2218 using an AES-CTR bit stream and which seeds and reseeds itself
2219 automatically using trusted system entropy sources.
2221 Some of its new features are:
2222 - Support for multiple DRBG instances with seed chaining.
2223 - The default RAND method makes use of a DRBG.
2224 - There is a public and private DRBG instance.
2225 - The DRBG instances are fork-safe.
2226 - Keep all global DRBG instances on the secure heap if it is enabled.
2227 - The public and private DRBG instance are per thread for lock free
2230 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2232 * Changed Configure so it only says what it does and doesn't dump
2233 so much data. Instead, ./configdata.pm should be used as a script
2234 to display all sorts of configuration data.
2238 * Added processing of "make variables" to Configure.
2242 * Added SHA512/224 and SHA512/256 algorithm support.
2246 * The last traces of Netware support, first removed in 1.1.0, have
2251 * Get rid of Makefile.shared, and in the process, make the processing
2252 of certain files (rc.obj, or the .def/.map/.opt files produced from
2253 the ordinal files) more visible and hopefully easier to trace and
2254 debug (or make silent).
2258 * Make it possible to have environment variable assignments as
2259 arguments to config / Configure.
2263 * Add multi-prime RSA (RFC 8017) support.
2267 * Add SM3 implemented according to GB/T 32905-2016
2268 *Jack Lloyd <jack.lloyd@ribose.com>,*
2269 *Ronald Tse <ronald.tse@ribose.com>,*
2270 *Erick Borsboom <erick.borsboom@ribose.com>*
2272 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2273 as documented in RFC6066.
2274 Based on a patch from Tomasz Moń
2276 *Filipe Raimundo da Silva*
2278 * Add SM4 implemented according to GB/T 32907-2016.
2279 *Jack Lloyd <jack.lloyd@ribose.com>,*
2280 *Ronald Tse <ronald.tse@ribose.com>,*
2281 *Erick Borsboom <erick.borsboom@ribose.com>*
2283 * Reimplement -newreq-nodes and ERR_error_string_n; the
2284 original author does not agree with the license change.
2288 * Add ARIA AEAD TLS support.
2292 * Some macro definitions to support VS6 have been removed. Visual
2293 Studio 6 has not worked since 1.1.0
2297 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2298 without clearing the errors.
2302 * Add "atfork" functions. If building on a system that without
2303 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2304 requirements. The RAND facility now uses/requires this.
2312 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2313 not possible to disable entirely. However, it's still possible to
2314 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2317 To disable, configure with 'no-ui-console'. 'no-ui' is still
2318 possible to use as an alias. Check at compile time with the
2319 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2320 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2324 * Add a STORE module, which implements a uniform and URI based reader of
2325 stores that can contain keys, certificates, CRLs and numerous other
2326 objects. The main API is loosely based on a few stdio functions,
2327 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2328 OSSL_STORE_error and OSSL_STORE_close.
2329 The implementation uses backends called "loaders" to implement arbitrary
2330 URI schemes. There is one built in "loader" for the 'file' scheme.
2334 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2335 then adjusted to work on FreeBSD 8.4 as well.
2336 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2337 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2341 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2342 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2343 error code calls like this:
2345 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2347 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2348 that can be encoded in C. For the foreseeable future, this will only
2351 *Richard Levitte and Tim Hudson*
2353 * Removed BSD cryptodev engine.
2357 * Add a build target 'build_all_generated', to build all generated files
2358 and only that. This can be used to prepare everything that requires
2359 things like perl for a system that lacks perl and then move everything
2360 to that system and do the rest of the build there.
2364 * In the UI interface, make it possible to duplicate the user data. This
2365 can be used by engines that need to retain the data for a longer time
2366 than just the call where this user data is passed.
2370 * Ignore the '-named_curve auto' value for compatibility of applications
2373 *Tomáš Mráz <tmraz@fedoraproject.org>*
2375 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2376 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2377 alerts across multiple records (some of which could be empty). In practice
2378 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2379 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2380 support this at all. Supporting it adds significant complexity to the
2381 record layer, and its removal is unlikely to cause interoperability
2386 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2387 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2388 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2393 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2394 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2396 *Richard Levitte, Andy Polyakov*
2398 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2403 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2404 platform rather than 'mingw'.
2408 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2409 success if they are asked to add an object which already exists
2410 in the store. This change cascades to other functions which load
2411 certificates and CRLs.
2415 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2416 facilitate stack unwinding even from assembly subroutines.
2420 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2421 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2425 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2426 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2427 which is the minimum version we support.
2431 * Certificate time validation (X509_cmp_time) enforces stricter
2432 compliance with RFC 5280. Fractional seconds and timezone offsets
2433 are no longer allowed.
2437 * Add support for ARIA
2441 * s_client will now send the Server Name Indication (SNI) extension by
2442 default unless the new "-noservername" option is used. The server name is
2443 based on the host provided to the "-connect" option unless overridden by
2444 using "-servername".
2448 * Add support for SipHash
2452 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2453 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2454 prevent issues where no progress is being made and the peer continually
2455 sends unrecognised record types, using up resources processing them.
2459 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2460 using the algorithm defined in
2461 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2465 * Heartbeat support has been removed; the ABI is changed for now.
2467 *Richard Levitte, Rich Salz*
2469 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2473 * The RSA "null" method, which was partially supported to avoid patent
2474 issues, has been replaced to always returns NULL.
2481 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2483 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2484 used even when parsing explicit parameters, when loading a encoded key
2485 or calling `EC_GROUP_new_from_ecpkparameters()`/
2486 `EC_GROUP_new_from_ecparameters()`.
2487 This prevents bypass of security hardening and performance gains,
2488 especially for curves with specialized EC_METHODs.
2489 By default, if a key encoded with explicit parameters is loaded and later
2490 encoded, the output is still encoded with explicit parameters, even if
2491 internally a "named" EC_GROUP is used for computation.
2495 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2496 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2497 NULL. After this change, only the cofactor parameter can be NULL. It also
2498 does some minimal sanity checks on the passed order.
2503 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2504 An attack is simple, if the first CMS_recipientInfo is valid but the
2505 second CMS_recipientInfo is chosen ciphertext. If the second
2506 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2507 encryption key will be replaced by garbage, and the message cannot be
2508 decoded, but if the RSA decryption fails, the correct encryption key is
2509 used and the recipient will not notice the attack.
2510 As a work around for this potential attack the length of the decrypted
2511 key must be equal to the cipher default key length, in case the
2512 certifiate is not given and all recipientInfo are tried out.
2513 The old behaviour can be re-enabled in the CMS code by setting the
2514 CMS_DEBUG_DECRYPT flag.
2519 * Use Windows installation paths in the mingw builds
2521 Mingw isn't a POSIX environment per se, which means that Windows
2522 paths should be used for installation.
2527 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2529 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2530 This changes the size when using the `genpkey` command when no size is given.
2531 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2532 generation commands to use 2048 bits by default.
2536 * Prevent over long nonces in ChaCha20-Poly1305.
2538 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2539 for every encryption operation. RFC 7539 specifies that the nonce value
2540 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2541 and front pads the nonce with 0 bytes if it is less than 12
2542 bytes. However it also incorrectly allows a nonce to be set of up to 16
2543 bytes. In this case only the last 12 bytes are significant and any
2544 additional leading bytes are ignored.
2546 It is a requirement of using this cipher that nonce values are
2547 unique. Messages encrypted using a reused nonce value are susceptible to
2548 serious confidentiality and integrity attacks. If an application changes
2549 the default nonce length to be longer than 12 bytes and then makes a
2550 change to the leading bytes of the nonce expecting the new value to be a
2551 new unique nonce then such an application could inadvertently encrypt
2552 messages with a reused nonce.
2554 Additionally the ignored bytes in a long nonce are not covered by the
2555 integrity guarantee of this cipher. Any application that relies on the
2556 integrity of these ignored leading bytes of a long nonce may be further
2557 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2558 is safe because no such use sets such a long nonce value. However user
2559 applications that use this cipher directly and set a non-default nonce
2560 length to be longer than 12 bytes may be vulnerable.
2562 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2568 * Added SCA hardening for modular field inversion in EC_GROUP through
2569 a new dedicated field_inv() pointer in EC_METHOD.
2570 This also addresses a leakage affecting conversions from projective
2571 to affine coordinates.
2573 *Billy Bob Brumley, Nicola Tuveri*
2575 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2576 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2580 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2584 * Remove the 'dist' target and add a tarball building script. The
2585 'dist' target has fallen out of use, and it shouldn't be
2586 necessary to configure just to create a source distribution.
2590 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2592 * Timing vulnerability in DSA signature generation
2594 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2595 timing side channel attack. An attacker could use variations in the signing
2596 algorithm to recover the private key.
2598 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2603 * Timing vulnerability in ECDSA signature generation
2605 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2606 timing side channel attack. An attacker could use variations in the signing
2607 algorithm to recover the private key.
2609 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2614 * Add coordinate blinding for EC_POINT and implement projective
2615 coordinate blinding for generic prime curves as a countermeasure to
2616 chosen point SCA attacks.
2618 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2620 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2622 * Client DoS due to large DH parameter
2624 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2625 malicious server can send a very large prime value to the client. This will
2626 cause the client to spend an unreasonably long period of time generating a
2627 key for this prime resulting in a hang until the client has finished. This
2628 could be exploited in a Denial Of Service attack.
2630 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2635 * Cache timing vulnerability in RSA Key Generation
2637 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2638 a cache timing side channel attack. An attacker with sufficient access to
2639 mount cache timing attacks during the RSA key generation process could
2640 recover the private key.
2642 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2643 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2648 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2649 parameter is no longer accepted, as it leads to a corrupt table. NULL
2650 pem_str is reserved for alias entries only.
2654 * Revert blinding in ECDSA sign and instead make problematic addition
2655 length-invariant. Switch even to fixed-length Montgomery multiplication.
2659 * Change generating and checking of primes so that the error rate of not
2660 being prime depends on the intended use based on the size of the input.
2661 For larger primes this will result in more rounds of Miller-Rabin.
2662 The maximal error rate for primes with more than 1080 bits is lowered
2665 *Kurt Roeckx, Annie Yousar*
2667 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2671 * Add blinding to ECDSA and DSA signatures to protect against side channel
2672 attacks discovered by Keegan Ryan (NCC Group).
2676 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2677 now allow empty (zero character) pass phrases.
2681 * Certificate time validation (X509_cmp_time) enforces stricter
2682 compliance with RFC 5280. Fractional seconds and timezone offsets
2683 are no longer allowed.
2687 * Fixed a text canonicalisation bug in CMS
2689 Where a CMS detached signature is used with text content the text goes
2690 through a canonicalisation process first prior to signing or verifying a
2691 signature. This process strips trailing space at the end of lines, converts
2692 line terminators to CRLF and removes additional trailing line terminators
2693 at the end of a file. A bug in the canonicalisation process meant that
2694 some characters, such as form-feed, were incorrectly treated as whitespace
2695 and removed. This is contrary to the specification (RFC5485). This fix
2696 could mean that detached text data signed with an earlier version of
2697 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2698 signed with a fixed OpenSSL may fail to verify with an earlier version of
2699 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2700 and use the "-binary" flag (for the "cms" command line application) or set
2701 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2705 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2707 * Constructed ASN.1 types with a recursive definition could exceed the stack
2709 Constructed ASN.1 types with a recursive definition (such as can be found
2710 in PKCS7) could eventually exceed the stack given malicious input with
2711 excessive recursion. This could result in a Denial Of Service attack. There
2712 are no such structures used within SSL/TLS that come from untrusted sources
2713 so this is considered safe.
2715 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2721 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2723 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2724 effectively reduced to only comparing the least significant bit of each
2725 byte. This allows an attacker to forge messages that would be considered as
2726 authenticated in an amount of tries lower than that guaranteed by the
2727 security claims of the scheme. The module can only be compiled by the
2728 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2730 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2736 * Add a build target 'build_all_generated', to build all generated files
2737 and only that. This can be used to prepare everything that requires
2738 things like perl for a system that lacks perl and then move everything
2739 to that system and do the rest of the build there.
2743 * Backport SSL_OP_NO_RENGOTIATION
2745 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2746 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2747 changes this is no longer possible in 1.1.0. Therefore the new
2748 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2749 1.1.0 to provide equivalent functionality.
2751 Note that if an application built against 1.1.0h headers (or above) is run
2752 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2753 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2757 * Removed the OS390-Unix config target. It relied on a script that doesn't
2762 * rsaz_1024_mul_avx2 overflow bug on x86_64
2764 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2765 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2766 Analysis suggests that attacks against RSA and DSA as a result of this
2767 defect would be very difficult to perform and are not believed likely.
2768 Attacks against DH1024 are considered just feasible, because most of the
2769 work necessary to deduce information about a private key may be performed
2770 offline. The amount of resources required for such an attack would be
2771 significant. However, for an attack on TLS to be meaningful, the server
2772 would have to share the DH1024 private key among multiple clients, which is
2773 no longer an option since CVE-2016-0701.
2775 This only affects processors that support the AVX2 but not ADX extensions
2776 like Intel Haswell (4th generation).
2778 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2779 was originally found via the OSS-Fuzz project.
2784 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2786 * bn_sqrx8x_internal carry bug on x86_64
2788 There is a carry propagating bug in the x86_64 Montgomery squaring
2789 procedure. No EC algorithms are affected. Analysis suggests that attacks
2790 against RSA and DSA as a result of this defect would be very difficult to
2791 perform and are not believed likely. Attacks against DH are considered just
2792 feasible (although very difficult) because most of the work necessary to
2793 deduce information about a private key may be performed offline. The amount
2794 of resources required for such an attack would be very significant and
2795 likely only accessible to a limited number of attackers. An attacker would
2796 additionally need online access to an unpatched system using the target
2797 private key in a scenario with persistent DH parameters and a private
2798 key that is shared between multiple clients.
2800 This only affects processors that support the BMI1, BMI2 and ADX extensions
2801 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2803 This issue was reported to OpenSSL by the OSS-Fuzz project.
2808 * Malformed X.509 IPAddressFamily could cause OOB read
2810 If an X.509 certificate has a malformed IPAddressFamily extension,
2811 OpenSSL could do a one-byte buffer overread. The most likely result
2812 would be an erroneous display of the certificate in text format.
2814 This issue was reported to OpenSSL by the OSS-Fuzz project.
2819 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2821 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2822 platform rather than 'mingw'.
2826 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2827 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2828 which is the minimum version we support.
2832 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2834 * Encrypt-Then-Mac renegotiation crash
2836 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2837 negotiated where it was not in the original handshake (or vice-versa) then
2838 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2839 and servers are affected.
2841 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2846 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2848 * Truncated packet could crash via OOB read
2850 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2851 cipher is being used, then a truncated packet can cause that host to
2852 perform an out-of-bounds read, usually resulting in a crash.
2854 This issue was reported to OpenSSL by Robert Święcki of Google.
2859 * Bad (EC)DHE parameters cause a client crash
2861 If a malicious server supplies bad parameters for a DHE or ECDHE key
2862 exchange then this can result in the client attempting to dereference a
2863 NULL pointer leading to a client crash. This could be exploited in a Denial
2866 This issue was reported to OpenSSL by Guido Vranken.
2871 * BN_mod_exp may produce incorrect results on x86_64
2873 There is a carry propagating bug in the x86_64 Montgomery squaring
2874 procedure. No EC algorithms are affected. Analysis suggests that attacks
2875 against RSA and DSA as a result of this defect would be very difficult to
2876 perform and are not believed likely. Attacks against DH are considered just
2877 feasible (although very difficult) because most of the work necessary to
2878 deduce information about a private key may be performed offline. The amount
2879 of resources required for such an attack would be very significant and
2880 likely only accessible to a limited number of attackers. An attacker would
2881 additionally need online access to an unpatched system using the target
2882 private key in a scenario with persistent DH parameters and a private
2883 key that is shared between multiple clients. For example this can occur by
2884 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2885 similar to CVE-2015-3193 but must be treated as a separate problem.
2887 This issue was reported to OpenSSL by the OSS-Fuzz project.
2892 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2894 * ChaCha20/Poly1305 heap-buffer-overflow
2896 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2897 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2898 crash. This issue is not considered to be exploitable beyond a DoS.
2900 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2905 * CMS Null dereference
2907 Applications parsing invalid CMS structures can crash with a NULL pointer
2908 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2909 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2910 structure callback if an attempt is made to free certain invalid encodings.
2911 Only CHOICE structures using a callback which do not handle NULL value are
2914 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2919 * Montgomery multiplication may produce incorrect results
2921 There is a carry propagating bug in the Broadwell-specific Montgomery
2922 multiplication procedure that handles input lengths divisible by, but
2923 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2924 and DH private keys are impossible. This is because the subroutine in
2925 question is not used in operations with the private key itself and an input
2926 of the attacker's direct choice. Otherwise the bug can manifest itself as
2927 transient authentication and key negotiation failures or reproducible
2928 erroneous outcome of public-key operations with specially crafted input.
2929 Among EC algorithms only Brainpool P-512 curves are affected and one
2930 presumably can attack ECDH key negotiation. Impact was not analyzed in
2931 detail, because pre-requisites for attack are considered unlikely. Namely
2932 multiple clients have to choose the curve in question and the server has to
2933 share the private key among them, neither of which is default behaviour.
2934 Even then only clients that chose the curve will be affected.
2936 This issue was publicly reported as transient failures and was not
2937 initially recognized as a security issue. Thanks to Richard Morgan for
2938 providing reproducible case.
2943 * Removed automatic addition of RPATH in shared libraries and executables,
2944 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2948 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2950 * Fix Use After Free for large message sizes
2952 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2953 message larger than approx 16k is received then the underlying buffer to
2954 store the incoming message is reallocated and moved. Unfortunately a
2955 dangling pointer to the old location is left which results in an attempt to
2956 write to the previously freed location. This is likely to result in a
2957 crash, however it could potentially lead to execution of arbitrary code.
2959 This issue only affects OpenSSL 1.1.0a.
2961 This issue was reported to OpenSSL by Robert Święcki.
2966 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2968 * OCSP Status Request extension unbounded memory growth
2970 A malicious client can send an excessively large OCSP Status Request
2971 extension. If that client continually requests renegotiation, sending a
2972 large OCSP Status Request extension each time, then there will be unbounded
2973 memory growth on the server. This will eventually lead to a Denial Of
2974 Service attack through memory exhaustion. Servers with a default
2975 configuration are vulnerable even if they do not support OCSP. Builds using
2976 the "no-ocsp" build time option are not affected.
2978 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2983 * SSL_peek() hang on empty record
2985 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2986 sends an empty record. This could be exploited by a malicious peer in a
2987 Denial Of Service attack.
2989 This issue was reported to OpenSSL by Alex Gaynor.
2994 * Excessive allocation of memory in tls_get_message_header() and
2995 dtls1_preprocess_fragment()
2997 A (D)TLS message includes 3 bytes for its length in the header for the
2998 message. This would allow for messages up to 16Mb in length. Messages of
2999 this length are excessive and OpenSSL includes a check to ensure that a
3000 peer is sending reasonably sized messages in order to avoid too much memory
3001 being consumed to service a connection. A flaw in the logic of version
3002 1.1.0 means that memory for the message is allocated too early, prior to
3003 the excessive message length check. Due to way memory is allocated in
3004 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3005 to service a connection. This could lead to a Denial of Service through
3006 memory exhaustion. However, the excessive message length check still takes
3007 place, and this would cause the connection to immediately fail. Assuming
3008 that the application calls SSL_free() on the failed connection in a timely
3009 manner then the 21Mb of allocated memory will then be immediately freed
3010 again. Therefore the excessive memory allocation will be transitory in
3011 nature. This then means that there is only a security impact if:
3013 1) The application does not call SSL_free() in a timely manner in the event
3014 that the connection fails
3016 2) The application is working in a constrained environment where there is
3017 very little free memory
3019 3) The attacker initiates multiple connection attempts such that there are
3020 multiple connections in a state where memory has been allocated for the
3021 connection; SSL_free() has not yet been called; and there is insufficient
3022 memory to service the multiple requests.
3024 Except in the instance of (1) above any Denial Of Service is likely to be
3025 transitory because as soon as the connection fails the memory is
3026 subsequently freed again in the SSL_free() call. However there is an
3027 increased risk during this period of application crashes due to the lack of
3028 memory - which would then mean a more serious Denial of Service.
3030 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3031 (CVE-2016-6307 and CVE-2016-6308)
3035 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3036 had to be removed. Primary reason is that vendor assembler can't
3037 assemble our modules with -KPIC flag. As result it, assembly
3038 support, was not even available as option. But its lack means
3039 lack of side-channel resistant code, which is incompatible with
3040 security by todays standards. Fortunately gcc is readily available
3041 prepackaged option, which we firmly point at...
3045 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3047 * Windows command-line tool supports UTF-8 opt-in option for arguments
3048 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3049 (to any value) allows Windows user to access PKCS#12 file generated
3050 with Windows CryptoAPI and protected with non-ASCII password, as well
3051 as files generated under UTF-8 locale on Linux also protected with
3056 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3057 have been disabled by default and removed from DEFAULT, just like RC4.
3058 See the RC4 item below to re-enable both.
3062 * The method for finding the storage location for the Windows RAND seed file
3063 has changed. First we check %RANDFILE%. If that is not set then we check
3064 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3065 all else fails we fall back to C:\.
3069 * The EVP_EncryptUpdate() function has had its return type changed from void
3070 to int. A return of 0 indicates and error while a return of 1 indicates
3075 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3076 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3077 off the constant time implementation for RSA, DSA and DH have been made
3078 no-ops and deprecated.
3082 * Windows RAND implementation was simplified to only get entropy by
3083 calling CryptGenRandom(). Various other RAND-related tickets
3086 *Joseph Wylie Yandle, Rich Salz*
3088 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3089 and `OPENSSL_LH_`, respectively. The old names are available
3090 with API compatibility. They new names are now completely documented.
3094 * Unify TYPE_up_ref(obj) methods signature.
3095 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3096 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3097 int (instead of void) like all others TYPE_up_ref() methods.
3098 So now these methods also check the return value of CRYPTO_atomic_add(),
3099 and the validity of object reference counter.
3101 *fdasilvayy@gmail.com*
3103 * With Windows Visual Studio builds, the .pdb files are installed
3104 alongside the installed libraries and executables. For a static
3105 library installation, ossl_static.pdb is the associate compiler
3106 generated .pdb file to be used when linking programs.
3110 * Remove openssl.spec. Packaging files belong with the packagers.
3114 * Automatic Darwin/OSX configuration has had a refresh, it will now
3115 recognise x86_64 architectures automatically. You can still decide
3116 to build for a different bitness with the environment variable
3117 KERNEL_BITS (can be 32 or 64), for example:
3119 KERNEL_BITS=32 ./config
3123 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3124 256 bit AES and HMAC with SHA256.
3128 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3132 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3136 * To enable users to have their own config files and build file templates,
3137 Configure looks in the directory indicated by the environment variable
3138 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3139 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3140 name and is used as is.
3144 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3145 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3146 X509_CERT_FILE_CTX was removed.
3150 * "shared" builds are now the default. To create only static libraries use
3151 the "no-shared" Configure option.
3155 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3156 All of these option have not worked for some while and are fundamental
3161 * Make various cleanup routines no-ops and mark them as deprecated. Most
3162 global cleanup functions are no longer required because they are handled
3163 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3164 Explicitly de-initing can cause problems (e.g. where a library that uses
3165 OpenSSL de-inits, but an application is still using it). The affected
3166 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3167 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3168 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3169 COMP_zlib_cleanup().
3173 * --strict-warnings no longer enables runtime debugging options
3174 such as REF_DEBUG. Instead, debug options are automatically
3175 enabled with '--debug' builds.
3177 *Andy Polyakov, Emilia Käsper*
3179 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3180 have been moved out of the public header files. New functions for managing
3181 these have been added.
3185 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3186 objects have been moved out of the public header files. New
3187 functions for managing these have been added.
3191 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3192 have been moved out of the public header files. New functions for managing
3193 these have been added.
3197 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3198 moved out of the public header files. New functions for managing these
3203 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3207 * Removed the mk1mf build scripts.
3211 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3212 it is always safe to #include a header now.
3216 * Removed the aged BC-32 config and all its supporting scripts
3220 * Removed support for Ultrix, Netware, and OS/2.
3224 * Add support for HKDF.
3226 *Alessandro Ghedini*
3228 * Add support for blake2b and blake2s
3232 * Added support for "pipelining". Ciphers that have the
3233 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3234 encryptions/decryptions simultaneously. There are currently no built-in
3235 ciphers with this property but the expectation is that engines will be able
3236 to offer it to significantly improve throughput. Support has been extended
3237 into libssl so that multiple records for a single connection can be
3238 processed in one go (for >=TLS 1.1).
3242 * Added the AFALG engine. This is an async capable engine which is able to
3243 offload work to the Linux kernel. In this initial version it only supports
3244 AES128-CBC. The kernel must be version 4.1.0 or greater.
3248 * OpenSSL now uses a new threading API. It is no longer necessary to
3249 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3250 are two supported threading models: pthreads and windows threads. It is
3251 also possible to configure OpenSSL at compile time for "no-threads". The
3252 old threading API should no longer be used. The functions have been
3253 replaced with "no-op" compatibility macros.
3255 *Alessandro Ghedini, Matt Caswell*
3257 * Modify behavior of ALPN to invoke callback after SNI/servername
3258 callback, such that updates to the SSL_CTX affect ALPN.
3262 * Add SSL_CIPHER queries for authentication and key-exchange.
3266 * Changes to the DEFAULT cipherlist:
3267 - Prefer (EC)DHE handshakes over plain RSA.
3268 - Prefer AEAD ciphers over legacy ciphers.
3269 - Prefer ECDSA over RSA when both certificates are available.
3270 - Prefer TLSv1.2 ciphers/PRF.
3271 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3276 * Change the ECC default curve list to be this, in order: x25519,
3277 secp256r1, secp521r1, secp384r1.
3281 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3282 disabled by default. They can be re-enabled using the
3283 enable-weak-ssl-ciphers option to Configure.
3287 * If the server has ALPN configured, but supports no protocols that the
3288 client advertises, send a fatal "no_application_protocol" alert.
3289 This behaviour is SHALL in RFC 7301, though it isn't universally
3290 implemented by other servers.
3294 * Add X25519 support.
3295 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3296 for public and private key encoding using the format documented in
3297 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3298 key generation and key derivation.
3300 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3305 * Deprecate SRP_VBASE_get_by_user.
3306 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3307 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3308 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3309 seed, even if the seed is configured.
3311 Users should use SRP_VBASE_get1_by_user instead. Note that in
3312 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3313 also that even though configuring the SRP seed attempts to hide
3314 invalid usernames by continuing the handshake with fake
3315 credentials, this behaviour is not constant time and no strong
3316 guarantees are made that the handshake is indistinguishable from
3317 that of a valid user.
3321 * Configuration change; it's now possible to build dynamic engines
3322 without having to build shared libraries and vice versa. This
3323 only applies to the engines in `engines/`, those in `crypto/engine/`
3324 will always be built into libcrypto (i.e. "static").
3326 Building dynamic engines is enabled by default; to disable, use
3327 the configuration option "disable-dynamic-engine".
3329 The only requirements for building dynamic engines are the
3330 presence of the DSO module and building with position independent
3331 code, so they will also automatically be disabled if configuring
3332 with "disable-dso" or "disable-pic".
3334 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3335 are also taken away from openssl/opensslconf.h, as they are
3340 * Configuration change; if there is a known flag to compile
3341 position independent code, it will always be applied on the
3342 libcrypto and libssl object files, and never on the application
3343 object files. This means other libraries that use routines from
3344 libcrypto / libssl can be made into shared libraries regardless
3345 of how OpenSSL was configured.
3347 If this isn't desirable, the configuration options "disable-pic"
3348 or "no-pic" can be used to disable the use of PIC. This will
3349 also disable building shared libraries and dynamic engines.
3353 * Removed JPAKE code. It was experimental and has no wide use.
3357 * The INSTALL_PREFIX Makefile variable has been renamed to
3358 DESTDIR. That makes for less confusion on what this variable
3359 is for. Also, the configuration option --install_prefix is
3364 * Heartbeat for TLS has been removed and is disabled by default
3365 for DTLS; configure with enable-heartbeats. Code that uses the
3366 old #define's might need to be updated.
3368 *Emilia Käsper, Rich Salz*
3370 * Rename REF_CHECK to REF_DEBUG.
3374 * New "unified" build system
3376 The "unified" build system is aimed to be a common system for all
3377 platforms we support. With it comes new support for VMS.
3379 This system builds supports building in a different directory tree
3380 than the source tree. It produces one Makefile (for unix family
3381 or lookalikes), or one descrip.mms (for VMS).
3383 The source of information to make the Makefile / descrip.mms is
3384 small files called 'build.info', holding the necessary
3385 information for each directory with source to compile, and a
3386 template in Configurations, like unix-Makefile.tmpl or
3389 With this change, the library names were also renamed on Windows
3390 and on VMS. They now have names that are closer to the standard
3391 on Unix, and include the major version number, and in certain
3392 cases, the architecture they are built for. See "Notes on shared
3393 libraries" in INSTALL.
3395 We rely heavily on the perl module Text::Template.
3399 * Added support for auto-initialisation and de-initialisation of the library.
3400 OpenSSL no longer requires explicit init or deinit routines to be called,
3401 except in certain circumstances. See the OPENSSL_init_crypto() and
3402 OPENSSL_init_ssl() man pages for further information.
3406 * The arguments to the DTLSv1_listen function have changed. Specifically the
3407 "peer" argument is now expected to be a BIO_ADDR object.
3409 * Rewrite of BIO networking library. The BIO library lacked consistent
3410 support of IPv6, and adding it required some more extensive
3411 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3412 which hold all types of addresses and chains of address information.
3413 It also introduces a new API, with functions like BIO_socket,
3414 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3415 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3416 have been adapted accordingly.
3420 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3425 * CRIME protection: disable compression by default, even if OpenSSL is
3426 compiled with zlib enabled. Applications can still enable compression
3427 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3428 using the SSL_CONF library to configure compression.
3432 * The signature of the session callback configured with
3433 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3434 was explicitly marked as `const unsigned char*` instead of
3439 * Always DPURIFY. Remove the use of uninitialized memory in the
3440 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3444 * Removed many obsolete configuration items, including
3445 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3446 MD2_CHAR, MD2_INT, MD2_LONG
3448 IDEA_SHORT, IDEA_LONG
3449 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3451 *Rich Salz, with advice from Andy Polyakov*
3453 * Many BN internals have been moved to an internal header file.
3455 *Rich Salz with help from Andy Polyakov*
3457 * Configuration and writing out the results from it has changed.
3458 Files such as Makefile include/openssl/opensslconf.h and are now
3459 produced through general templates, such as Makefile.in and
3460 crypto/opensslconf.h.in and some help from the perl module
3463 Also, the center of configuration information is no longer
3464 Makefile. Instead, Configure produces a perl module in
3465 configdata.pm which holds most of the config data (in the hash
3466 table %config), the target data that comes from the target
3467 configuration in one of the `Configurations/*.conf` files (in
3472 * To clarify their intended purposes, the Configure options
3473 --prefix and --openssldir change their semantics, and become more
3474 straightforward and less interdependent.
3476 --prefix shall be used exclusively to give the location INSTALLTOP
3477 where programs, scripts, libraries, include files and manuals are
3478 going to be installed. The default is now /usr/local.
3480 --openssldir shall be used exclusively to give the default
3481 location OPENSSLDIR where certificates, private keys, CRLs are
3482 managed. This is also where the default openssl.cnf gets
3484 If the directory given with this option is a relative path, the
3485 values of both the --prefix value and the --openssldir value will
3486 be combined to become OPENSSLDIR.
3487 The default for --openssldir is INSTALLTOP/ssl.
3489 Anyone who uses --openssldir to specify where OpenSSL is to be
3490 installed MUST change to use --prefix instead.
3494 * The GOST engine was out of date and therefore it has been removed. An up
3495 to date GOST engine is now being maintained in an external repository.
3496 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3497 support for GOST ciphersuites (these are only activated if a GOST engine
3502 * EGD is no longer supported by default; use enable-egd when
3505 *Ben Kaduk and Rich Salz*
3507 * The distribution now has Makefile.in files, which are used to
3508 create Makefile's when Configure is run. *Configure must be run
3509 before trying to build now.*
3513 * The return value for SSL_CIPHER_description() for error conditions
3518 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3520 Obtaining and performing DNSSEC validation of TLSA records is
3521 the application's responsibility. The application provides
3522 the TLSA records of its choice to OpenSSL, and these are then
3523 used to authenticate the peer.
3525 The TLSA records need not even come from DNS. They can, for
3526 example, be used to implement local end-entity certificate or
3527 trust-anchor "pinning", where the "pin" data takes the form
3528 of TLSA records, which can augment or replace verification
3529 based on the usual WebPKI public certification authorities.
3533 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3534 continues to support deprecated interfaces in default builds.
3535 However, applications are strongly advised to compile their
3536 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3537 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3538 or the 1.1.0 releases.
3540 In environments in which all applications have been ported to
3541 not use any deprecated interfaces OpenSSL's Configure script
3542 should be used with the --api=1.1.0 option to entirely remove
3543 support for the deprecated features from the library and
3544 unconditionally disable them in the installed headers.
3545 Essentially the same effect can be achieved with the "no-deprecated"
3546 argument to Configure, except that this will always restrict
3547 the build to just the latest API, rather than a fixed API
3550 As applications are ported to future revisions of the API,
3551 they should update their compile-time OPENSSL_API_COMPAT define
3552 accordingly, but in most cases should be able to continue to
3553 compile with later releases.
3555 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3556 0x10000000L and 0x00908000L, respectively. However those
3557 versions did not support the OPENSSL_API_COMPAT feature, and
3558 so applications are not typically tested for explicit support
3559 of just the undeprecated features of either release.
3563 * Add support for setting the minimum and maximum supported protocol.
3564 It can bet set via the SSL_set_min_proto_version() and
3565 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3566 MaxProtocol. It's recommended to use the new APIs to disable
3567 protocols instead of disabling individual protocols using
3568 SSL_set_options() or SSL_CONF's Protocol. This change also
3569 removes support for disabling TLS 1.2 in the OpenSSL TLS
3570 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3574 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3578 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3579 and integrates ECDSA and ECDH functionality into EC. Implementations can
3580 now redirect key generation and no longer need to convert to or from
3583 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3584 include the ec.h header file instead.
3588 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3589 ciphers who are no longer supported and drops support the ephemeral RSA key
3590 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3594 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3595 opaque. For HMAC_CTX, the following constructors and destructors
3598 HMAC_CTX *HMAC_CTX_new(void);
3599 void HMAC_CTX_free(HMAC_CTX *ctx);
3601 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3602 destroy such methods has been added. See EVP_MD_meth_new(3) and
3603 EVP_CIPHER_meth_new(3) for documentation.
3606 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3607 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3608 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3609 an already created structure.
3610 2) For consistency with the majority of our object creators and
3611 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3612 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3613 for deprecated builds.
3617 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3618 cryptographic operations to be performed asynchronously as long as an
3619 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3620 further details. Libssl has also had this capability integrated with the
3621 introduction of the new mode SSL_MODE_ASYNC and associated error
3622 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3623 pages. This work was developed in partnership with Intel Corp.
3627 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3628 always enabled now. If you want to disable the support you should
3629 exclude it using the list of supported ciphers. This also means that the
3630 "-no_ecdhe" option has been removed from s_server.
3634 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3635 SSL_{CTX_}set1_curves() which can set a list.
3639 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3640 curve you want to support using SSL_{CTX_}set1_curves().
3644 * State machine rewrite. The state machine code has been significantly
3645 refactored in order to remove much duplication of code and solve issues
3646 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3647 further details). This change does have some associated API changes.
3648 Notably the SSL_state() function has been removed and replaced by
3649 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3650 SSL_set_state() has been removed altogether. The previous handshake states
3651 defined in ssl.h and ssl3.h have also been removed.
3655 * All instances of the string "ssleay" in the public API were replaced
3656 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3657 Some error codes related to internal RSA_eay API's were renamed.
3661 * The demo files in crypto/threads were moved to demo/threads.
3665 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3668 *Matt Caswell, Rich Salz*
3670 * New ASN.1 embed macro.
3672 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3673 structure is not allocated: it is part of the parent. That is instead of
3681 This reduces memory fragmentation and make it impossible to accidentally
3682 set a mandatory field to NULL.
3684 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3685 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3686 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3691 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3695 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3696 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3697 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3698 DES and RC4 ciphersuites.
3702 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3703 This changes the decoding behaviour for some invalid messages,
3704 though the change is mostly in the more lenient direction, and
3705 legacy behaviour is preserved as much as possible.
3709 * Fix no-stdio build.
3710 *David Woodhouse <David.Woodhouse@intel.com> and also*
3711 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3713 * New testing framework
3714 The testing framework has been largely rewritten and is now using
3715 perl and the perl modules Test::Harness and an extended variant of
3716 Test::More called OpenSSL::Test to do its work. All test scripts in
3717 test/ have been rewritten into test recipes, and all direct calls to
3718 executables in test/Makefile have become individual recipes using the
3719 simplified testing OpenSSL::Test::Simple.
3721 For documentation on our testing modules, do:
3723 perldoc test/testlib/OpenSSL/Test/Simple.pm
3724 perldoc test/testlib/OpenSSL/Test.pm
3728 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3729 are used; the latter aborts on memory leaks (usually checked on exit).
3730 Some undocumented "set malloc, etc., hooks" functions were removed
3731 and others were changed. All are now documented.
3735 * In DSA_generate_parameters_ex, if the provided seed is too short,
3738 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3740 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3741 from RFC4279, RFC4785, RFC5487, RFC5489.
3743 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3744 original RSA_PSK patch.
3748 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3749 era flag was never set throughout the codebase (only read). Also removed
3750 SSL3_FLAGS_POP_BUFFER which was only used if
3751 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3755 * Changed the default name options in the "ca", "crl", "req" and "x509"
3756 to be "oneline" instead of "compat".
3760 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3761 not aware of clients that still exhibit this bug, and the workaround
3762 hasn't been working properly for a while.
3766 * The return type of BIO_number_read() and BIO_number_written() as well as
3767 the corresponding num_read and num_write members in the BIO structure has
3768 changed from unsigned long to uint64_t. On platforms where an unsigned
3769 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3774 * Given the pervasive nature of TLS extensions it is inadvisable to run
3775 OpenSSL without support for them. It also means that maintaining
3776 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3777 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3781 * Removed support for the two export grade static DH ciphersuites
3782 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3783 were newly added (along with a number of other static DH ciphersuites) to
3784 1.0.2. However the two export ones have *never* worked since they were
3785 introduced. It seems strange in any case to be adding new export
3786 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3790 * Version negotiation has been rewritten. In particular SSLv23_method(),
3791 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3792 and turned into macros which simply call the new preferred function names
3793 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3794 should use the new names instead. Also as part of this change the ssl23.h
3795 header file has been removed.
3799 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3800 code and the associated standard is no longer considered fit-for-purpose.
3804 * RT2547 was closed. When generating a private key, try to make the
3805 output file readable only by the owner. This behavior change might
3806 be noticeable when interacting with other software.
3808 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3813 * Added HTTP GET support to the ocsp command.
3817 * Changed default digest for the dgst and enc commands from MD5 to
3822 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3826 * Added support for TLS extended master secret from
3827 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3828 initial patch which was a great help during development.
3832 * All libssl internal structures have been removed from the public header
3833 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3834 now redundant). Users should not attempt to access internal structures
3835 directly. Instead they should use the provided API functions.
3839 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3840 Access to deprecated functions can be re-enabled by running config with
3841 "enable-deprecated". In addition applications wishing to use deprecated
3842 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3843 will, by default, disable some transitive includes that previously existed
3844 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3848 * Added support for OCB mode. OpenSSL has been granted a patent license
3849 compatible with the OpenSSL license for use of OCB. Details are available
3850 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3851 for OCB can be removed by calling config with no-ocb.
3855 * SSLv2 support has been removed. It still supports receiving a SSLv2
3856 compatible client hello.
3860 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3861 done while fixing the error code for the key-too-small case.
3863 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3865 * CA.sh has been removed; use CA.pl instead.
3869 * Removed old DES API.
3873 * Remove various unsupported platforms:
3879 Sinix/ReliantUNIX RM400
3884 16-bit platforms such as WIN16
3888 * Clean up OPENSSL_NO_xxx #define's
3889 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3890 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3891 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3892 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3893 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3894 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3895 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3896 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3897 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3898 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3902 * Cleaned up dead code
3903 Remove all but one '#ifdef undef' which is to be looked at.
3907 * Clean up calling of xxx_free routines.
3908 Just like free(), fix most of the xxx_free routines to accept
3909 NULL. Remove the non-null checks from callers. Save much code.
3913 * Add secure heap for storage of private keys (when possible).
3914 Add BIO_s_secmem(), CBIGNUM, etc.
3915 Contributed by Akamai Technologies under our Corporate CLA.
3919 * Experimental support for a new, fast, unbiased prime candidate generator,
3920 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3922 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3924 * New output format NSS in the sess_id command line tool. This allows
3925 exporting the session id and the master key in NSS keylog format.
3927 *Martin Kaiser <martin@kaiser.cx>*
3929 * Harmonize version and its documentation. -f flag is used to display
3932 *mancha <mancha1@zoho.com>*
3934 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3935 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3937 *mancha <mancha1@zoho.com>*
3939 * Fix some double frees. These are not thought to be exploitable.
3941 *mancha <mancha1@zoho.com>*
3943 * A missing bounds check in the handling of the TLS heartbeat extension
3944 can be used to reveal up to 64k of memory to a connected client or
3947 Thanks for Neel Mehta of Google Security for discovering this bug and to
3948 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3949 preparing the fix ([CVE-2014-0160])
3951 *Adam Langley, Bodo Moeller*
3953 * Fix for the attack described in the paper "Recovering OpenSSL
3954 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3955 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3956 <http://eprint.iacr.org/2014/140>
3958 Thanks to Yuval Yarom and Naomi Benger for discovering this
3959 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3961 *Yuval Yarom and Naomi Benger*
3963 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3964 this fixes a limitation in previous versions of OpenSSL.
3968 * Experimental encrypt-then-mac support.
3970 Experimental support for encrypt then mac from
3971 draft-gutmann-tls-encrypt-then-mac-02.txt
3973 To enable it set the appropriate extension number (0x42 for the test
3974 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3976 For non-compliant peers (i.e. just about everything) this should have no
3979 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3983 * Add EVP support for key wrapping algorithms, to avoid problems with
3984 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3985 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3986 algorithms and include tests cases.
3990 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3995 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3996 MGF1 digest and OAEP label.
4000 * Make openssl verify return errors.
4002 *Chris Palmer <palmer@google.com> and Ben Laurie*
4004 * New function ASN1_TIME_diff to calculate the difference between two
4005 ASN1_TIME structures or one structure and the current time.
4009 * Update fips_test_suite to support multiple command line options. New
4010 test to induce all self test errors in sequence and check expected
4015 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4016 sign or verify all in one operation.
4020 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4021 test programs and fips_test_suite. Includes functionality to parse
4022 the minimal script output of fipsalgest.pl directly.
4026 * Add authorisation parameter to FIPS_module_mode_set().
4030 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4034 * Use separate DRBG fields for internal and external flags. New function
4035 FIPS_drbg_health_check() to perform on demand health checking. Add
4036 generation tests to fips_test_suite with reduced health check interval to
4037 demonstrate periodic health checking. Add "nodh" option to
4038 fips_test_suite to skip very slow DH test.
4042 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4047 * More extensive health check for DRBG checking many more failure modes.
4048 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4049 combination: call this in fips_test_suite.
4053 * Add support for canonical generation of DSA parameter 'g'. See
4056 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4057 POST to handle HMAC cases.
4061 * Add functions FIPS_module_version() and FIPS_module_version_text()
4062 to return numerical and string versions of the FIPS module number.
4066 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4067 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4068 outside the validated module in the FIPS capable OpenSSL.
4072 * Minor change to DRBG entropy callback semantics. In some cases
4073 there is no multiple of the block length between min_len and
4074 max_len. Allow the callback to return more than max_len bytes
4075 of entropy but discard any extra: it is the callback's responsibility
4076 to ensure that the extra data discarded does not impact the
4077 requested amount of entropy.
4081 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4082 information in FIPS186-3, SP800-57 and SP800-131A.
4086 * CCM support via EVP. Interface is very similar to GCM case except we
4087 must supply all data in one chunk (i.e. no update, final) and the
4088 message length must be supplied if AAD is used. Add algorithm test
4093 * Initial version of POST overhaul. Add POST callback to allow the status
4094 of POST to be monitored and/or failures induced. Modify fips_test_suite
4095 to use callback. Always run all selftests even if one fails.
4099 * XTS support including algorithm test driver in the fips_gcmtest program.
4100 Note: this does increase the maximum key length from 32 to 64 bytes but
4101 there should be no binary compatibility issues as existing applications
4102 will never use XTS mode.
4106 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4107 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4108 performs algorithm blocking for unapproved PRNG types. Also do not
4109 set PRNG type in FIPS_mode_set(): leave this to the application.
4110 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4111 the standard OpenSSL PRNG: set additional data to a date time vector.
4115 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4116 This shouldn't present any incompatibility problems because applications
4117 shouldn't be using these directly and any that are will need to rethink
4118 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4122 * Extensive self tests and health checking required by SP800-90 DRBG.
4123 Remove strength parameter from FIPS_drbg_instantiate and always
4124 instantiate at maximum supported strength.
4128 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4132 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4136 * New function DH_compute_key_padded() to compute a DH key and pad with
4137 leading zeroes if needed: this complies with SP800-56A et al.
4141 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4142 anything, incomplete, subject to change and largely untested at present.
4146 * Modify fipscanisteronly build option to only build the necessary object
4147 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4151 * Add experimental option FIPSSYMS to give all symbols in
4152 fipscanister.o and FIPS or fips prefix. This will avoid
4153 conflicts with future versions of OpenSSL. Add perl script
4154 util/fipsas.pl to preprocess assembly language source files
4155 and rename any affected symbols.
4159 * Add selftest checks and algorithm block of non-fips algorithms in
4160 FIPS mode. Remove DES2 from selftests.
4164 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4165 return internal method without any ENGINE dependencies. Add new
4166 tiny fips sign and verify functions.
4170 * New build option no-ec2m to disable characteristic 2 code.
4174 * New build option "fipscanisteronly". This only builds fipscanister.o
4175 and (currently) associated fips utilities. Uses the file Makefile.fips
4176 instead of Makefile.org as the prototype.
4180 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4181 Update fips_gcmtest to use IV generator.
4185 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4186 setting output buffer to NULL. The `*Final` function must be
4187 called although it will not retrieve any additional data. The tag
4188 can be set or retrieved with a ctrl. The IV length is by default 12
4189 bytes (96 bits) but can be set to an alternative value. If the IV
4190 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4195 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4196 underlying do_cipher function handles all cipher semantics itself
4197 including padding and finalisation. This is useful if (for example)
4198 an ENGINE cipher handles block padding itself. The behaviour of
4199 do_cipher is subtly changed if this flag is set: the return value
4200 is the number of characters written to the output buffer (zero is
4201 no longer an error code) or a negative error code. Also if the
4202 input buffer is NULL and length 0 finalisation should be performed.
4206 * If a candidate issuer certificate is already part of the constructed
4207 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4211 * Improve forward-security support: add functions
4213 void SSL_CTX_set_not_resumable_session_callback(
4214 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4215 void SSL_set_not_resumable_session_callback(
4216 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4218 for use by SSL/TLS servers; the callback function will be called whenever a
4219 new session is created, and gets to decide whether the session may be
4220 cached to make it resumable (return 0) or not (return 1). (As by the
4221 SSL/TLS protocol specifications, the session_id sent by the server will be
4222 empty to indicate that the session is not resumable; also, the server will
4223 not generate RFC 4507 (RFC 5077) session tickets.)
4225 A simple reasonable callback implementation is to return is_forward_secure.
4226 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4227 by the SSL/TLS server library, indicating whether it can provide forward
4230 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4232 * New -verify_name option in command line utilities to set verification
4237 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4238 Add CMAC pkey methods.
4242 * Experimental renegotiation in s_server -www mode. If the client
4243 browses /reneg connection is renegotiated. If /renegcert it is
4244 renegotiated requesting a certificate.
4248 * Add an "external" session cache for debugging purposes to s_server. This
4249 should help trace issues which normally are only apparent in deployed
4250 multi-process servers.
4254 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4255 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4256 BIO_set_cipher() and some obscure PEM functions were changed so they
4257 can now return an error. The RAND changes required a change to the
4258 RAND_METHOD structure.
4262 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4263 a gcc attribute to warn if the result of a function is ignored. This
4264 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4265 whose return value is often ignored.
4269 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4270 These allow SCTs (signed certificate timestamps) to be requested and
4271 validated when establishing a connection.
4273 *Rob Percival <robpercival@google.com>*
4278 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4280 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4281 used even when parsing explicit parameters, when loading a encoded key
4282 or calling `EC_GROUP_new_from_ecpkparameters()`/
4283 `EC_GROUP_new_from_ecparameters()`.
4284 This prevents bypass of security hardening and performance gains,
4285 especially for curves with specialized EC_METHODs.
4286 By default, if a key encoded with explicit parameters is loaded and later
4287 encoded, the output is still encoded with explicit parameters, even if
4288 internally a "named" EC_GROUP is used for computation.
4292 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4293 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4294 NULL. After this change, only the cofactor parameter can be NULL. It also
4295 does some minimal sanity checks on the passed order.
4300 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4301 An attack is simple, if the first CMS_recipientInfo is valid but the
4302 second CMS_recipientInfo is chosen ciphertext. If the second
4303 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4304 encryption key will be replaced by garbage, and the message cannot be
4305 decoded, but if the RSA decryption fails, the correct encryption key is
4306 used and the recipient will not notice the attack.
4307 As a work around for this potential attack the length of the decrypted
4308 key must be equal to the cipher default key length, in case the
4309 certifiate is not given and all recipientInfo are tried out.
4310 The old behaviour can be re-enabled in the CMS code by setting the
4311 CMS_DEBUG_DECRYPT flag.
4316 * Document issue with installation paths in diverse Windows builds
4318 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4319 binaries and run-time config file.
4324 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4326 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4327 This changes the size when using the `genpkey` command when no size is given.
4328 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4329 generation commands to use 2048 bits by default.
4333 * Add FIPS support for Android Arm 64-bit
4335 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4336 Module in Version 2.0.10. For some reason, the corresponding target
4337 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4338 built with FIPS support on Android Arm 64-bit. This omission has been
4341 *Matthias St. Pierre*
4343 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4345 * 0-byte record padding oracle
4347 If an application encounters a fatal protocol error and then calls
4348 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4349 then OpenSSL can respond differently to the calling application if a 0 byte
4350 record is received with invalid padding compared to if a 0 byte record is
4351 received with an invalid MAC. If the application then behaves differently
4352 based on that in a way that is detectable to the remote peer, then this
4353 amounts to a padding oracle that could be used to decrypt data.
4355 In order for this to be exploitable "non-stitched" ciphersuites must be in
4356 use. Stitched ciphersuites are optimised implementations of certain
4357 commonly used ciphersuites. Also the application must call SSL_shutdown()
4358 twice even if a protocol error has occurred (applications should not do
4359 this but some do anyway).
4361 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4362 Aviram, with additional investigation by Steven Collison and Andrew
4363 Hourselt. It was reported to OpenSSL on 10th December 2018.
4368 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4372 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4374 * Microarchitecture timing vulnerability in ECC scalar multiplication
4376 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4377 shown to be vulnerable to a microarchitecture timing side channel attack.
4378 An attacker with sufficient access to mount local timing attacks during
4379 ECDSA signature generation could recover the private key.
4381 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4382 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4388 * Timing vulnerability in DSA signature generation
4390 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4391 timing side channel attack. An attacker could use variations in the signing
4392 algorithm to recover the private key.
4394 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4399 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4400 Module, accidentally introduced while backporting security fixes from the
4401 development branch and hindering the use of ECC in FIPS mode.
4405 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4407 * Client DoS due to large DH parameter
4409 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4410 malicious server can send a very large prime value to the client. This will
4411 cause the client to spend an unreasonably long period of time generating a
4412 key for this prime resulting in a hang until the client has finished. This
4413 could be exploited in a Denial Of Service attack.
4415 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4420 * Cache timing vulnerability in RSA Key Generation
4422 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4423 a cache timing side channel attack. An attacker with sufficient access to
4424 mount cache timing attacks during the RSA key generation process could
4425 recover the private key.
4427 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4428 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4433 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4434 parameter is no longer accepted, as it leads to a corrupt table. NULL
4435 pem_str is reserved for alias entries only.
4439 * Revert blinding in ECDSA sign and instead make problematic addition
4440 length-invariant. Switch even to fixed-length Montgomery multiplication.
4444 * Change generating and checking of primes so that the error rate of not
4445 being prime depends on the intended use based on the size of the input.
4446 For larger primes this will result in more rounds of Miller-Rabin.
4447 The maximal error rate for primes with more than 1080 bits is lowered
4450 *Kurt Roeckx, Annie Yousar*
4452 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4456 * Add blinding to ECDSA and DSA signatures to protect against side channel
4457 attacks discovered by Keegan Ryan (NCC Group).
4461 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4462 now allow empty (zero character) pass phrases.
4466 * Certificate time validation (X509_cmp_time) enforces stricter
4467 compliance with RFC 5280. Fractional seconds and timezone offsets
4468 are no longer allowed.
4472 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4474 * Constructed ASN.1 types with a recursive definition could exceed the stack
4476 Constructed ASN.1 types with a recursive definition (such as can be found
4477 in PKCS7) could eventually exceed the stack given malicious input with
4478 excessive recursion. This could result in a Denial Of Service attack. There
4479 are no such structures used within SSL/TLS that come from untrusted sources
4480 so this is considered safe.
4482 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4488 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4490 * Read/write after SSL object in error state
4492 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4493 mechanism. The intent was that if a fatal error occurred during a handshake
4494 then OpenSSL would move into the error state and would immediately fail if
4495 you attempted to continue the handshake. This works as designed for the
4496 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4497 SSL_connect()), however due to a bug it does not work correctly if
4498 SSL_read() or SSL_write() is called directly. In that scenario, if the
4499 handshake fails then a fatal error will be returned in the initial function
4500 call. If SSL_read()/SSL_write() is subsequently called by the application
4501 for the same SSL object then it will succeed and the data is passed without
4502 being decrypted/encrypted directly from the SSL/TLS record layer.
4504 In order to exploit this issue an application bug would have to be present
4505 that resulted in a call to SSL_read()/SSL_write() being issued after having
4506 already received a fatal error.
4508 This issue was reported to OpenSSL by David Benjamin (Google).
4513 * rsaz_1024_mul_avx2 overflow bug on x86_64
4515 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4516 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4517 Analysis suggests that attacks against RSA and DSA as a result of this
4518 defect would be very difficult to perform and are not believed likely.
4519 Attacks against DH1024 are considered just feasible, because most of the
4520 work necessary to deduce information about a private key may be performed
4521 offline. The amount of resources required for such an attack would be
4522 significant. However, for an attack on TLS to be meaningful, the server
4523 would have to share the DH1024 private key among multiple clients, which is
4524 no longer an option since CVE-2016-0701.
4526 This only affects processors that support the AVX2 but not ADX extensions
4527 like Intel Haswell (4th generation).
4529 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4530 was originally found via the OSS-Fuzz project.
4535 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4537 * bn_sqrx8x_internal carry bug on x86_64
4539 There is a carry propagating bug in the x86_64 Montgomery squaring
4540 procedure. No EC algorithms are affected. Analysis suggests that attacks
4541 against RSA and DSA as a result of this defect would be very difficult to
4542 perform and are not believed likely. Attacks against DH are considered just
4543 feasible (although very difficult) because most of the work necessary to
4544 deduce information about a private key may be performed offline. The amount
4545 of resources required for such an attack would be very significant and
4546 likely only accessible to a limited number of attackers. An attacker would
4547 additionally need online access to an unpatched system using the target
4548 private key in a scenario with persistent DH parameters and a private
4549 key that is shared between multiple clients.
4551 This only affects processors that support the BMI1, BMI2 and ADX extensions
4552 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4554 This issue was reported to OpenSSL by the OSS-Fuzz project.
4559 * Malformed X.509 IPAddressFamily could cause OOB read
4561 If an X.509 certificate has a malformed IPAddressFamily extension,
4562 OpenSSL could do a one-byte buffer overread. The most likely result
4563 would be an erroneous display of the certificate in text format.
4565 This issue was reported to OpenSSL by the OSS-Fuzz project.
4570 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4572 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4573 platform rather than 'mingw'.
4577 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4579 * Truncated packet could crash via OOB read
4581 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4582 cipher is being used, then a truncated packet can cause that host to
4583 perform an out-of-bounds read, usually resulting in a crash.
4585 This issue was reported to OpenSSL by Robert Święcki of Google.
4590 * BN_mod_exp may produce incorrect results on x86_64
4592 There is a carry propagating bug in the x86_64 Montgomery squaring
4593 procedure. No EC algorithms are affected. Analysis suggests that attacks
4594 against RSA and DSA as a result of this defect would be very difficult to
4595 perform and are not believed likely. Attacks against DH are considered just
4596 feasible (although very difficult) because most of the work necessary to
4597 deduce information about a private key may be performed offline. The amount
4598 of resources required for such an attack would be very significant and
4599 likely only accessible to a limited number of attackers. An attacker would
4600 additionally need online access to an unpatched system using the target
4601 private key in a scenario with persistent DH parameters and a private
4602 key that is shared between multiple clients. For example this can occur by
4603 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4604 similar to CVE-2015-3193 but must be treated as a separate problem.
4606 This issue was reported to OpenSSL by the OSS-Fuzz project.
4611 * Montgomery multiplication may produce incorrect results
4613 There is a carry propagating bug in the Broadwell-specific Montgomery
4614 multiplication procedure that handles input lengths divisible by, but
4615 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4616 and DH private keys are impossible. This is because the subroutine in
4617 question is not used in operations with the private key itself and an input
4618 of the attacker's direct choice. Otherwise the bug can manifest itself as
4619 transient authentication and key negotiation failures or reproducible
4620 erroneous outcome of public-key operations with specially crafted input.
4621 Among EC algorithms only Brainpool P-512 curves are affected and one
4622 presumably can attack ECDH key negotiation. Impact was not analyzed in
4623 detail, because pre-requisites for attack are considered unlikely. Namely
4624 multiple clients have to choose the curve in question and the server has to
4625 share the private key among them, neither of which is default behaviour.
4626 Even then only clients that chose the curve will be affected.
4628 This issue was publicly reported as transient failures and was not
4629 initially recognized as a security issue. Thanks to Richard Morgan for
4630 providing reproducible case.
4635 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4636 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4637 prevent issues where no progress is being made and the peer continually
4638 sends unrecognised record types, using up resources processing them.
4642 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4644 * Missing CRL sanity check
4646 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4647 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4648 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4650 This issue only affects the OpenSSL 1.0.2i
4655 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4657 * OCSP Status Request extension unbounded memory growth
4659 A malicious client can send an excessively large OCSP Status Request
4660 extension. If that client continually requests renegotiation, sending a
4661 large OCSP Status Request extension each time, then there will be unbounded
4662 memory growth on the server. This will eventually lead to a Denial Of
4663 Service attack through memory exhaustion. Servers with a default
4664 configuration are vulnerable even if they do not support OCSP. Builds using
4665 the "no-ocsp" build time option are not affected.
4667 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4672 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4675 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4681 * OOB write in MDC2_Update()
4683 An overflow can occur in MDC2_Update() either if called directly or
4684 through the EVP_DigestUpdate() function using MDC2. If an attacker
4685 is able to supply very large amounts of input data after a previous
4686 call to EVP_EncryptUpdate() with a partial block then a length check
4687 can overflow resulting in a heap corruption.
4689 The amount of data needed is comparable to SIZE_MAX which is impractical
4692 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4697 * Malformed SHA512 ticket DoS
4699 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4700 DoS attack where a malformed ticket will result in an OOB read which will
4703 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4704 a custom server callback and ticket lookup mechanism.
4706 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4711 * OOB write in BN_bn2dec()
4713 The function BN_bn2dec() does not check the return value of BN_div_word().
4714 This can cause an OOB write if an application uses this function with an
4715 overly large BIGNUM. This could be a problem if an overly large certificate
4716 or CRL is printed out from an untrusted source. TLS is not affected because
4717 record limits will reject an oversized certificate before it is parsed.
4719 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4724 * OOB read in TS_OBJ_print_bio()
4726 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4727 the total length the OID text representation would use and not the amount
4728 of data written. This will result in OOB reads when large OIDs are
4731 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4736 * Pointer arithmetic undefined behaviour
4738 Avoid some undefined pointer arithmetic
4740 A common idiom in the codebase is to check limits in the following manner:
4743 Where "p" points to some malloc'd data of SIZE bytes and
4746 "len" here could be from some externally supplied data (e.g. from a TLS
4749 The rules of C pointer arithmetic are such that "p + len" is only well
4750 defined where len <= SIZE. Therefore the above idiom is actually
4751 undefined behaviour.
4753 For example this could cause problems if some malloc implementation
4754 provides an address for "p" such that "p + len" actually overflows for
4755 values of len that are too big and therefore p + len < limit.
4757 This issue was reported to OpenSSL by Guido Vranken
4762 * Constant time flag not preserved in DSA signing
4764 Operations in the DSA signing algorithm should run in constant time in
4765 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4766 implementation means that a non-constant time codepath is followed for
4767 certain operations. This has been demonstrated through a cache-timing
4768 attack to be sufficient for an attacker to recover the private DSA key.
4770 This issue was reported by César Pereida (Aalto University), Billy Brumley
4771 (Tampere University of Technology), and Yuval Yarom (The University of
4772 Adelaide and NICTA).
4777 * DTLS buffered message DoS
4779 In a DTLS connection where handshake messages are delivered out-of-order
4780 those messages that OpenSSL is not yet ready to process will be buffered
4781 for later use. Under certain circumstances, a flaw in the logic means that
4782 those messages do not get removed from the buffer even though the handshake
4783 has been completed. An attacker could force up to approx. 15 messages to
4784 remain in the buffer when they are no longer required. These messages will
4785 be cleared when the DTLS connection is closed. The default maximum size for
4786 a message is 100k. Therefore the attacker could force an additional 1500k
4787 to be consumed per connection. By opening many simulataneous connections an
4788 attacker could cause a DoS attack through memory exhaustion.
4790 This issue was reported to OpenSSL by Quan Luo.
4795 * DTLS replay protection DoS
4797 A flaw in the DTLS replay attack protection mechanism means that records
4798 that arrive for future epochs update the replay protection "window" before
4799 the MAC for the record has been validated. This could be exploited by an
4800 attacker by sending a record for the next epoch (which does not have to
4801 decrypt or have a valid MAC), with a very large sequence number. This means
4802 that all subsequent legitimate packets are dropped causing a denial of
4803 service for a specific DTLS connection.
4805 This issue was reported to OpenSSL by the OCAP audit team.
4810 * Certificate message OOB reads
4812 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4813 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4814 theoretical DoS risk but this has not been observed in practice on common
4817 The messages affected are client certificate, client certificate request
4818 and server certificate. As a result the attack can only be performed
4819 against a client or a server which enables client authentication.
4821 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4826 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4828 * Prevent padding oracle in AES-NI CBC MAC check
4830 A MITM attacker can use a padding oracle attack to decrypt traffic
4831 when the connection uses an AES CBC cipher and the server support
4834 This issue was introduced as part of the fix for Lucky 13 padding
4835 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4836 constant time by making sure that always the same bytes are read and
4837 compared against either the MAC or padding bytes. But it no longer
4838 checked that there was enough data to have both the MAC and padding
4841 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4846 * Fix EVP_EncodeUpdate overflow
4848 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4849 Base64 encoding of binary data. If an attacker is able to supply very large
4850 amounts of input data then a length check can overflow resulting in a heap
4853 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4854 the `PEM_write_bio*` family of functions. These are mainly used within the
4855 OpenSSL command line applications, so any application which processes data
4856 from an untrusted source and outputs it as a PEM file should be considered
4857 vulnerable to this issue. User applications that call these APIs directly
4858 with large amounts of untrusted data may also be vulnerable.
4860 This issue was reported by Guido Vranken.
4865 * Fix EVP_EncryptUpdate overflow
4867 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4868 is able to supply very large amounts of input data after a previous call to
4869 EVP_EncryptUpdate() with a partial block then a length check can overflow
4870 resulting in a heap corruption. Following an analysis of all OpenSSL
4871 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4872 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4873 the first called function after an EVP_EncryptInit(), and therefore that
4874 specific call must be safe. The second form is where the length passed to
4875 EVP_EncryptUpdate() can be seen from the code to be some small value and
4876 therefore there is no possibility of an overflow. Since all instances are
4877 one of these two forms, it is believed that there can be no overflows in
4878 internal code due to this problem. It should be noted that
4879 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4880 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4881 of these calls have also been analysed too and it is believed there are no
4882 instances in internal usage where an overflow could occur.
4884 This issue was reported by Guido Vranken.
4889 * Prevent ASN.1 BIO excessive memory allocation
4891 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4892 a short invalid encoding can cause allocation of large amounts of memory
4893 potentially consuming excessive resources or exhausting memory.
4895 Any application parsing untrusted data through d2i BIO functions is
4896 affected. The memory based functions such as d2i_X509() are *not* affected.
4897 Since the memory based functions are used by the TLS library, TLS
4898 applications are not affected.
4900 This issue was reported by Brian Carpenter.
4907 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4908 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4909 in arbitrary stack data being returned in the buffer.
4911 This issue was reported by Guido Vranken.
4916 * Modify behavior of ALPN to invoke callback after SNI/servername
4917 callback, such that updates to the SSL_CTX affect ALPN.
4921 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4926 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4927 methods are enabled and ssl2 is disabled the methods return NULL.
4931 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4933 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4934 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4935 provide any "EXPORT" or "LOW" strength ciphers.
4939 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4940 is by default disabled at build-time. Builds that are not configured with
4941 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4942 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4943 will need to explicitly call either of:
4945 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4947 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4949 as appropriate. Even if either of those is used, or the application
4950 explicitly uses the version-specific SSLv2_method() or its client and
4951 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4952 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4953 ciphers, and SSLv2 56-bit DES are no longer available.
4958 * Fix a double-free in DSA code
4960 A double free bug was discovered when OpenSSL parses malformed DSA private
4961 keys and could lead to a DoS attack or memory corruption for applications
4962 that receive DSA private keys from untrusted sources. This scenario is
4965 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4971 * Disable SRP fake user seed to address a server memory leak.
4973 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4975 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4976 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4977 was changed to ignore the "fake user" SRP seed, even if the seed
4980 Users should use SRP_VBASE_get1_by_user instead. Note that in
4981 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4982 also that even though configuring the SRP seed attempts to hide
4983 invalid usernames by continuing the handshake with fake
4984 credentials, this behaviour is not constant time and no strong
4985 guarantees are made that the handshake is indistinguishable from
4986 that of a valid user.
4991 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4993 In the BN_hex2bn function the number of hex digits is calculated using an
4994 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4995 large values of `i` this can result in `bn_expand` not allocating any
4996 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4997 field as NULL leading to a subsequent NULL ptr deref. For very large values
4998 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4999 In this case memory is allocated to the internal BIGNUM data field, but it
5000 is insufficiently sized leading to heap corruption. A similar issue exists
5001 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5002 is ever called by user applications with very large untrusted hex/dec data.
5003 This is anticipated to be a rare occurrence.
5005 All OpenSSL internal usage of these functions use data that is not expected
5006 to be untrusted, e.g. config file data or application command line
5007 arguments. If user developed applications generate config file data based
5008 on untrusted data then it is possible that this could also lead to security
5009 consequences. This is also anticipated to be rare.
5011 This issue was reported to OpenSSL by Guido Vranken.
5016 * Fix memory issues in `BIO_*printf` functions
5018 The internal `fmtstr` function used in processing a "%s" format string in
5019 the `BIO_*printf` functions could overflow while calculating the length of a
5020 string and cause an OOB read when printing very long strings.
5022 Additionally the internal `doapr_outch` function can attempt to write to an
5023 OOB memory location (at an offset from the NULL pointer) in the event of a
5024 memory allocation failure. In 1.0.2 and below this could be caused where
5025 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5026 could be in processing a very long "%s" format string. Memory leaks can
5029 The first issue may mask the second issue dependent on compiler behaviour.
5030 These problems could enable attacks where large amounts of untrusted data
5031 is passed to the `BIO_*printf` functions. If applications use these functions
5032 in this way then they could be vulnerable. OpenSSL itself uses these
5033 functions when printing out human-readable dumps of ASN.1 data. Therefore
5034 applications that print this data could be vulnerable if the data is from
5035 untrusted sources. OpenSSL command line applications could also be
5036 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5037 as command line arguments.
5039 Libssl is not considered directly vulnerable. Additionally certificates etc
5040 received via remote connections via libssl are also unlikely to be able to
5041 trigger these issues because of message size limits enforced within libssl.
5043 This issue was reported to OpenSSL Guido Vranken.
5048 * Side channel attack on modular exponentiation
5050 A side-channel attack was found which makes use of cache-bank conflicts on
5051 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5052 of RSA keys. The ability to exploit this issue is limited as it relies on
5053 an attacker who has control of code in a thread running on the same
5054 hyper-threaded core as the victim thread which is performing decryptions.
5056 This issue was reported to OpenSSL by Yuval Yarom, The University of
5057 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5058 Nadia Heninger, University of Pennsylvania with more information at
5059 <http://cachebleed.info>.
5064 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5065 if no keysize is specified with default_bits. This fixes an
5066 omission in an earlier change that changed all RSA/DSA key generation
5067 commands to use 2048 bits by default.
5071 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5073 * DH small subgroups
5075 Historically OpenSSL only ever generated DH parameters based on "safe"
5076 primes. More recently (in version 1.0.2) support was provided for
5077 generating X9.42 style parameter files such as those required for RFC 5114
5078 support. The primes used in such files may not be "safe". Where an
5079 application is using DH configured with parameters based on primes that are
5080 not "safe" then an attacker could use this fact to find a peer's private
5081 DH exponent. This attack requires that the attacker complete multiple
5082 handshakes in which the peer uses the same private DH exponent. For example
5083 this could be used to discover a TLS server's private DH exponent if it's
5084 reusing the private DH exponent or it's using a static DH ciphersuite.
5086 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5087 TLS. It is not on by default. If the option is not set then the server
5088 reuses the same private DH exponent for the life of the server process and
5089 would be vulnerable to this attack. It is believed that many popular
5090 applications do set this option and would therefore not be at risk.
5092 The fix for this issue adds an additional check where a "q" parameter is
5093 available (as is the case in X9.42 based parameters). This detects the
5094 only known attack, and is the only possible defense for static DH
5095 ciphersuites. This could have some performance impact.
5097 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5098 default and cannot be disabled. This could have some performance impact.
5100 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5105 * SSLv2 doesn't block disabled ciphers
5107 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5108 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5109 been disabled, provided that the SSLv2 protocol was not also disabled via
5112 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5113 and Sebastian Schinzel.
5118 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5120 * BN_mod_exp may produce incorrect results on x86_64
5122 There is a carry propagating bug in the x86_64 Montgomery squaring
5123 procedure. No EC algorithms are affected. Analysis suggests that attacks
5124 against RSA and DSA as a result of this defect would be very difficult to
5125 perform and are not believed likely. Attacks against DH are considered just
5126 feasible (although very difficult) because most of the work necessary to
5127 deduce information about a private key may be performed offline. The amount
5128 of resources required for such an attack would be very significant and
5129 likely only accessible to a limited number of attackers. An attacker would
5130 additionally need online access to an unpatched system using the target
5131 private key in a scenario with persistent DH parameters and a private
5132 key that is shared between multiple clients. For example this can occur by
5133 default in OpenSSL DHE based SSL/TLS ciphersuites.
5135 This issue was reported to OpenSSL by Hanno Böck.
5140 * Certificate verify crash with missing PSS parameter
5142 The signature verification routines will crash with a NULL pointer
5143 dereference if presented with an ASN.1 signature using the RSA PSS
5144 algorithm and absent mask generation function parameter. Since these
5145 routines are used to verify certificate signature algorithms this can be
5146 used to crash any certificate verification operation and exploited in a
5147 DoS attack. Any application which performs certificate verification is
5148 vulnerable including OpenSSL clients and servers which enable client
5151 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5156 * X509_ATTRIBUTE memory leak
5158 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5159 memory. This structure is used by the PKCS#7 and CMS routines so any
5160 application which reads PKCS#7 or CMS data from untrusted sources is
5161 affected. SSL/TLS is not affected.
5163 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5169 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5170 This changes the decoding behaviour for some invalid messages,
5171 though the change is mostly in the more lenient direction, and
5172 legacy behaviour is preserved as much as possible.
5176 * In DSA_generate_parameters_ex, if the provided seed is too short,
5179 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5181 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5183 * Alternate chains certificate forgery
5185 During certificate verification, OpenSSL will attempt to find an
5186 alternative certificate chain if the first attempt to build such a chain
5187 fails. An error in the implementation of this logic can mean that an
5188 attacker could cause certain checks on untrusted certificates to be
5189 bypassed, such as the CA flag, enabling them to use a valid leaf
5190 certificate to act as a CA and "issue" an invalid certificate.
5192 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5197 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5199 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5200 incompatibility in the handling of HMAC. The previous ABI has now been
5205 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5207 * Malformed ECParameters causes infinite loop
5209 When processing an ECParameters structure OpenSSL enters an infinite loop
5210 if the curve specified is over a specially malformed binary polynomial
5213 This can be used to perform denial of service against any
5214 system which processes public keys, certificate requests or
5215 certificates. This includes TLS clients and TLS servers with
5216 client authentication enabled.
5218 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5223 * Exploitable out-of-bounds read in X509_cmp_time
5225 X509_cmp_time does not properly check the length of the ASN1_TIME
5226 string and can read a few bytes out of bounds. In addition,
5227 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5230 An attacker can use this to craft malformed certificates and CRLs of
5231 various sizes and potentially cause a segmentation fault, resulting in
5232 a DoS on applications that verify certificates or CRLs. TLS clients
5233 that verify CRLs are affected. TLS clients and servers with client
5234 authentication enabled may be affected if they use custom verification
5237 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5238 independently by Hanno Böck.
5243 * PKCS7 crash with missing EnvelopedContent
5245 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5246 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5247 with missing content and trigger a NULL pointer dereference on parsing.
5249 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5250 structures from untrusted sources are affected. OpenSSL clients and
5251 servers are not affected.
5253 This issue was reported to OpenSSL by Michal Zalewski (Google).
5258 * CMS verify infinite loop with unknown hash function
5260 When verifying a signedData message the CMS code can enter an infinite loop
5261 if presented with an unknown hash function OID. This can be used to perform
5262 denial of service against any system which verifies signedData messages using
5264 This issue was reported to OpenSSL by Johannes Bauer.
5269 * Race condition handling NewSessionTicket
5271 If a NewSessionTicket is received by a multi-threaded client when attempting to
5272 reuse a previous ticket then a race condition can occur potentially leading to
5273 a double free of the ticket data.
5278 * Only support 256-bit or stronger elliptic curves with the
5279 'ecdh_auto' setting (server) or by default (client). Of supported
5280 curves, prefer P-256 (both).
5284 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5286 * ClientHello sigalgs DoS fix
5288 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5289 invalid signature algorithms extension a NULL pointer dereference will
5290 occur. This can be exploited in a DoS attack against the server.
5292 This issue was was reported to OpenSSL by David Ramos of Stanford
5296 *Stephen Henson and Matt Caswell*
5298 * Multiblock corrupted pointer fix
5300 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5301 feature only applies on 64 bit x86 architecture platforms that support AES
5302 NI instructions. A defect in the implementation of "multiblock" can cause
5303 OpenSSL's internal write buffer to become incorrectly set to NULL when
5304 using non-blocking IO. Typically, when the user application is using a
5305 socket BIO for writing, this will only result in a failed connection.
5306 However if some other BIO is used then it is likely that a segmentation
5307 fault will be triggered, thus enabling a potential DoS attack.
5309 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5314 * Segmentation fault in DTLSv1_listen fix
5316 The DTLSv1_listen function is intended to be stateless and processes the
5317 initial ClientHello from many peers. It is common for user code to loop
5318 over the call to DTLSv1_listen until a valid ClientHello is received with
5319 an associated cookie. A defect in the implementation of DTLSv1_listen means
5320 that state is preserved in the SSL object from one invocation to the next
5321 that can lead to a segmentation fault. Errors processing the initial
5322 ClientHello can trigger this scenario. An example of such an error could be
5323 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5326 This issue was reported to OpenSSL by Per Allansson.
5331 * Segmentation fault in ASN1_TYPE_cmp fix
5333 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5334 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5335 certificate signature algorithm consistency this can be used to crash any
5336 certificate verification operation and exploited in a DoS attack. Any
5337 application which performs certificate verification is vulnerable including
5338 OpenSSL clients and servers which enable client authentication.
5343 * Segmentation fault for invalid PSS parameters fix
5345 The signature verification routines will crash with a NULL pointer
5346 dereference if presented with an ASN.1 signature using the RSA PSS
5347 algorithm and invalid parameters. Since these routines are used to verify
5348 certificate signature algorithms this can be used to crash any
5349 certificate verification operation and exploited in a DoS attack. Any
5350 application which performs certificate verification is vulnerable including
5351 OpenSSL clients and servers which enable client authentication.
5353 This issue was was reported to OpenSSL by Brian Carpenter.
5358 * ASN.1 structure reuse memory corruption fix
5360 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5361 memory corruption via an invalid write. Such reuse is and has been
5362 strongly discouraged and is believed to be rare.
5364 Applications that parse structures containing CHOICE or ANY DEFINED BY
5365 components may be affected. Certificate parsing (d2i_X509 and related
5366 functions) are however not affected. OpenSSL clients and servers are
5372 * PKCS7 NULL pointer dereferences fix
5374 The PKCS#7 parsing code does not handle missing outer ContentInfo
5375 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5376 missing content and trigger a NULL pointer dereference on parsing.
5378 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5379 otherwise parse PKCS#7 structures from untrusted sources are
5380 affected. OpenSSL clients and servers are not affected.
5382 This issue was reported to OpenSSL by Michal Zalewski (Google).
5387 * DoS via reachable assert in SSLv2 servers fix
5389 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5390 servers that both support SSLv2 and enable export cipher suites by sending
5391 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5393 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5394 (OpenSSL development team).
5399 * Empty CKE with client auth and DHE fix
5401 If client auth is used then a server can seg fault in the event of a DHE
5402 ciphersuite being selected and a zero length ClientKeyExchange message
5403 being sent by the client. This could be exploited in a DoS attack.
5408 * Handshake with unseeded PRNG fix
5410 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5411 with an unseeded PRNG. The conditions are:
5412 - The client is on a platform where the PRNG has not been seeded
5413 automatically, and the user has not seeded manually
5414 - A protocol specific client method version has been used (i.e. not
5415 SSL_client_methodv23)
5416 - A ciphersuite is used that does not require additional random data from
5417 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5419 If the handshake succeeds then the client random that has been used will
5420 have been generated from a PRNG with insufficient entropy and therefore the
5421 output may be predictable.
5423 For example using the following command with an unseeded openssl will
5424 succeed on an unpatched platform:
5426 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5431 * Use After Free following d2i_ECPrivatekey error fix
5433 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5434 could cause a use after free condition. This, in turn, could cause a double
5435 free in several private key parsing functions (such as d2i_PrivateKey
5436 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5437 for applications that receive EC private keys from untrusted
5438 sources. This scenario is considered rare.
5440 This issue was discovered by the BoringSSL project and fixed in their
5446 * X509_to_X509_REQ NULL pointer deref fix
5448 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5449 the certificate key is invalid. This function is rarely used in practice.
5451 This issue was discovered by Brian Carpenter.
5456 * Removed the export ciphers from the DEFAULT ciphers
5460 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5462 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5463 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5464 So far those who have to target multiple platforms would compromise
5465 and argue that binary targeting say ARMv5 would still execute on
5466 ARMv8. "Universal" build resolves this compromise by providing
5467 near-optimal performance even on newer platforms.
5471 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5472 (other platforms pending).
5474 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5476 * Add support for the SignedCertificateTimestampList certificate and
5477 OCSP response extensions from RFC6962.
5481 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5482 for corner cases. (Certain input points at infinity could lead to
5483 bogus results, with non-infinity inputs mapped to infinity too.)
5487 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5488 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5489 common cases are optimized and there still is room for further
5490 improvements. Vector Permutation AES for Altivec is also added.
5494 * Add support for little-endian ppc64 Linux target.
5496 *Marcelo Cerri (IBM)*
5498 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5499 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5500 are optimized and there still is room for further improvements.
5501 Both 32- and 64-bit modes are supported.
5503 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5505 * Improved ARMv7 NEON support.
5509 * Support for SPARC Architecture 2011 crypto extensions, first
5510 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5511 SHA256/512, MD5, GHASH and modular exponentiation.
5513 *Andy Polyakov, David Miller*
5515 * Accelerated modular exponentiation for Intel processors, a.k.a.
5518 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5520 * Support for new and upcoming Intel processors, including AVX2,
5521 BMI and SHA ISA extensions. This includes additional "stitched"
5522 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5525 This work was sponsored by Intel Corp.
5529 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5530 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5531 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5535 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5536 this fixes a limitation in previous versions of OpenSSL.
5540 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5541 MGF1 digest and OAEP label.
5545 * Add EVP support for key wrapping algorithms, to avoid problems with
5546 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5547 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5548 algorithms and include tests cases.
5552 * Add functions to allocate and set the fields of an ECDSA_METHOD
5555 *Douglas E. Engert, Steve Henson*
5557 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5558 difference in days and seconds between two tm or ASN1_TIME structures.
5562 * Add -rev test option to s_server to just reverse order of characters
5563 received by client and send back to server. Also prints an abbreviated
5564 summary of the connection parameters.
5568 * New option -brief for s_client and s_server to print out a brief summary
5569 of connection parameters.
5573 * Add callbacks for arbitrary TLS extensions.
5575 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5577 * New option -crl_download in several openssl utilities to download CRLs
5578 from CRLDP extension in certificates.
5582 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5586 * New function X509_CRL_diff to generate a delta CRL from the difference
5587 of two full CRLs. Add support to "crl" utility.
5591 * New functions to set lookup_crls function and to retrieve
5592 X509_STORE from X509_STORE_CTX.
5596 * Print out deprecated issuer and subject unique ID fields in
5601 * Extend OCSP I/O functions so they can be used for simple general purpose
5602 HTTP as well as OCSP. New wrapper function which can be used to download
5603 CRLs using the OCSP API.
5607 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5611 * `SSL_CONF*` functions. These provide a common framework for application
5612 configuration using configuration files or command lines.
5616 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5617 message callback and prints the results. Needs compile time option
5618 "enable-ssl-trace". New options to s_client and s_server to enable
5623 * New ctrl and macro to retrieve supported points extensions.
5624 Print out extension in s_server and s_client.
5628 * New functions to retrieve certificate signature and signature
5633 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5638 * New Suite B modes for TLS code. These use and enforce the requirements
5639 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5640 only use Suite B curves. The Suite B modes can be set by using the
5641 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5645 * New chain verification flags for Suite B levels of security. Check
5646 algorithms are acceptable when flags are set in X509_verify_cert.
5650 * Make tls1_check_chain return a set of flags indicating checks passed
5651 by a certificate chain. Add additional tests to handle client
5652 certificates: checks for matching certificate type and issuer name
5657 * If an attempt is made to use a signature algorithm not in the peer
5658 preference list abort the handshake. If client has no suitable
5659 signature algorithms in response to a certificate request do not
5660 use the certificate.
5664 * If server EC tmp key is not in client preference list abort handshake.
5668 * Add support for certificate stores in CERT structure. This makes it
5669 possible to have different stores per SSL structure or one store in
5670 the parent SSL_CTX. Include distinct stores for certificate chain
5671 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5672 to build and store a certificate chain in CERT structure: returning
5673 an error if the chain cannot be built: this will allow applications
5674 to test if a chain is correctly configured.
5676 Note: if the CERT based stores are not set then the parent SSL_CTX
5677 store is used to retain compatibility with existing behaviour.
5681 * New function ssl_set_client_disabled to set a ciphersuite disabled
5682 mask based on the current session, check mask when sending client
5683 hello and checking the requested ciphersuite.
5687 * New ctrls to retrieve and set certificate types in a certificate
5688 request message. Print out received values in s_client. If certificate
5689 types is not set with custom values set sensible values based on
5690 supported signature algorithms.
5694 * Support for distinct client and server supported signature algorithms.
5698 * Add certificate callback. If set this is called whenever a certificate
5699 is required by client or server. An application can decide which
5700 certificate chain to present based on arbitrary criteria: for example
5701 supported signature algorithms. Add very simple example to s_server.
5702 This fixes many of the problems and restrictions of the existing client
5703 certificate callback: for example you can now clear an existing
5704 certificate and specify the whole chain.
5708 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5709 the certificate can be used for (if anything). Set valid_flags field
5710 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5711 to have similar checks in it.
5713 Add new "cert_flags" field to CERT structure and include a "strict mode".
5714 This enforces some TLS certificate requirements (such as only permitting
5715 certificate signature algorithms contained in the supported algorithms
5716 extension) which some implementations ignore: this option should be used
5717 with caution as it could cause interoperability issues.
5721 * Update and tidy signature algorithm extension processing. Work out
5722 shared signature algorithms based on preferences and peer algorithms
5723 and print them out in s_client and s_server. Abort handshake if no
5724 shared signature algorithms.
5728 * Add new functions to allow customised supported signature algorithms
5729 for SSL and SSL_CTX structures. Add options to s_client and s_server
5734 * New function SSL_certs_clear() to delete all references to certificates
5735 from an SSL structure. Before this once a certificate had been added
5736 it couldn't be removed.
5740 * Integrate hostname, email address and IP address checking with certificate
5741 verification. New verify options supporting checking in openssl utility.
5745 * Fixes and wildcard matching support to hostname and email checking
5746 functions. Add manual page.
5748 *Florian Weimer (Red Hat Product Security Team)*
5750 * New functions to check a hostname email or IP address against a
5751 certificate. Add options x509 utility to print results of checks against
5756 * Fix OCSP checking.
5758 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5760 * Initial experimental support for explicitly trusted non-root CAs.
5761 OpenSSL still tries to build a complete chain to a root but if an
5762 intermediate CA has a trust setting included that is used. The first
5763 setting is used: whether to trust (e.g., -addtrust option to the x509
5768 * Add -trusted_first option which attempts to find certificates in the
5769 trusted store even if an untrusted chain is also supplied.
5773 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5774 platform support for Linux and Android.
5778 * Support for linux-x32, ILP32 environment in x86_64 framework.
5782 * Experimental multi-implementation support for FIPS capable OpenSSL.
5783 When in FIPS mode the approved implementations are used as normal,
5784 when not in FIPS mode the internal unapproved versions are used instead.
5785 This means that the FIPS capable OpenSSL isn't forced to use the
5786 (often lower performance) FIPS implementations outside FIPS mode.
5790 * Transparently support X9.42 DH parameters when calling
5791 PEM_read_bio_DHparameters. This means existing applications can handle
5792 the new parameter format automatically.
5796 * Initial experimental support for X9.42 DH parameter format: mainly
5797 to support use of 'q' parameter for RFC5114 parameters.
5801 * Add DH parameters from RFC5114 including test data to dhtest.
5805 * Support for automatic EC temporary key parameter selection. If enabled
5806 the most preferred EC parameters are automatically used instead of
5807 hardcoded fixed parameters. Now a server just has to call:
5808 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5809 support ECDH and use the most appropriate parameters.
5813 * Enhance and tidy EC curve and point format TLS extension code. Use
5814 static structures instead of allocation if default values are used.
5815 New ctrls to set curves we wish to support and to retrieve shared curves.
5816 Print out shared curves in s_server. New options to s_server and s_client
5817 to set list of supported curves.
5821 * New ctrls to retrieve supported signature algorithms and
5822 supported curve values as an array of NIDs. Extend openssl utility
5823 to print out received values.
5827 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5828 between NIDs and the more common NIST names such as "P-256". Enhance
5829 ecparam utility and ECC method to recognise the NIST names for curves.
5833 * Enhance SSL/TLS certificate chain handling to support different
5834 chains for each certificate instead of one chain in the parent SSL_CTX.
5838 * Support for fixed DH ciphersuite client authentication: where both
5839 server and client use DH certificates with common parameters.
5843 * Support for fixed DH ciphersuites: those requiring DH server
5848 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5850 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5851 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5852 X509_CINF_get_signature were reverted post internal team review.
5857 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5859 * OCSP Status Request extension unbounded memory growth
5861 A malicious client can send an excessively large OCSP Status Request
5862 extension. If that client continually requests renegotiation, sending a
5863 large OCSP Status Request extension each time, then there will be unbounded
5864 memory growth on the server. This will eventually lead to a Denial Of
5865 Service attack through memory exhaustion. Servers with a default
5866 configuration are vulnerable even if they do not support OCSP. Builds using
5867 the "no-ocsp" build time option are not affected.
5869 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5874 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5877 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5883 * OOB write in MDC2_Update()
5885 An overflow can occur in MDC2_Update() either if called directly or
5886 through the EVP_DigestUpdate() function using MDC2. If an attacker
5887 is able to supply very large amounts of input data after a previous
5888 call to EVP_EncryptUpdate() with a partial block then a length check
5889 can overflow resulting in a heap corruption.
5891 The amount of data needed is comparable to SIZE_MAX which is impractical
5894 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5899 * Malformed SHA512 ticket DoS
5901 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5902 DoS attack where a malformed ticket will result in an OOB read which will
5905 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5906 a custom server callback and ticket lookup mechanism.
5908 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5913 * OOB write in BN_bn2dec()
5915 The function BN_bn2dec() does not check the return value of BN_div_word().
5916 This can cause an OOB write if an application uses this function with an
5917 overly large BIGNUM. This could be a problem if an overly large certificate
5918 or CRL is printed out from an untrusted source. TLS is not affected because
5919 record limits will reject an oversized certificate before it is parsed.
5921 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5926 * OOB read in TS_OBJ_print_bio()
5928 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5929 the total length the OID text representation would use and not the amount
5930 of data written. This will result in OOB reads when large OIDs are
5933 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5938 * Pointer arithmetic undefined behaviour
5940 Avoid some undefined pointer arithmetic
5942 A common idiom in the codebase is to check limits in the following manner:
5945 Where "p" points to some malloc'd data of SIZE bytes and
5948 "len" here could be from some externally supplied data (e.g. from a TLS
5951 The rules of C pointer arithmetic are such that "p + len" is only well
5952 defined where len <= SIZE. Therefore the above idiom is actually
5953 undefined behaviour.
5955 For example this could cause problems if some malloc implementation
5956 provides an address for "p" such that "p + len" actually overflows for
5957 values of len that are too big and therefore p + len < limit.
5959 This issue was reported to OpenSSL by Guido Vranken
5964 * Constant time flag not preserved in DSA signing
5966 Operations in the DSA signing algorithm should run in constant time in
5967 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5968 implementation means that a non-constant time codepath is followed for
5969 certain operations. This has been demonstrated through a cache-timing
5970 attack to be sufficient for an attacker to recover the private DSA key.
5972 This issue was reported by César Pereida (Aalto University), Billy Brumley
5973 (Tampere University of Technology), and Yuval Yarom (The University of
5974 Adelaide and NICTA).
5979 * DTLS buffered message DoS
5981 In a DTLS connection where handshake messages are delivered out-of-order
5982 those messages that OpenSSL is not yet ready to process will be buffered
5983 for later use. Under certain circumstances, a flaw in the logic means that
5984 those messages do not get removed from the buffer even though the handshake
5985 has been completed. An attacker could force up to approx. 15 messages to
5986 remain in the buffer when they are no longer required. These messages will
5987 be cleared when the DTLS connection is closed. The default maximum size for
5988 a message is 100k. Therefore the attacker could force an additional 1500k
5989 to be consumed per connection. By opening many simulataneous connections an
5990 attacker could cause a DoS attack through memory exhaustion.
5992 This issue was reported to OpenSSL by Quan Luo.
5997 * DTLS replay protection DoS
5999 A flaw in the DTLS replay attack protection mechanism means that records
6000 that arrive for future epochs update the replay protection "window" before
6001 the MAC for the record has been validated. This could be exploited by an
6002 attacker by sending a record for the next epoch (which does not have to
6003 decrypt or have a valid MAC), with a very large sequence number. This means
6004 that all subsequent legitimate packets are dropped causing a denial of
6005 service for a specific DTLS connection.
6007 This issue was reported to OpenSSL by the OCAP audit team.
6012 * Certificate message OOB reads
6014 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6015 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6016 theoretical DoS risk but this has not been observed in practice on common
6019 The messages affected are client certificate, client certificate request
6020 and server certificate. As a result the attack can only be performed
6021 against a client or a server which enables client authentication.
6023 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6028 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6030 * Prevent padding oracle in AES-NI CBC MAC check
6032 A MITM attacker can use a padding oracle attack to decrypt traffic
6033 when the connection uses an AES CBC cipher and the server support
6036 This issue was introduced as part of the fix for Lucky 13 padding
6037 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6038 constant time by making sure that always the same bytes are read and
6039 compared against either the MAC or padding bytes. But it no longer
6040 checked that there was enough data to have both the MAC and padding
6043 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6048 * Fix EVP_EncodeUpdate overflow
6050 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6051 Base64 encoding of binary data. If an attacker is able to supply very large
6052 amounts of input data then a length check can overflow resulting in a heap
6055 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6056 the `PEM_write_bio*` family of functions. These are mainly used within the
6057 OpenSSL command line applications, so any application which processes data
6058 from an untrusted source and outputs it as a PEM file should be considered
6059 vulnerable to this issue. User applications that call these APIs directly
6060 with large amounts of untrusted data may also be vulnerable.
6062 This issue was reported by Guido Vranken.
6067 * Fix EVP_EncryptUpdate overflow
6069 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6070 is able to supply very large amounts of input data after a previous call to
6071 EVP_EncryptUpdate() with a partial block then a length check can overflow
6072 resulting in a heap corruption. Following an analysis of all OpenSSL
6073 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6074 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6075 the first called function after an EVP_EncryptInit(), and therefore that
6076 specific call must be safe. The second form is where the length passed to
6077 EVP_EncryptUpdate() can be seen from the code to be some small value and
6078 therefore there is no possibility of an overflow. Since all instances are
6079 one of these two forms, it is believed that there can be no overflows in
6080 internal code due to this problem. It should be noted that
6081 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6082 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6083 of these calls have also been analysed too and it is believed there are no
6084 instances in internal usage where an overflow could occur.
6086 This issue was reported by Guido Vranken.
6091 * Prevent ASN.1 BIO excessive memory allocation
6093 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6094 a short invalid encoding can casuse allocation of large amounts of memory
6095 potentially consuming excessive resources or exhausting memory.
6097 Any application parsing untrusted data through d2i BIO functions is
6098 affected. The memory based functions such as d2i_X509() are *not* affected.
6099 Since the memory based functions are used by the TLS library, TLS
6100 applications are not affected.
6102 This issue was reported by Brian Carpenter.
6109 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6110 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6111 in arbitrary stack data being returned in the buffer.
6113 This issue was reported by Guido Vranken.
6118 * Modify behavior of ALPN to invoke callback after SNI/servername
6119 callback, such that updates to the SSL_CTX affect ALPN.
6123 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6128 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6129 methods are enabled and ssl2 is disabled the methods return NULL.
6133 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6135 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6136 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6137 provide any "EXPORT" or "LOW" strength ciphers.
6141 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6142 is by default disabled at build-time. Builds that are not configured with
6143 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6144 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6145 will need to explicitly call either of:
6147 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6149 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6151 as appropriate. Even if either of those is used, or the application
6152 explicitly uses the version-specific SSLv2_method() or its client and
6153 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6154 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6155 ciphers, and SSLv2 56-bit DES are no longer available.
6160 * Fix a double-free in DSA code
6162 A double free bug was discovered when OpenSSL parses malformed DSA private
6163 keys and could lead to a DoS attack or memory corruption for applications
6164 that receive DSA private keys from untrusted sources. This scenario is
6167 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6173 * Disable SRP fake user seed to address a server memory leak.
6175 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6177 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6178 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6179 was changed to ignore the "fake user" SRP seed, even if the seed
6182 Users should use SRP_VBASE_get1_by_user instead. Note that in
6183 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6184 also that even though configuring the SRP seed attempts to hide
6185 invalid usernames by continuing the handshake with fake
6186 credentials, this behaviour is not constant time and no strong
6187 guarantees are made that the handshake is indistinguishable from
6188 that of a valid user.
6193 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6195 In the BN_hex2bn function the number of hex digits is calculated using an
6196 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6197 large values of `i` this can result in `bn_expand` not allocating any
6198 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6199 field as NULL leading to a subsequent NULL ptr deref. For very large values
6200 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6201 In this case memory is allocated to the internal BIGNUM data field, but it
6202 is insufficiently sized leading to heap corruption. A similar issue exists
6203 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6204 is ever called by user applications with very large untrusted hex/dec data.
6205 This is anticipated to be a rare occurrence.
6207 All OpenSSL internal usage of these functions use data that is not expected
6208 to be untrusted, e.g. config file data or application command line
6209 arguments. If user developed applications generate config file data based
6210 on untrusted data then it is possible that this could also lead to security
6211 consequences. This is also anticipated to be rare.
6213 This issue was reported to OpenSSL by Guido Vranken.
6218 * Fix memory issues in `BIO_*printf` functions
6220 The internal `fmtstr` function used in processing a "%s" format string in
6221 the `BIO_*printf` functions could overflow while calculating the length of a
6222 string and cause an OOB read when printing very long strings.
6224 Additionally the internal `doapr_outch` function can attempt to write to an
6225 OOB memory location (at an offset from the NULL pointer) in the event of a
6226 memory allocation failure. In 1.0.2 and below this could be caused where
6227 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6228 could be in processing a very long "%s" format string. Memory leaks can
6231 The first issue may mask the second issue dependent on compiler behaviour.
6232 These problems could enable attacks where large amounts of untrusted data
6233 is passed to the `BIO_*printf` functions. If applications use these functions
6234 in this way then they could be vulnerable. OpenSSL itself uses these
6235 functions when printing out human-readable dumps of ASN.1 data. Therefore
6236 applications that print this data could be vulnerable if the data is from
6237 untrusted sources. OpenSSL command line applications could also be
6238 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6239 as command line arguments.
6241 Libssl is not considered directly vulnerable. Additionally certificates etc
6242 received via remote connections via libssl are also unlikely to be able to
6243 trigger these issues because of message size limits enforced within libssl.
6245 This issue was reported to OpenSSL Guido Vranken.
6250 * Side channel attack on modular exponentiation
6252 A side-channel attack was found which makes use of cache-bank conflicts on
6253 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6254 of RSA keys. The ability to exploit this issue is limited as it relies on
6255 an attacker who has control of code in a thread running on the same
6256 hyper-threaded core as the victim thread which is performing decryptions.
6258 This issue was reported to OpenSSL by Yuval Yarom, The University of
6259 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6260 Nadia Heninger, University of Pennsylvania with more information at
6261 <http://cachebleed.info>.
6266 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6267 if no keysize is specified with default_bits. This fixes an
6268 omission in an earlier change that changed all RSA/DSA key generation
6269 commands to use 2048 bits by default.
6273 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6275 * Protection for DH small subgroup attacks
6277 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6278 switched on by default and cannot be disabled. This could have some
6283 * SSLv2 doesn't block disabled ciphers
6285 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6286 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6287 been disabled, provided that the SSLv2 protocol was not also disabled via
6290 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6291 and Sebastian Schinzel.
6296 * Reject DH handshakes with parameters shorter than 1024 bits.
6300 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6302 * Certificate verify crash with missing PSS parameter
6304 The signature verification routines will crash with a NULL pointer
6305 dereference if presented with an ASN.1 signature using the RSA PSS
6306 algorithm and absent mask generation function parameter. Since these
6307 routines are used to verify certificate signature algorithms this can be
6308 used to crash any certificate verification operation and exploited in a
6309 DoS attack. Any application which performs certificate verification is
6310 vulnerable including OpenSSL clients and servers which enable client
6313 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6318 * X509_ATTRIBUTE memory leak
6320 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6321 memory. This structure is used by the PKCS#7 and CMS routines so any
6322 application which reads PKCS#7 or CMS data from untrusted sources is
6323 affected. SSL/TLS is not affected.
6325 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6331 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6332 This changes the decoding behaviour for some invalid messages,
6333 though the change is mostly in the more lenient direction, and
6334 legacy behaviour is preserved as much as possible.
6338 * In DSA_generate_parameters_ex, if the provided seed is too short,
6339 use a random seed, as already documented.
6341 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6343 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6345 * Alternate chains certificate forgery
6347 During certificate verfification, OpenSSL will attempt to find an
6348 alternative certificate chain if the first attempt to build such a chain
6349 fails. An error in the implementation of this logic can mean that an
6350 attacker could cause certain checks on untrusted certificates to be
6351 bypassed, such as the CA flag, enabling them to use a valid leaf
6352 certificate to act as a CA and "issue" an invalid certificate.
6354 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6360 * Race condition handling PSK identify hint
6362 If PSK identity hints are received by a multi-threaded client then
6363 the values are wrongly updated in the parent SSL_CTX structure. This can
6364 result in a race condition potentially leading to a double free of the
6370 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6372 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6373 incompatibility in the handling of HMAC. The previous ABI has now been
6376 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6378 * Malformed ECParameters causes infinite loop
6380 When processing an ECParameters structure OpenSSL enters an infinite loop
6381 if the curve specified is over a specially malformed binary polynomial
6384 This can be used to perform denial of service against any
6385 system which processes public keys, certificate requests or
6386 certificates. This includes TLS clients and TLS servers with
6387 client authentication enabled.
6389 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6394 * Exploitable out-of-bounds read in X509_cmp_time
6396 X509_cmp_time does not properly check the length of the ASN1_TIME
6397 string and can read a few bytes out of bounds. In addition,
6398 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6401 An attacker can use this to craft malformed certificates and CRLs of
6402 various sizes and potentially cause a segmentation fault, resulting in
6403 a DoS on applications that verify certificates or CRLs. TLS clients
6404 that verify CRLs are affected. TLS clients and servers with client
6405 authentication enabled may be affected if they use custom verification
6408 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6409 independently by Hanno Böck.
6414 * PKCS7 crash with missing EnvelopedContent
6416 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6417 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6418 with missing content and trigger a NULL pointer dereference on parsing.
6420 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6421 structures from untrusted sources are affected. OpenSSL clients and
6422 servers are not affected.
6424 This issue was reported to OpenSSL by Michal Zalewski (Google).
6429 * CMS verify infinite loop with unknown hash function
6431 When verifying a signedData message the CMS code can enter an infinite loop
6432 if presented with an unknown hash function OID. This can be used to perform
6433 denial of service against any system which verifies signedData messages using
6435 This issue was reported to OpenSSL by Johannes Bauer.
6440 * Race condition handling NewSessionTicket
6442 If a NewSessionTicket is received by a multi-threaded client when attempting to
6443 reuse a previous ticket then a race condition can occur potentially leading to
6444 a double free of the ticket data.
6449 * Reject DH handshakes with parameters shorter than 768 bits.
6451 *Kurt Roeckx and Emilia Kasper*
6453 * dhparam: generate 2048-bit parameters by default.
6455 *Kurt Roeckx and Emilia Kasper*
6457 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6459 * Segmentation fault in ASN1_TYPE_cmp fix
6461 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6462 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6463 certificate signature algorithm consistency this can be used to crash any
6464 certificate verification operation and exploited in a DoS attack. Any
6465 application which performs certificate verification is vulnerable including
6466 OpenSSL clients and servers which enable client authentication.
6471 * ASN.1 structure reuse memory corruption fix
6473 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6474 memory corruption via an invalid write. Such reuse is and has been
6475 strongly discouraged and is believed to be rare.
6477 Applications that parse structures containing CHOICE or ANY DEFINED BY
6478 components may be affected. Certificate parsing (d2i_X509 and related
6479 functions) are however not affected. OpenSSL clients and servers are
6485 * PKCS7 NULL pointer dereferences fix
6487 The PKCS#7 parsing code does not handle missing outer ContentInfo
6488 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6489 missing content and trigger a NULL pointer dereference on parsing.
6491 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6492 otherwise parse PKCS#7 structures from untrusted sources are
6493 affected. OpenSSL clients and servers are not affected.
6495 This issue was reported to OpenSSL by Michal Zalewski (Google).
6500 * DoS via reachable assert in SSLv2 servers fix
6502 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6503 servers that both support SSLv2 and enable export cipher suites by sending
6504 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6506 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6507 (OpenSSL development team).
6512 * Use After Free following d2i_ECPrivatekey error fix
6514 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6515 could cause a use after free condition. This, in turn, could cause a double
6516 free in several private key parsing functions (such as d2i_PrivateKey
6517 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6518 for applications that receive EC private keys from untrusted
6519 sources. This scenario is considered rare.
6521 This issue was discovered by the BoringSSL project and fixed in their
6527 * X509_to_X509_REQ NULL pointer deref fix
6529 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6530 the certificate key is invalid. This function is rarely used in practice.
6532 This issue was discovered by Brian Carpenter.
6537 * Removed the export ciphers from the DEFAULT ciphers
6541 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6543 * Build fixes for the Windows and OpenVMS platforms
6545 *Matt Caswell and Richard Levitte*
6547 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6549 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6550 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6551 dereference. This could lead to a Denial Of Service attack. Thanks to
6552 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6557 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6558 dtls1_buffer_record function under certain conditions. In particular this
6559 could occur if an attacker sent repeated DTLS records with the same
6560 sequence number but for the next epoch. The memory leak could be exploited
6561 by an attacker in a Denial of Service attack through memory exhaustion.
6562 Thanks to Chris Mueller for reporting this issue.
6567 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6568 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6569 method would be set to NULL which could later result in a NULL pointer
6570 dereference. Thanks to Frank Schmirler for reporting this issue.
6575 * Abort handshake if server key exchange message is omitted for ephemeral
6578 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6579 reporting this issue.
6584 * Remove non-export ephemeral RSA code on client and server. This code
6585 violated the TLS standard by allowing the use of temporary RSA keys in
6586 non-export ciphersuites and could be used by a server to effectively
6587 downgrade the RSA key length used to a value smaller than the server
6588 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6589 INRIA or reporting this issue.
6594 * Fixed issue where DH client certificates are accepted without verification.
6595 An OpenSSL server will accept a DH certificate for client authentication
6596 without the certificate verify message. This effectively allows a client to
6597 authenticate without the use of a private key. This only affects servers
6598 which trust a client certificate authority which issues certificates
6599 containing DH keys: these are extremely rare and hardly ever encountered.
6600 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6606 * Ensure that the session ID context of an SSL is updated when its
6607 SSL_CTX is updated via SSL_set_SSL_CTX.
6609 The session ID context is typically set from the parent SSL_CTX,
6610 and can vary with the CTX.
6614 * Fix various certificate fingerprint issues.
6616 By using non-DER or invalid encodings outside the signed portion of a
6617 certificate the fingerprint can be changed without breaking the signature.
6618 Although no details of the signed portion of the certificate can be changed
6619 this can cause problems with some applications: e.g. those using the
6620 certificate fingerprint for blacklists.
6622 1. Reject signatures with non zero unused bits.
6624 If the BIT STRING containing the signature has non zero unused bits reject
6625 the signature. All current signature algorithms require zero unused bits.
6627 2. Check certificate algorithm consistency.
6629 Check the AlgorithmIdentifier inside TBS matches the one in the
6630 certificate signature. NB: this will result in signature failure
6631 errors for some broken certificates.
6633 Thanks to Konrad Kraszewski from Google for reporting this issue.
6635 3. Check DSA/ECDSA signatures use DER.
6637 Re-encode DSA/ECDSA signatures and compare with the original received
6638 signature. Return an error if there is a mismatch.
6640 This will reject various cases including garbage after signature
6641 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6642 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6643 (negative or with leading zeroes).
6645 Further analysis was conducted and fixes were developed by Stephen Henson
6646 of the OpenSSL core team.
6652 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6653 results on some platforms, including x86_64. This bug occurs at random
6654 with a very low probability, and is not known to be exploitable in any
6655 way, though its exact impact is difficult to determine. Thanks to Pieter
6656 Wuille (Blockstream) who reported this issue and also suggested an initial
6657 fix. Further analysis was conducted by the OpenSSL development team and
6658 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6659 the OpenSSL core team.
6664 * Do not resume sessions on the server if the negotiated protocol
6665 version does not match the session's version. Resuming with a different
6666 version, while not strictly forbidden by the RFC, is of questionable
6667 sanity and breaks all known clients.
6669 *David Benjamin, Emilia Käsper*
6671 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6672 early CCS messages during renegotiation. (Note that because
6673 renegotiation is encrypted, this early CCS was not exploitable.)
6677 * Tighten client-side session ticket handling during renegotiation:
6678 ensure that the client only accepts a session ticket if the server sends
6679 the extension anew in the ServerHello. Previously, a TLS client would
6680 reuse the old extension state and thus accept a session ticket if one was
6681 announced in the initial ServerHello.
6683 Similarly, ensure that the client requires a session ticket if one
6684 was advertised in the ServerHello. Previously, a TLS client would
6685 ignore a missing NewSessionTicket message.
6689 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6693 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6694 sends a carefully crafted handshake message, to cause OpenSSL to fail
6695 to free up to 64k of memory causing a memory leak. This could be
6696 exploited in a Denial Of Service attack. This issue affects OpenSSL
6697 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6698 whether SRTP is used or configured. Implementations of OpenSSL that
6699 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6701 The fix was developed by the OpenSSL team.
6706 * Session Ticket Memory Leak.
6708 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6709 integrity of that ticket is first verified. In the event of a session
6710 ticket integrity check failing, OpenSSL will fail to free memory
6711 causing a memory leak. By sending a large number of invalid session
6712 tickets an attacker could exploit this issue in a Denial Of Service
6718 * Build option no-ssl3 is incomplete.
6720 When OpenSSL is configured with "no-ssl3" as a build option, servers
6721 could accept and complete a SSL 3.0 handshake, and clients could be
6722 configured to send them.
6725 *Akamai and the OpenSSL team*
6727 * Add support for TLS_FALLBACK_SCSV.
6728 Client applications doing fallback retries should call
6729 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6732 *Adam Langley, Bodo Moeller*
6734 * Add additional DigestInfo checks.
6736 Re-encode DigestInto in DER and check against the original when
6737 verifying RSA signature: this will reject any improperly encoded
6738 DigestInfo structures.
6740 Note: this is a precautionary measure and no attacks are currently known.
6744 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6746 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6747 SRP code can be overrun an internal buffer. Add sanity check that
6748 g, A, B < N to SRP code.
6750 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6751 Group for discovering this issue.
6756 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6757 TLS 1.0 instead of higher protocol versions when the ClientHello message
6758 is badly fragmented. This allows a man-in-the-middle attacker to force a
6759 downgrade to TLS 1.0 even if both the server and the client support a
6760 higher protocol version, by modifying the client's TLS records.
6762 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6763 researching this issue.
6768 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6769 to a denial of service attack. A malicious server can crash the client
6770 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6771 ciphersuite and sending carefully crafted handshake messages.
6773 Thanks to Felix Gröbert (Google) for discovering and researching this
6779 * By sending carefully crafted DTLS packets an attacker could cause openssl
6780 to leak memory. This can be exploited through a Denial of Service attack.
6781 Thanks to Adam Langley for discovering and researching this issue.
6786 * An attacker can force openssl to consume large amounts of memory whilst
6787 processing DTLS handshake messages. This can be exploited through a
6788 Denial of Service attack.
6789 Thanks to Adam Langley for discovering and researching this issue.
6794 * An attacker can force an error condition which causes openssl to crash
6795 whilst processing DTLS packets due to memory being freed twice. This
6796 can be exploited through a Denial of Service attack.
6797 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6803 * If a multithreaded client connects to a malicious server using a resumed
6804 session and the server sends an ec point format extension it could write
6805 up to 255 bytes to freed memory.
6807 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6813 * A malicious server can crash an OpenSSL client with a null pointer
6814 dereference (read) by specifying an SRP ciphersuite even though it was not
6815 properly negotiated with the client. This can be exploited through a
6816 Denial of Service attack.
6818 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6819 discovering and researching this issue.
6824 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6825 X509_name_oneline, X509_name_print_ex et al. to leak some information
6826 from the stack. Applications may be affected if they echo pretty printing
6827 output to the attacker.
6829 Thanks to Ivan Fratric (Google) for discovering this issue.
6832 *Emilia Käsper, and Steve Henson*
6834 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6835 for corner cases. (Certain input points at infinity could lead to
6836 bogus results, with non-infinity inputs mapped to infinity too.)
6840 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6842 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6843 handshake can force the use of weak keying material in OpenSSL
6844 SSL/TLS clients and servers.
6846 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6847 researching this issue. ([CVE-2014-0224])
6849 *KIKUCHI Masashi, Steve Henson*
6851 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6852 OpenSSL DTLS client the code can be made to recurse eventually crashing
6855 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6858 *Imre Rad, Steve Henson*
6860 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6861 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6862 client or server. This is potentially exploitable to run arbitrary
6863 code on a vulnerable client or server.
6865 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6867 *Jüri Aedla, Steve Henson*
6869 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6870 are subject to a denial of service attack.
6872 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6873 this issue. ([CVE-2014-3470])
6875 *Felix Gröbert, Ivan Fratric, Steve Henson*
6877 * Harmonize version and its documentation. -f flag is used to display
6880 *mancha <mancha1@zoho.com>*
6882 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6883 in i2d_ECPrivateKey.
6885 *mancha <mancha1@zoho.com>*
6887 * Fix some double frees. These are not thought to be exploitable.
6889 *mancha <mancha1@zoho.com>*
6891 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6893 * A missing bounds check in the handling of the TLS heartbeat extension
6894 can be used to reveal up to 64k of memory to a connected client or
6897 Thanks for Neel Mehta of Google Security for discovering this bug and to
6898 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6899 preparing the fix ([CVE-2014-0160])
6901 *Adam Langley, Bodo Moeller*
6903 * Fix for the attack described in the paper "Recovering OpenSSL
6904 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6905 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6906 <http://eprint.iacr.org/2014/140>
6908 Thanks to Yuval Yarom and Naomi Benger for discovering this
6909 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6911 *Yuval Yarom and Naomi Benger*
6913 * TLS pad extension: draft-agl-tls-padding-03
6915 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6916 TLS client Hello record length value would otherwise be > 255 and
6917 less that 512 pad with a dummy extension containing zeroes so it
6918 is at least 512 bytes long.
6920 *Adam Langley, Steve Henson*
6922 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6924 * Fix for TLS record tampering bug. A carefully crafted invalid
6925 handshake could crash OpenSSL with a NULL pointer exception.
6926 Thanks to Anton Johansson for reporting this issues.
6929 * Keep original DTLS digest and encryption contexts in retransmission
6930 structures so we can use the previous session parameters if they need
6931 to be resent. ([CVE-2013-6450])
6935 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6936 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6937 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6938 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6939 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6940 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6942 *Rob Stradling, Adam Langley*
6944 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6946 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6947 supporting platforms or when small records were transferred.
6949 *Andy Polyakov, Steve Henson*
6951 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6953 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6955 This addresses the flaw in CBC record processing discovered by
6956 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6957 at: <http://www.isg.rhul.ac.uk/tls/>
6959 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6960 Security Group at Royal Holloway, University of London
6961 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6962 Emilia Käsper for the initial patch.
6965 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6967 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6968 ciphersuites which can be exploited in a denial of service attack.
6969 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6970 and detecting this bug and to Wolfgang Ettlinger
6971 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6976 * Return an error when checking OCSP signatures when key is NULL.
6977 This fixes a DoS attack. ([CVE-2013-0166])
6981 * Make openssl verify return errors.
6983 *Chris Palmer <palmer@google.com> and Ben Laurie*
6985 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6986 the right response is stapled. Also change SSL_get_certificate()
6987 so it returns the certificate actually sent.
6988 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6990 *Rob Stradling <rob.stradling@comodo.com>*
6992 * Fix possible deadlock when decoding public keys.
6996 * Don't use TLS 1.0 record version number in initial client hello
7001 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7003 * Sanity check record length before skipping explicit IV in TLS
7004 1.2, 1.1 and DTLS to fix DoS attack.
7006 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7007 fuzzing as a service testing platform.
7012 * Initialise tkeylen properly when encrypting CMS messages.
7013 Thanks to Solar Designer of Openwall for reporting this issue.
7017 * In FIPS mode don't try to use composite ciphers as they are not
7022 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7024 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7025 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7026 mean any application compiled against OpenSSL 1.0.0 headers setting
7027 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7028 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7029 0x10000000L Any application which was previously compiled against
7030 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7031 will need to be recompiled as a result. Letting be results in
7032 inability to disable specifically TLS 1.1 and in client context,
7033 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7037 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7038 disable just protocol X, but all protocols above X *if* there are
7039 protocols *below* X still enabled. In more practical terms it means
7040 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7041 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7042 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7047 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7049 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7050 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7051 in CRYPTO_realloc_clean.
7053 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7054 issue and to Adam Langley <agl@chromium.org> for fixing it.
7057 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7059 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7063 * Workarounds for some broken servers that "hang" if a client hello
7064 record length exceeds 255 bytes.
7066 1. Do not use record version number > TLS 1.0 in initial client
7067 hello: some (but not all) hanging servers will now work.
7068 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7069 the number of ciphers sent in the client hello. This should be
7070 set to an even number, such as 50, for example by passing:
7071 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7072 Most broken servers should now work.
7073 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7074 TLS 1.2 client support entirely.
7078 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7082 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7084 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7085 STRING form instead of a DigestInfo.
7089 * The format used for MDC2 RSA signatures is inconsistent between EVP
7090 and the RSA_sign/RSA_verify functions. This was made more apparent when
7091 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7092 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7093 the correct format in RSA_verify so both forms transparently work.
7097 * Some servers which support TLS 1.0 can choke if we initially indicate
7098 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7099 encrypted premaster secret. As a workaround use the maximum permitted
7100 client version in client hello, this should keep such servers happy
7101 and still work with previous versions of OpenSSL.
7105 * Add support for TLS/DTLS heartbeats.
7107 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7109 * Add support for SCTP.
7111 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7113 * Improved PRNG seeding for VOS.
7115 *Paul Green <Paul.Green@stratus.com>*
7117 * Extensive assembler packs updates, most notably:
7119 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7120 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7121 - x86_64: bit-sliced AES implementation;
7122 - ARM: NEON support, contemporary platforms optimizations;
7123 - s390x: z196 support;
7124 - `*`: GHASH and GF(2^m) multiplication implementations;
7128 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7129 (removal of unnecessary code)
7131 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7133 * Add TLS key material exporter from RFC 5705.
7137 * Add DTLS-SRTP negotiation from RFC 5764.
7141 * Add Next Protocol Negotiation,
7142 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7143 disabled with a no-npn flag to config or Configure. Code donated
7146 *Adam Langley <agl@google.com> and Ben Laurie*
7148 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7149 NIST-P256, NIST-P521, with constant-time single point multiplication on
7150 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7151 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7152 Code made available under Apache License version 2.0.
7154 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7155 line to include this in your build of OpenSSL, and run "make depend" (or
7156 "make update"). This enables the following EC_METHODs:
7158 EC_GFp_nistp224_method()
7159 EC_GFp_nistp256_method()
7160 EC_GFp_nistp521_method()
7162 EC_GROUP_new_by_curve_name() will automatically use these (while
7163 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7166 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7168 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7169 all platforms. Move ssize_t definition from e_os.h to the public
7170 header file e_os2.h as it now appears in public header file cms.h
7174 * New -sigopt option to the ca, req and x509 utilities. Additional
7175 signature parameters can be passed using this option and in
7180 * Add RSA PSS signing function. This will generate and set the
7181 appropriate AlgorithmIdentifiers for PSS based on those in the
7182 corresponding EVP_MD_CTX structure. No application support yet.
7186 * Support for companion algorithm specific ASN1 signing routines.
7187 New function ASN1_item_sign_ctx() signs a pre-initialised
7188 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7189 the appropriate parameters.
7193 * Add new algorithm specific ASN1 verification initialisation function
7194 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7195 handling will be the same no matter what EVP_PKEY_METHOD is used.
7196 Add a PSS handler to support verification of PSS signatures: checked
7197 against a number of sample certificates.
7201 * Add signature printing for PSS. Add PSS OIDs.
7203 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7205 * Add algorithm specific signature printing. An individual ASN1 method
7206 can now print out signatures instead of the standard hex dump.
7208 More complex signatures (e.g. PSS) can print out more meaningful
7209 information. Include DSA version that prints out the signature
7214 * Password based recipient info support for CMS library: implementing
7219 * Split password based encryption into PBES2 and PBKDF2 functions. This
7220 neatly separates the code into cipher and PBE sections and is required
7221 for some algorithms that split PBES2 into separate pieces (such as
7222 password based CMS).
7226 * Session-handling fixes:
7227 - Fix handling of connections that are resuming with a session ID,
7228 but also support Session Tickets.
7229 - Fix a bug that suppressed issuing of a new ticket if the client
7230 presented a ticket with an expired session.
7231 - Try to set the ticket lifetime hint to something reasonable.
7232 - Make tickets shorter by excluding irrelevant information.
7233 - On the client side, don't ignore renewed tickets.
7235 *Adam Langley, Bodo Moeller (Google)*
7237 * Fix PSK session representation.
7241 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7243 This work was sponsored by Intel.
7247 * Add GCM support to TLS library. Some custom code is needed to split
7248 the IV between the fixed (from PRF) and explicit (from TLS record)
7249 portions. This adds all GCM ciphersuites supported by RFC5288 and
7250 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7251 add a special AESGCM string for GCM only.
7255 * Expand range of ctrls for AES GCM. Permit setting invocation
7256 field on decrypt and retrieval of invocation field only on encrypt.
7260 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7261 As required by RFC5289 these ciphersuites cannot be used if for
7262 versions of TLS earlier than 1.2.
7266 * For FIPS capable OpenSSL interpret a NULL default public key method
7267 as unset and return the appropriate default but do *not* set the default.
7268 This means we can return the appropriate method in applications that
7269 switch between FIPS and non-FIPS modes.
7273 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7274 ENGINE is used then we cannot handle that in the FIPS module so we
7275 keep original code iff non-FIPS operations are allowed.
7279 * Add -attime option to openssl utilities.
7281 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7283 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7287 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7288 FIPS EC methods unconditionally for now.
7292 * New build option no-ec2m to disable characteristic 2 code.
7296 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7297 all cases can be covered as some introduce binary incompatibilities.
7301 * Redirect RSA operations to FIPS module including keygen,
7302 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7306 * Add similar low level API blocking to ciphers.
7310 * Low level digest APIs are not approved in FIPS mode: any attempt
7311 to use these will cause a fatal error. Applications that *really* want
7312 to use them can use the `private_*` version instead.
7316 * Redirect cipher operations to FIPS module for FIPS builds.
7320 * Redirect digest operations to FIPS module for FIPS builds.
7324 * Update build system to add "fips" flag which will link in fipscanister.o
7325 for static and shared library builds embedding a signature if needed.
7329 * Output TLS supported curves in preference order instead of numerical
7330 order. This is currently hardcoded for the highest order curves first.
7331 This should be configurable so applications can judge speed vs strength.
7335 * Add TLS v1.2 server support for client authentication.
7339 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7344 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7345 FIPS modules versions.
7349 * Add TLS v1.2 client side support for client authentication. Keep cache
7350 of handshake records longer as we don't know the hash algorithm to use
7351 until after the certificate request message is received.
7355 * Initial TLS v1.2 client support. Add a default signature algorithms
7356 extension including all the algorithms we support. Parse new signature
7357 format in client key exchange. Relax some ECC signing restrictions for
7358 TLS v1.2 as indicated in RFC5246.
7362 * Add server support for TLS v1.2 signature algorithms extension. Switch
7363 to new signature format when needed using client digest preference.
7364 All server ciphersuites should now work correctly in TLS v1.2. No client
7365 support yet and no support for client certificates.
7369 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7370 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7371 ciphersuites. At present only RSA key exchange ciphersuites work with
7372 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7373 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7374 and version checking.
7378 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7379 with this defined it will not be affected by any changes to ssl internal
7380 structures. Add several utility functions to allow openssl application
7381 to work with OPENSSL_NO_SSL_INTERN defined.
7385 * A long standing patch to add support for SRP from EdelWeb (Peter
7386 Sylvester and Christophe Renou) was integrated.
7387 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7388 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7391 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7395 * Permit abbreviated handshakes when renegotiating using the function
7396 SSL_renegotiate_abbreviated().
7398 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7400 * Add call to ENGINE_register_all_complete() to
7401 ENGINE_load_builtin_engines(), so some implementations get used
7402 automatically instead of needing explicit application support.
7406 * Add support for TLS key exporter as described in RFC5705.
7408 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7410 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7411 a few changes are required:
7413 Add SSL_OP_NO_TLSv1_1 flag.
7414 Add TLSv1_1 methods.
7415 Update version checking logic to handle version 1.1.
7416 Add explicit IV handling (ported from DTLS code).
7417 Add command line options to s_client/s_server.
7424 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7426 * X509_ATTRIBUTE memory leak
7428 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7429 memory. This structure is used by the PKCS#7 and CMS routines so any
7430 application which reads PKCS#7 or CMS data from untrusted sources is
7431 affected. SSL/TLS is not affected.
7433 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7439 * Race condition handling PSK identify hint
7441 If PSK identity hints are received by a multi-threaded client then
7442 the values are wrongly updated in the parent SSL_CTX structure. This can
7443 result in a race condition potentially leading to a double free of the
7449 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7451 * Malformed ECParameters causes infinite loop
7453 When processing an ECParameters structure OpenSSL enters an infinite loop
7454 if the curve specified is over a specially malformed binary polynomial
7457 This can be used to perform denial of service against any
7458 system which processes public keys, certificate requests or
7459 certificates. This includes TLS clients and TLS servers with
7460 client authentication enabled.
7462 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7467 * Exploitable out-of-bounds read in X509_cmp_time
7469 X509_cmp_time does not properly check the length of the ASN1_TIME
7470 string and can read a few bytes out of bounds. In addition,
7471 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7474 An attacker can use this to craft malformed certificates and CRLs of
7475 various sizes and potentially cause a segmentation fault, resulting in
7476 a DoS on applications that verify certificates or CRLs. TLS clients
7477 that verify CRLs are affected. TLS clients and servers with client
7478 authentication enabled may be affected if they use custom verification
7481 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7482 independently by Hanno Böck.
7487 * PKCS7 crash with missing EnvelopedContent
7489 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7490 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7491 with missing content and trigger a NULL pointer dereference on parsing.
7493 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7494 structures from untrusted sources are affected. OpenSSL clients and
7495 servers are not affected.
7497 This issue was reported to OpenSSL by Michal Zalewski (Google).
7502 * CMS verify infinite loop with unknown hash function
7504 When verifying a signedData message the CMS code can enter an infinite loop
7505 if presented with an unknown hash function OID. This can be used to perform
7506 denial of service against any system which verifies signedData messages using
7508 This issue was reported to OpenSSL by Johannes Bauer.
7513 * Race condition handling NewSessionTicket
7515 If a NewSessionTicket is received by a multi-threaded client when attempting to
7516 reuse a previous ticket then a race condition can occur potentially leading to
7517 a double free of the ticket data.
7522 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7524 * Segmentation fault in ASN1_TYPE_cmp fix
7526 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7527 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7528 certificate signature algorithm consistency this can be used to crash any
7529 certificate verification operation and exploited in a DoS attack. Any
7530 application which performs certificate verification is vulnerable including
7531 OpenSSL clients and servers which enable client authentication.
7536 * ASN.1 structure reuse memory corruption fix
7538 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7539 memory corruption via an invalid write. Such reuse is and has been
7540 strongly discouraged and is believed to be rare.
7542 Applications that parse structures containing CHOICE or ANY DEFINED BY
7543 components may be affected. Certificate parsing (d2i_X509 and related
7544 functions) are however not affected. OpenSSL clients and servers are
7550 * PKCS7 NULL pointer dereferences fix
7552 The PKCS#7 parsing code does not handle missing outer ContentInfo
7553 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7554 missing content and trigger a NULL pointer dereference on parsing.
7556 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7557 otherwise parse PKCS#7 structures from untrusted sources are
7558 affected. OpenSSL clients and servers are not affected.
7560 This issue was reported to OpenSSL by Michal Zalewski (Google).
7565 * DoS via reachable assert in SSLv2 servers fix
7567 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7568 servers that both support SSLv2 and enable export cipher suites by sending
7569 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7571 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7572 (OpenSSL development team).
7577 * Use After Free following d2i_ECPrivatekey error fix
7579 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7580 could cause a use after free condition. This, in turn, could cause a double
7581 free in several private key parsing functions (such as d2i_PrivateKey
7582 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7583 for applications that receive EC private keys from untrusted
7584 sources. This scenario is considered rare.
7586 This issue was discovered by the BoringSSL project and fixed in their
7592 * X509_to_X509_REQ NULL pointer deref fix
7594 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7595 the certificate key is invalid. This function is rarely used in practice.
7597 This issue was discovered by Brian Carpenter.
7602 * Removed the export ciphers from the DEFAULT ciphers
7606 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7608 * Build fixes for the Windows and OpenVMS platforms
7610 *Matt Caswell and Richard Levitte*
7612 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7614 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7615 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7616 dereference. This could lead to a Denial Of Service attack. Thanks to
7617 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7622 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7623 dtls1_buffer_record function under certain conditions. In particular this
7624 could occur if an attacker sent repeated DTLS records with the same
7625 sequence number but for the next epoch. The memory leak could be exploited
7626 by an attacker in a Denial of Service attack through memory exhaustion.
7627 Thanks to Chris Mueller for reporting this issue.
7632 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7633 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7634 method would be set to NULL which could later result in a NULL pointer
7635 dereference. Thanks to Frank Schmirler for reporting this issue.
7640 * Abort handshake if server key exchange message is omitted for ephemeral
7643 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7644 reporting this issue.
7649 * Remove non-export ephemeral RSA code on client and server. This code
7650 violated the TLS standard by allowing the use of temporary RSA keys in
7651 non-export ciphersuites and could be used by a server to effectively
7652 downgrade the RSA key length used to a value smaller than the server
7653 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7654 INRIA or reporting this issue.
7659 * Fixed issue where DH client certificates are accepted without verification.
7660 An OpenSSL server will accept a DH certificate for client authentication
7661 without the certificate verify message. This effectively allows a client to
7662 authenticate without the use of a private key. This only affects servers
7663 which trust a client certificate authority which issues certificates
7664 containing DH keys: these are extremely rare and hardly ever encountered.
7665 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7671 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7672 results on some platforms, including x86_64. This bug occurs at random
7673 with a very low probability, and is not known to be exploitable in any
7674 way, though its exact impact is difficult to determine. Thanks to Pieter
7675 Wuille (Blockstream) who reported this issue and also suggested an initial
7676 fix. Further analysis was conducted by the OpenSSL development team and
7677 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7678 the OpenSSL core team.
7683 * Fix various certificate fingerprint issues.
7685 By using non-DER or invalid encodings outside the signed portion of a
7686 certificate the fingerprint can be changed without breaking the signature.
7687 Although no details of the signed portion of the certificate can be changed
7688 this can cause problems with some applications: e.g. those using the
7689 certificate fingerprint for blacklists.
7691 1. Reject signatures with non zero unused bits.
7693 If the BIT STRING containing the signature has non zero unused bits reject
7694 the signature. All current signature algorithms require zero unused bits.
7696 2. Check certificate algorithm consistency.
7698 Check the AlgorithmIdentifier inside TBS matches the one in the
7699 certificate signature. NB: this will result in signature failure
7700 errors for some broken certificates.
7702 Thanks to Konrad Kraszewski from Google for reporting this issue.
7704 3. Check DSA/ECDSA signatures use DER.
7706 Reencode DSA/ECDSA signatures and compare with the original received
7707 signature. Return an error if there is a mismatch.
7709 This will reject various cases including garbage after signature
7710 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7711 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7712 (negative or with leading zeroes).
7714 Further analysis was conducted and fixes were developed by Stephen Henson
7715 of the OpenSSL core team.
7721 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7723 * Session Ticket Memory Leak.
7725 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7726 integrity of that ticket is first verified. In the event of a session
7727 ticket integrity check failing, OpenSSL will fail to free memory
7728 causing a memory leak. By sending a large number of invalid session
7729 tickets an attacker could exploit this issue in a Denial Of Service
7735 * Build option no-ssl3 is incomplete.
7737 When OpenSSL is configured with "no-ssl3" as a build option, servers
7738 could accept and complete a SSL 3.0 handshake, and clients could be
7739 configured to send them.
7742 *Akamai and the OpenSSL team*
7744 * Add support for TLS_FALLBACK_SCSV.
7745 Client applications doing fallback retries should call
7746 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7749 *Adam Langley, Bodo Moeller*
7751 * Add additional DigestInfo checks.
7753 Reencode DigestInto in DER and check against the original when
7754 verifying RSA signature: this will reject any improperly encoded
7755 DigestInfo structures.
7757 Note: this is a precautionary measure and no attacks are currently known.
7761 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7763 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7764 to a denial of service attack. A malicious server can crash the client
7765 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7766 ciphersuite and sending carefully crafted handshake messages.
7768 Thanks to Felix Gröbert (Google) for discovering and researching this
7774 * By sending carefully crafted DTLS packets an attacker could cause openssl
7775 to leak memory. This can be exploited through a Denial of Service attack.
7776 Thanks to Adam Langley for discovering and researching this issue.
7781 * An attacker can force openssl to consume large amounts of memory whilst
7782 processing DTLS handshake messages. This can be exploited through a
7783 Denial of Service attack.
7784 Thanks to Adam Langley for discovering and researching this issue.
7789 * An attacker can force an error condition which causes openssl to crash
7790 whilst processing DTLS packets due to memory being freed twice. This
7791 can be exploited through a Denial of Service attack.
7792 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7798 * If a multithreaded client connects to a malicious server using a resumed
7799 session and the server sends an ec point format extension it could write
7800 up to 255 bytes to freed memory.
7802 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7808 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7809 X509_name_oneline, X509_name_print_ex et al. to leak some information
7810 from the stack. Applications may be affected if they echo pretty printing
7811 output to the attacker.
7813 Thanks to Ivan Fratric (Google) for discovering this issue.
7816 *Emilia Käsper, and Steve Henson*
7818 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7819 for corner cases. (Certain input points at infinity could lead to
7820 bogus results, with non-infinity inputs mapped to infinity too.)
7824 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7826 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7827 handshake can force the use of weak keying material in OpenSSL
7828 SSL/TLS clients and servers.
7830 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7831 researching this issue. ([CVE-2014-0224])
7833 *KIKUCHI Masashi, Steve Henson*
7835 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7836 OpenSSL DTLS client the code can be made to recurse eventually crashing
7839 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7842 *Imre Rad, Steve Henson*
7844 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7845 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7846 client or server. This is potentially exploitable to run arbitrary
7847 code on a vulnerable client or server.
7849 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7851 *Jüri Aedla, Steve Henson*
7853 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7854 are subject to a denial of service attack.
7856 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7857 this issue. ([CVE-2014-3470])
7859 *Felix Gröbert, Ivan Fratric, Steve Henson*
7861 * Harmonize version and its documentation. -f flag is used to display
7864 *mancha <mancha1@zoho.com>*
7866 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7867 in i2d_ECPrivateKey.
7869 *mancha <mancha1@zoho.com>*
7871 * Fix some double frees. These are not thought to be exploitable.
7873 *mancha <mancha1@zoho.com>*
7875 * Fix for the attack described in the paper "Recovering OpenSSL
7876 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7877 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7878 <http://eprint.iacr.org/2014/140>
7880 Thanks to Yuval Yarom and Naomi Benger for discovering this
7881 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7883 *Yuval Yarom and Naomi Benger*
7885 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7887 * Keep original DTLS digest and encryption contexts in retransmission
7888 structures so we can use the previous session parameters if they need
7889 to be resent. ([CVE-2013-6450])
7893 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7894 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7895 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7896 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7897 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7898 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7900 *Rob Stradling, Adam Langley*
7902 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7904 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7906 This addresses the flaw in CBC record processing discovered by
7907 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7908 at: <http://www.isg.rhul.ac.uk/tls/>
7910 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7911 Security Group at Royal Holloway, University of London
7912 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7913 Emilia Käsper for the initial patch.
7916 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7918 * Return an error when checking OCSP signatures when key is NULL.
7919 This fixes a DoS attack. ([CVE-2013-0166])
7923 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7924 the right response is stapled. Also change SSL_get_certificate()
7925 so it returns the certificate actually sent.
7926 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7927 (This is a backport)
7929 *Rob Stradling <rob.stradling@comodo.com>*
7931 * Fix possible deadlock when decoding public keys.
7935 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7937 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7940 * Sanity check record length before skipping explicit IV in DTLS
7943 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7944 fuzzing as a service testing platform.
7949 * Initialise tkeylen properly when encrypting CMS messages.
7950 Thanks to Solar Designer of Openwall for reporting this issue.
7954 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7956 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7957 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7958 in CRYPTO_realloc_clean.
7960 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7961 issue and to Adam Langley <agl@chromium.org> for fixing it.
7964 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7966 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7968 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7969 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7970 content decryption and always return the same error. Note: this attack
7971 needs on average 2^20 messages so it only affects automated senders. The
7972 old behaviour can be re-enabled in the CMS code by setting the
7973 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7974 an MMA defence is not necessary.
7975 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7976 this issue. ([CVE-2012-0884])
7980 * Fix CVE-2011-4619: make sure we really are receiving a
7981 client hello before rejecting multiple SGC restarts. Thanks to
7982 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7986 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7988 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7989 Thanks to Antonio Martin, Enterprise Secure Access Research and
7990 Development, Cisco Systems, Inc. for discovering this bug and
7991 preparing a fix. ([CVE-2012-0050])
7995 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7997 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7998 of the Vaudenay padding oracle attack on CBC mode encryption
7999 which enables an efficient plaintext recovery attack against
8000 the OpenSSL implementation of DTLS. Their attack exploits timing
8001 differences arising during decryption processing. A research
8002 paper describing this attack can be found at:
8003 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8004 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8005 Security Group at Royal Holloway, University of London
8006 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8007 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8008 for preparing the fix. ([CVE-2011-4108])
8010 *Robin Seggelmann, Michael Tuexen*
8012 * Clear bytes used for block padding of SSL 3.0 records.
8015 *Adam Langley (Google)*
8017 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8018 Kadianakis <desnacked@gmail.com> for discovering this issue and
8019 Adam Langley for preparing the fix. ([CVE-2011-4619])
8021 *Adam Langley (Google)*
8023 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8025 *Andrey Kulikov <amdeich@gmail.com>*
8027 * Prevent malformed RFC3779 data triggering an assertion failure.
8028 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8029 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8031 *Rob Austein <sra@hactrn.net>*
8033 * Improved PRNG seeding for VOS.
8035 *Paul Green <Paul.Green@stratus.com>*
8037 * Fix ssl_ciph.c set-up race.
8039 *Adam Langley (Google)*
8041 * Fix spurious failures in ecdsatest.c.
8043 *Emilia Käsper (Google)*
8045 * Fix the BIO_f_buffer() implementation (which was mixing different
8046 interpretations of the `..._len` fields).
8048 *Adam Langley (Google)*
8050 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8051 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8052 threads won't reuse the same blinding coefficients.
8054 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8055 lock to call BN_BLINDING_invert_ex, and avoids one use of
8056 BN_BLINDING_update for each BN_BLINDING structure (previously,
8057 the last update always remained unused).
8059 *Emilia Käsper (Google)*
8061 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8063 *Bob Buckholz (Google)*
8065 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8067 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8068 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8070 *Kaspar Brand <ossl@velox.ch>*
8072 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8073 for multi-threaded use of ECDH. ([CVE-2011-3210])
8075 *Adam Langley (Google)*
8077 * Fix x509_name_ex_d2i memory leak on bad inputs.
8081 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8082 signature public key algorithm by using OID xref utilities instead.
8083 Before this you could only use some ECC ciphersuites with SHA1 only.
8087 * Add protection against ECDSA timing attacks as mentioned in the paper
8088 by Billy Bob Brumley and Nicola Tuveri, see:
8089 <http://eprint.iacr.org/2011/232.pdf>
8091 *Billy Bob Brumley and Nicola Tuveri*
8093 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8095 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8097 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8099 * Fix bug in string printing code: if *any* escaping is enabled we must
8100 escape the escape character (backslash) or the resulting string is
8105 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8107 * Disable code workaround for ancient and obsolete Netscape browsers
8108 and servers: an attacker can use it in a ciphersuite downgrade attack.
8109 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8113 * Fixed J-PAKE implementation error, originally discovered by
8114 Sebastien Martini, further info and confirmation from Stefan
8115 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8119 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8121 * Fix extension code to avoid race conditions which can result in a buffer
8122 overrun vulnerability: resumed sessions must not be modified as they can
8123 be shared by multiple threads. CVE-2010-3864
8127 * Fix WIN32 build system to correctly link an ENGINE directory into
8132 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8134 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8137 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8139 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8141 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8142 context. The operation can be customised via the ctrl mechanism in
8143 case ENGINEs want to include additional functionality.
8147 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8151 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8152 output hashes compatible with older versions of OpenSSL.
8154 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8156 * Fix compression algorithm handling: if resuming a session use the
8157 compression algorithm of the resumed session instead of determining
8158 it from client hello again. Don't allow server to change algorithm.
8162 * Add load_crls() function to commands tidying load_certs() too. Add option
8163 to verify utility to allow additional CRLs to be included.
8167 * Update OCSP request code to permit adding custom headers to the request:
8168 some responders need this.
8172 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8175 *Julia Lawall <julia@diku.dk>*
8177 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8178 needlessly dereferenced structures, used obsolete functions and
8179 didn't handle all updated verify codes correctly.
8183 * Disable MD2 in the default configuration.
8187 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8188 indicate the initial BIO being pushed or popped. This makes it possible
8189 to determine whether the BIO is the one explicitly called or as a result
8190 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8191 it handles reference counts correctly and doesn't zero out the I/O bio
8192 when it is not being explicitly popped. WARNING: applications which
8193 included workarounds for the old buggy behaviour will need to be modified
8194 or they could free up already freed BIOs.
8198 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8199 renaming to all platforms (within the 0.9.8 branch, this was
8200 done conditionally on Netware platforms to avoid a name clash).
8202 *Guenter <lists@gknw.net>*
8204 * Add ECDHE and PSK support to DTLS.
8206 *Michael Tuexen <tuexen@fh-muenster.de>*
8208 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8213 * Add "missing" function EVP_MD_flags() (without this the only way to
8214 retrieve a digest flags is by accessing the structure directly. Update
8215 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8216 or cipher is registered as in the "from" argument. Print out all
8217 registered digests in the dgst usage message instead of manually
8218 attempting to work them out.
8222 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8223 this allows the use of compression and extensions. Change default cipher
8224 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8225 by default unless an application cipher string requests it.
8229 * Alter match criteria in PKCS12_parse(). It used to try to use local
8230 key ids to find matching certificates and keys but some PKCS#12 files
8231 don't follow the (somewhat unwritten) rules and this strategy fails.
8232 Now just gather all certificates together and the first private key
8233 then look for the first certificate that matches the key.
8237 * Support use of registered digest and cipher names for dgst and cipher
8238 commands instead of having to add each one as a special case. So now
8245 openssl dgst -sha256 foo
8247 and this works for ENGINE based algorithms too.
8251 * Update Gost ENGINE to support parameter files.
8253 *Victor B. Wagner <vitus@cryptocom.ru>*
8255 * Support GeneralizedTime in ca utility.
8257 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8259 * Enhance the hash format used for certificate directory links. The new
8260 form uses the canonical encoding (meaning equivalent names will work
8261 even if they aren't identical) and uses SHA1 instead of MD5. This form
8262 is incompatible with the older format and as a result c_rehash should
8263 be used to rebuild symbolic links.
8267 * Make PKCS#8 the default write format for private keys, replacing the
8268 traditional format. This form is standardised, more secure and doesn't
8269 include an implicit MD5 dependency.
8273 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8274 committed to OpenSSL should pass this lot as a minimum.
8278 * Add session ticket override functionality for use by EAP-FAST.
8280 *Jouni Malinen <j@w1.fi>*
8282 * Modify HMAC functions to return a value. Since these can be implemented
8283 in an ENGINE errors can occur.
8287 * Type-checked OBJ_bsearch_ex.
8291 * Type-checked OBJ_bsearch. Also some constification necessitated
8292 by type-checking. Still to come: TXT_DB, bsearch(?),
8293 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8298 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8299 seconds to a tm structure directly, instead of going through OS
8300 specific date routines. This avoids any issues with OS routines such
8301 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8302 and X509_time_adj_ex() to cover the extended range. The existing
8303 X509_time_adj() is still usable and will no longer have any date issues.
8307 * Delta CRL support. New use deltas option which will attempt to locate
8308 and search any appropriate delta CRLs available.
8310 This work was sponsored by Google.
8314 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8315 code and add additional score elements. Validate alternate CRL paths
8316 as part of the CRL checking and indicate a new error "CRL path validation
8317 error" in this case. Applications wanting additional details can use
8318 the verify callback and check the new "parent" field. If this is not
8319 NULL CRL path validation is taking place. Existing applications won't
8320 see this because it requires extended CRL support which is off by
8323 This work was sponsored by Google.
8327 * Support for freshest CRL extension.
8329 This work was sponsored by Google.
8333 * Initial indirect CRL support. Currently only supported in the CRLs
8334 passed directly and not via lookup. Process certificate issuer
8335 CRL entry extension and lookup CRL entries by bother issuer name
8336 and serial number. Check and process CRL issuer entry in IDP extension.
8338 This work was sponsored by Google.
8342 * Add support for distinct certificate and CRL paths. The CRL issuer
8343 certificate is validated separately in this case. Only enabled if
8344 an extended CRL support flag is set: this flag will enable additional
8345 CRL functionality in future.
8347 This work was sponsored by Google.
8351 * Add support for policy mappings extension.
8353 This work was sponsored by Google.
8357 * Fixes to pathlength constraint, self issued certificate handling,
8358 policy processing to align with RFC3280 and PKITS tests.
8360 This work was sponsored by Google.
8364 * Support for name constraints certificate extension. DN, email, DNS
8365 and URI types are currently supported.
8367 This work was sponsored by Google.
8371 * To cater for systems that provide a pointer-based thread ID rather
8372 than numeric, deprecate the current numeric thread ID mechanism and
8373 replace it with a structure and associated callback type. This
8374 mechanism allows a numeric "hash" to be extracted from a thread ID in
8375 either case, and on platforms where pointers are larger than 'long',
8376 mixing is done to help ensure the numeric 'hash' is usable even if it
8377 can't be guaranteed unique. The default mechanism is to use "&errno"
8378 as a pointer-based thread ID to distinguish between threads.
8380 Applications that want to provide their own thread IDs should now use
8381 CRYPTO_THREADID_set_callback() to register a callback that will call
8382 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8384 Note that ERR_remove_state() is now deprecated, because it is tied
8385 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8386 to free the current thread's error state should be replaced by
8387 ERR_remove_thread_state(NULL).
8389 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8390 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8391 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8392 application was previously providing a numeric thread callback that
8393 was inappropriate for distinguishing threads, then uniqueness might
8394 have been obtained with &errno that happened immediately in the
8395 intermediate development versions of OpenSSL; this is no longer the
8396 case, the numeric thread callback will now override the automatic use
8399 *Geoff Thorpe, with help from Bodo Moeller*
8401 * Initial support for different CRL issuing certificates. This covers a
8402 simple case where the self issued certificates in the chain exist and
8403 the real CRL issuer is higher in the existing chain.
8405 This work was sponsored by Google.
8409 * Removed effectively defunct crypto/store from the build.
8413 * Revamp of STACK to provide stronger type-checking. Still to come:
8414 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8415 ASN1_STRING, CONF_VALUE.
8419 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8420 RAM on SSL connections. This option can save about 34k per idle SSL.
8424 * Revamp of LHASH to provide stronger type-checking. Still to come:
8425 STACK, TXT_DB, bsearch, qsort.
8429 * Initial support for Cryptographic Message Syntax (aka CMS) based
8430 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8431 support for data, signedData, compressedData, digestedData and
8432 encryptedData, envelopedData types included. Scripts to check against
8433 RFC4134 examples draft and interop and consistency checks of many
8434 content types and variants.
8438 * Add options to enc utility to support use of zlib compression BIO.
8442 * Extend mk1mf to support importing of options and assembly language
8443 files from Configure script, currently only included in VC-WIN32.
8444 The assembly language rules can now optionally generate the source
8445 files from the associated perl scripts.
8449 * Implement remaining functionality needed to support GOST ciphersuites.
8450 Interop testing has been performed using CryptoPro implementations.
8452 *Victor B. Wagner <vitus@cryptocom.ru>*
8454 * s390x assembler pack.
8458 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8463 * Implement Opaque PRF Input TLS extension as specified in
8464 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8465 official specification yet and no extension type assignment by
8466 IANA exists, this extension (for now) will have to be explicitly
8467 enabled when building OpenSSL by providing the extension number
8468 to use. For example, specify an option
8470 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8472 to the "config" or "Configure" script to enable the extension,
8473 assuming extension number 0x9527 (which is a completely arbitrary
8474 and unofficial assignment based on the MD5 hash of the Internet
8475 Draft). Note that by doing so, you potentially lose
8476 interoperability with other TLS implementations since these might
8477 be using the same extension number for other purposes.
8479 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8480 opaque PRF input value to use in the handshake. This will create
8481 an internal copy of the length-'len' string at 'src', and will
8482 return non-zero for success.
8484 To get more control and flexibility, provide a callback function
8487 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8488 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8492 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8495 Callback function 'cb' will be called in handshakes, and is
8496 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8497 Argument 'arg' is for application purposes (the value as given to
8498 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8499 be provided to the callback function). The callback function
8500 has to return non-zero to report success: usually 1 to use opaque
8501 PRF input just if possible, or 2 to enforce use of the opaque PRF
8502 input. In the latter case, the library will abort the handshake
8503 if opaque PRF input is not successfully negotiated.
8505 Arguments 'peerinput' and 'len' given to the callback function
8506 will always be NULL and 0 in the case of a client. A server will
8507 see the client's opaque PRF input through these variables if
8508 available (NULL and 0 otherwise). Note that if the server
8509 provides an opaque PRF input, the length must be the same as the
8510 length of the client's opaque PRF input.
8512 Note that the callback function will only be called when creating
8513 a new session (session resumption can resume whatever was
8514 previously negotiated), and will not be called in SSL 2.0
8515 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8516 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8517 for applications that need to enforce opaque PRF input.
8521 * Update ssl code to support digests other than SHA1+MD5 for handshake
8524 *Victor B. Wagner <vitus@cryptocom.ru>*
8526 * Add RFC4507 support to OpenSSL. This includes the corrections in
8527 RFC4507bis. The encrypted ticket format is an encrypted encoded
8528 SSL_SESSION structure, that way new session features are automatically
8531 If a client application caches session in an SSL_SESSION structure
8532 support is transparent because tickets are now stored in the encoded
8535 The SSL_CTX structure automatically generates keys for ticket
8536 protection in servers so again support should be possible
8537 with no application modification.
8539 If a client or server wishes to disable RFC4507 support then the option
8540 SSL_OP_NO_TICKET can be set.
8542 Add a TLS extension debugging callback to allow the contents of any client
8543 or server extensions to be examined.
8545 This work was sponsored by Google.
8549 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8550 OpenSSL should now compile cleanly on gcc 4.2
8552 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8554 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8555 support including streaming MAC support: this is required for GOST
8556 ciphersuite support.
8558 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8560 * Add option -stream to use PKCS#7 streaming in smime utility. New
8561 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8562 to output in BER and PEM format.
8566 * Experimental support for use of HMAC via EVP_PKEY interface. This
8567 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8568 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8569 ENGINE support for HMAC keys which are unextractable. New -mac and
8570 -macopt options to dgst utility.
8574 * New option -sigopt to dgst utility. Update dgst to use
8575 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8576 alternative signing parameters such as X9.31 or PSS in the dgst
8581 * Change ssl_cipher_apply_rule(), the internal function that does
8582 the work each time a ciphersuite string requests enabling
8583 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8584 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8585 the order of disabled ciphersuites such that those ciphersuites
8586 that most recently went from enabled to disabled not only stay
8587 in order with respect to each other, but also have higher priority
8588 than other disabled ciphersuites the next time ciphersuites are
8591 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8592 the same ciphersuites as with "HIGH" alone, but in a specific
8593 order where the PSK ciphersuites come first (since they are the
8594 most recently disabled ciphersuites when "HIGH" is parsed).
8596 Also, change ssl_create_cipher_list() (using this new
8597 functionality) such that between otherwise identical
8598 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8603 * Change ssl_create_cipher_list() so that it automatically
8604 arranges the ciphersuites in reasonable order before starting
8605 to process the rule string. Thus, the definition for "DEFAULT"
8606 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8607 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8608 This makes it much easier to arrive at a reasonable default order
8609 in applications for which anonymous ciphers are OK (meaning
8610 that you can't actually use DEFAULT).
8612 *Bodo Moeller; suggested by Victor Duchovni*
8614 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8615 processing) into multiple integers instead of setting
8616 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8617 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8618 (These masks as well as the individual bit definitions are hidden
8619 away into the non-exported interface ssl/ssl_locl.h, so this
8620 change to the definition of the SSL_CIPHER structure shouldn't
8621 affect applications.) This give us more bits for each of these
8622 categories, so there is no longer a need to coagulate AES128 and
8623 AES256 into a single algorithm bit, and to coagulate Camellia128
8624 and Camellia256 into a single algorithm bit, which has led to all
8627 Thus, among other things, the kludge introduced in 0.9.7m and
8628 0.9.8e for masking out AES256 independently of AES128 or masking
8629 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8631 With the change, we also introduce new ciphersuite aliases that
8632 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8637 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8638 Use the leftmost N bytes of the signature input if the input is
8639 larger than the prime q (with N being the size in bytes of q).
8643 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8644 it yet and it is largely untested.
8648 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8652 * Initial incomplete changes to avoid need for function casts in OpenSSL
8653 some compilers (gcc 4.2 and later) reject their use. Safestack is
8654 reimplemented. Update ASN1 to avoid use of legacy functions.
8658 * Win32/64 targets are linked with Winsock2.
8662 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8663 to external functions. This can be used to increase CRL handling
8664 efficiency especially when CRLs are very large by (for example) storing
8665 the CRL revoked certificates in a database.
8669 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8670 new CRLs added to a directory can be used. New command line option
8671 -verify_return_error to s_client and s_server. This causes real errors
8672 to be returned by the verify callback instead of carrying on no matter
8673 what. This reflects the way a "real world" verify callback would behave.
8677 * GOST engine, supporting several GOST algorithms and public key formats.
8678 Kindly donated by Cryptocom.
8682 * Partial support for Issuing Distribution Point CRL extension. CRLs
8683 partitioned by DP are handled but no indirect CRL or reason partitioning
8684 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8685 selected via a scoring technique which handles IDP and AKID in CRLs.
8689 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8690 will ultimately be used for all verify operations: this will remove the
8691 X509_STORE dependency on certificate verification and allow alternative
8692 lookup methods. X509_STORE based implementations of these two callbacks.
8696 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8697 Modify get_crl() to find a valid (unexpired) CRL if possible.
8701 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8702 this would be called X509_CRL_cmp() but that name is already used by
8703 a function that just compares CRL issuer names. Cache several CRL
8704 extensions in X509_CRL structure and cache CRLDP in X509.
8708 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8709 this maps equivalent X509_NAME structures into a consistent structure.
8710 Name comparison can then be performed rapidly using memcmp().
8714 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8719 * Allow digests to supply their own micalg string for S/MIME type using
8720 the ctrl EVP_MD_CTRL_MICALG.
8724 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8725 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8726 ctrl. It can then customise the structure before and/or after signing
8731 * New function OBJ_add_sigid() to allow application defined signature OIDs
8732 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8733 to free up any added signature OIDs.
8737 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8738 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8739 digest and cipher tables. New options added to openssl utility:
8740 list-message-digest-algorithms and list-cipher-algorithms.
8744 * Change the array representation of binary polynomials: the list
8745 of degrees of non-zero coefficients is now terminated with -1.
8746 Previously it was terminated with 0, which was also part of the
8747 value; thus, the array representation was not applicable to
8748 polynomials where t^0 has coefficient zero. This change makes
8749 the array representation useful in a more general context.
8753 * Various modifications and fixes to SSL/TLS cipher string
8754 handling. For ECC, the code now distinguishes between fixed ECDH
8755 with RSA certificates on the one hand and with ECDSA certificates
8756 on the other hand, since these are separate ciphersuites. The
8757 unused code for Fortezza ciphersuites has been removed.
8759 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8760 (not "ECDHE"). For consistency with the code for DH
8761 certificates, use of ECDH certificates is now considered ECDH
8762 authentication, not RSA or ECDSA authentication (the latter is
8763 merely the CA's signing algorithm and not actively used in the
8766 The temporary ciphersuite alias "ECCdraft" is no longer
8767 available, and ECC ciphersuites are no longer excluded from "ALL"
8768 and "DEFAULT". The following aliases now exist for RFC 4492
8769 ciphersuites, most of these by analogy with the DH case:
8771 kECDHr - ECDH cert, signed with RSA
8772 kECDHe - ECDH cert, signed with ECDSA
8773 kECDH - ECDH cert (signed with either RSA or ECDSA)
8774 kEECDH - ephemeral ECDH
8775 ECDH - ECDH cert or ephemeral ECDH
8781 AECDH - anonymous ECDH
8782 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8786 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8787 Use correct micalg parameters depending on digest(s) in signed message.
8791 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8792 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8796 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8797 an engine to register a method. Add ENGINE lookups for methods and
8798 functional reference processing.
8802 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8803 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8808 * New -resign option to smime utility. This adds one or more signers
8809 to an existing PKCS#7 signedData structure. Also -md option to use an
8810 alternative message digest algorithm for signing.
8814 * Tidy up PKCS#7 routines and add new functions to make it easier to
8815 create PKCS7 structures containing multiple signers. Update smime
8816 application to support multiple signers.
8820 * New -macalg option to pkcs12 utility to allow setting of an alternative
8825 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8826 Reorganize PBE internals to lookup from a static table using NIDs,
8827 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8828 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8829 PRF which will be automatically used with PBES2.
8833 * Replace the algorithm specific calls to generate keys in "req" with the
8838 * Update PKCS#7 enveloped data routines to use new API. This is now
8839 supported by any public key method supporting the encrypt operation. A
8840 ctrl is added to allow the public key algorithm to examine or modify
8841 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8846 * Add a ctrl to asn1 method to allow a public key algorithm to express
8847 a default digest type to use. In most cases this will be SHA1 but some
8848 algorithms (such as GOST) need to specify an alternative digest. The
8849 return value indicates how strong the preference is 1 means optional and
8850 2 is mandatory (that is it is the only supported type). Modify
8851 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8852 use the default md. Update openssl utilities to use the default digest
8853 type for signing if it is not explicitly indicated.
8857 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8858 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8859 signing method from the key type. This effectively removes the link
8860 between digests and public key types.
8864 * Add an OID cross reference table and utility functions. Its purpose is to
8865 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8866 rsaEncryption. This will allow some of the algorithm specific hackery
8867 needed to use the correct OID to be removed.
8871 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8872 structures for PKCS7_sign(). They are now set up by the relevant public
8877 * Add provisional EC pkey method with support for ECDSA and ECDH.
8881 * Add support for key derivation (agreement) in the API, DH method and
8886 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8887 public and private key formats. As a side effect these add additional
8888 command line functionality not previously available: DSA signatures can be
8889 generated and verified using pkeyutl and DH key support and generation in
8896 *Oliver Tappe <zooey@hirschkaefer.de>*
8898 * New make target "install_html_docs" installs HTML renditions of the
8901 *Oliver Tappe <zooey@hirschkaefer.de>*
8903 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8904 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8905 support key and parameter generation and add initial key generation
8906 functionality for RSA.
8910 * Add functions for main EVP_PKEY_method operations. The undocumented
8911 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8912 `EVP_PKEY_{encrypt,decrypt}_old`.
8916 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8917 key API, doesn't do much yet.
8921 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8922 public key algorithms. New option to openssl utility:
8923 "list-public-key-algorithms" to print out info.
8927 * Implement the Supported Elliptic Curves Extension for
8928 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8932 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8933 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8937 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8938 utilities such as rsa, dsa, dsaparam etc except they process any key
8943 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8944 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8945 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8950 * Initial support for pluggable public key ASN1.
8951 De-spaghettify the public key ASN1 handling. Move public and private
8952 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8953 algorithm specific handling to a single module within the relevant
8954 algorithm directory. Add functions to allow (near) opaque processing
8955 of public and private key structures.
8959 * Implement the Supported Point Formats Extension for
8960 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8964 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8965 for the psk identity [hint] and the psk callback functions to the
8966 SSL_SESSION, SSL and SSL_CTX structure.
8969 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8973 SSL_CTX_use_psk_identity_hint
8974 SSL_get_psk_identity_hint
8975 SSL_get_psk_identity
8976 SSL_use_psk_identity_hint
8978 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8980 * Add RFC 3161 compliant time stamp request creation, response generation
8981 and response verification functionality.
8983 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8985 * Add initial support for TLS extensions, specifically for the server_name
8986 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8987 have new members for a host name. The SSL data structure has an
8988 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8989 stored in that context to allow for session resumption, even after the
8990 SSL has been switched to a new SSL_CTX in reaction to a client's
8991 server_name extension.
8993 New functions (subject to change):
8995 SSL_get_servername()
8996 SSL_get_servername_type()
8999 New CTRL codes and macros (subject to change):
9001 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9002 - SSL_CTX_set_tlsext_servername_callback()
9003 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9004 - SSL_CTX_set_tlsext_servername_arg()
9005 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9007 openssl s_client has a new '-servername ...' option.
9009 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9010 '-key2 ...', '-servername_fatal' (subject to change). This allows
9011 testing the HostName extension for a specific single host name ('-cert'
9012 and '-key' remain fallbacks for handshakes without HostName
9013 negotiation). If the unrecognized_name alert has to be sent, this by
9014 default is a warning; it becomes fatal with the '-servername_fatal'
9017 *Peter Sylvester, Remy Allais, Christophe Renou*
9019 * Whirlpool hash implementation is added.
9023 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9024 bn(64,32). Because of instruction set limitations it doesn't have
9025 any negative impact on performance. This was done mostly in order
9026 to make it possible to share assembler modules, such as bn_mul_mont
9027 implementations, between 32- and 64-bit builds without hassle.
9031 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9032 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9037 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9038 dedicated Montgomery multiplication procedure, is introduced.
9039 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9040 "64-bit" performance on certain 32-bit targets.
9044 * New option SSL_OP_NO_COMP to disable use of compression selectively
9045 in SSL structures. New SSL ctrl to set maximum send fragment size.
9046 Save memory by setting the I/O buffer sizes dynamically instead of
9047 using the maximum available value.
9051 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9052 in addition to the text details.
9056 * Very, very preliminary EXPERIMENTAL support for printing of general
9057 ASN1 structures. This currently produces rather ugly output and doesn't
9058 handle several customised structures at all.
9062 * Integrated support for PVK file format and some related formats such
9063 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9064 these in the 'rsa' and 'dsa' utilities.
9068 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9072 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9073 place for the (very old) "NETSCAPE" format certificates which are now
9074 handled using new ASN1 code equivalents.
9078 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9079 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9080 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9084 * Modify CRL distribution points extension code to print out previously
9085 unsupported fields. Enhance extension setting code to allow setting of
9090 * Add print and set support for Issuing Distribution Point CRL extension.
9094 * Change 'Configure' script to enable Camellia by default.
9101 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9103 * When rejecting SSL/TLS records due to an incorrect version number, never
9104 update s->server with a new major version number. As of
9105 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9106 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9107 the previous behavior could result in a read attempt at NULL when
9108 receiving specific incorrect SSL/TLS records once record payload
9109 protection is active. ([CVE-2010-0740])
9111 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9113 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9114 could be crashed if the relevant tables were not present (e.g. chrooted).
9116 *Tomas Hoger <thoger@redhat.com>*
9118 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9120 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9122 *Martin Olsson, Neel Mehta*
9124 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9125 accommodate for stack sorting, always a write lock!).
9129 * On some versions of WIN32 Heap32Next is very slow. This can cause
9130 excessive delays in the RAND_poll(): over a minute. As a workaround
9131 include a time check in the inner Heap32Next loop too.
9135 * The code that handled flushing of data in SSL/TLS originally used the
9136 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9137 the problem outlined in PR#1949. The fix suggested there however can
9138 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9139 of Apache). So instead simplify the code to flush unconditionally.
9140 This should be fine since flushing with no data to flush is a no op.
9144 * Handle TLS versions 2.0 and later properly and correctly use the
9145 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9146 off ancient servers have a habit of sticking around for a while...
9150 * Modify compression code so it frees up structures without using the
9151 ex_data callbacks. This works around a problem where some applications
9152 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9153 restarting) then use compression (e.g. SSL with compression) later.
9154 This results in significant per-connection memory leaks and
9155 has caused some security issues including CVE-2008-1678 and
9160 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9161 change when encrypting or decrypting.
9165 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9166 connect and renegotiate with servers which do not support RI.
9167 Until RI is more widely deployed this option is enabled by default.
9171 * Add "missing" ssl ctrls to clear options and mode.
9175 * If client attempts to renegotiate and doesn't support RI respond with
9176 a no_renegotiation alert as required by RFC5746. Some renegotiating
9177 TLS clients will continue a connection gracefully when they receive
9178 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9179 waiting for a server hello which it will never receive. Now we treat a
9180 received no_renegotiation alert as a fatal error. This is because
9181 applications requesting a renegotiation might well expect it to succeed
9182 and would have no code in place to handle the server denying it so the
9183 only safe thing to do is to terminate the connection.
9187 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9188 peer supports secure renegotiation and 0 otherwise. Print out peer
9189 renegotiation support in s_client/s_server.
9193 * Replace the highly broken and deprecated SPKAC certification method with
9194 the updated NID creation version. This should correctly handle UTF8.
9198 * Implement RFC5746. Re-enable renegotiation but require the extension
9199 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9200 turns out to be a bad idea. It has been replaced by
9201 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9202 SSL_CTX_set_options(). This is really not recommended unless you
9203 know what you are doing.
9205 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9207 * Fixes to stateless session resumption handling. Use initial_ctx when
9208 issuing and attempting to decrypt tickets in case it has changed during
9209 servername handling. Use a non-zero length session ID when attempting
9210 stateless session resumption: this makes it possible to determine if
9211 a resumption has occurred immediately after receiving server hello
9212 (several places in OpenSSL subtly assume this) instead of later in
9217 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9218 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9219 fixes for a few places where the return code is not checked
9222 *Julia Lawall <julia@diku.dk>*
9224 * Add --strict-warnings option to Configure script to include devteam
9225 warnings in other configurations.
9229 * Add support for --libdir option and LIBDIR variable in makefiles. This
9230 makes it possible to install openssl libraries in locations which
9231 have names other than "lib", for example "/usr/lib64" which some
9234 *Steve Henson, based on patch from Jeremy Utley*
9236 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9237 X690 8.9.12 and can produce some misleading textual output of OIDs.
9239 *Steve Henson, reported by Dan Kaminsky*
9241 * Delete MD2 from algorithm tables. This follows the recommendation in
9242 several standards that it is not used in new applications due to
9243 several cryptographic weaknesses. For binary compatibility reasons
9244 the MD2 API is still compiled in by default.
9248 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9253 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9254 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9257 *Guenter <lists@gknw.net>*
9259 * Fix the server certificate chain building code to use X509_verify_cert(),
9260 it used to have an ad-hoc builder which was unable to cope with anything
9261 other than a simple chain.
9263 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9265 * Don't check self signed certificate signatures in X509_verify_cert()
9266 by default (a flag can override this): it just wastes time without
9267 adding any security. As a useful side effect self signed root CAs
9268 with non-FIPS digests are now usable in FIPS mode.
9272 * In dtls1_process_out_of_seq_message() the check if the current message
9273 is already buffered was missing. For every new message was memory
9274 allocated, allowing an attacker to perform an denial of service attack
9275 with sending out of seq handshake messages until there is no memory
9276 left. Additionally every future message was buffered, even if the
9277 sequence number made no sense and would be part of another handshake.
9278 So only messages with sequence numbers less than 10 in advance will be
9279 buffered. ([CVE-2009-1378])
9281 *Robin Seggelmann, discovered by Daniel Mentz*
9283 * Records are buffered if they arrive with a future epoch to be
9284 processed after finishing the corresponding handshake. There is
9285 currently no limitation to this buffer allowing an attacker to perform
9286 a DOS attack with sending records with future epochs until there is no
9287 memory left. This patch adds the pqueue_size() function to determine
9288 the size of a buffer and limits the record buffer to 100 entries.
9291 *Robin Seggelmann, discovered by Daniel Mentz*
9293 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9294 parent structure is freed. ([CVE-2009-1379])
9298 * Handle non-blocking I/O properly in SSL_shutdown() call.
9300 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9302 * Add `2.5.4.*` OIDs
9304 *Ilya O. <vrghost@gmail.com>*
9306 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9308 * Disable renegotiation completely - this fixes a severe security
9309 problem ([CVE-2009-3555]) at the cost of breaking all
9310 renegotiation. Renegotiation can be re-enabled by setting
9311 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9312 run-time. This is really not recommended unless you know what
9317 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9319 * Don't set val to NULL when freeing up structures, it is freed up by
9320 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9321 zeroing past the valid field. ([CVE-2009-0789])
9323 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9325 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9326 checked correctly. This would allow some invalid signed attributes to
9327 appear to verify correctly. ([CVE-2009-0591])
9329 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9331 * Reject UniversalString and BMPString types with invalid lengths. This
9332 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9333 a legal length. ([CVE-2009-0590])
9337 * Set S/MIME signing as the default purpose rather than setting it
9338 unconditionally. This allows applications to override it at the store
9343 * Permit restricted recursion of ASN1 strings. This is needed in practice
9344 to handle some structures.
9348 * Improve efficiency of mem_gets: don't search whole buffer each time
9351 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9353 * New -hex option for openssl rand.
9357 * Print out UTF8String and NumericString when parsing ASN1.
9361 * Support NumericString type for name components.
9365 * Allow CC in the environment to override the automatically chosen
9366 compiler. Note that nothing is done to ensure flags work with the
9371 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9373 * Properly check EVP_VerifyFinal() and similar return values
9376 *Ben Laurie, Bodo Moeller, Google Security Team*
9378 * Enable TLS extensions by default.
9382 * Allow the CHIL engine to be loaded, whether the application is
9383 multithreaded or not. (This does not release the developer from the
9384 obligation to set up the dynamic locking callbacks.)
9386 *Sander Temme <sander@temme.net>*
9388 * Use correct exit code if there is an error in dgst command.
9390 *Steve Henson; problem pointed out by Roland Dirlewanger*
9392 * Tweak Configure so that you need to say "experimental-jpake" to enable
9393 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9397 * Add experimental JPAKE support, including demo authentication in
9398 s_client and s_server.
9402 * Set the comparison function in v3_addr_canonize().
9404 *Rob Austein <sra@hactrn.net>*
9406 * Add support for XMPP STARTTLS in s_client.
9408 *Philip Paeps <philip@freebsd.org>*
9410 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9411 to ensure that even with this option, only ciphersuites in the
9412 server's preference list will be accepted. (Note that the option
9413 applies only when resuming a session, so the earlier behavior was
9414 just about the algorithm choice for symmetric cryptography.)
9418 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9420 * Fix NULL pointer dereference if a DTLS server received
9421 ChangeCipherSpec as first record ([CVE-2009-1386]).
9425 * Fix a state transition in s3_srvr.c and d1_srvr.c
9426 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9430 * The fix in 0.9.8c that supposedly got rid of unsafe
9431 double-checked locking was incomplete for RSA blinding,
9432 addressing just one layer of what turns out to have been
9433 doubly unsafe triple-checked locking.
9435 So now fix this for real by retiring the MONT_HELPER macro
9436 in crypto/rsa/rsa_eay.c.
9438 *Bodo Moeller; problem pointed out by Marius Schilder*
9440 * Various precautionary measures:
9442 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9444 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9445 (NB: This would require knowledge of the secret session ticket key
9446 to exploit, in which case you'd be SOL either way.)
9448 - Change bn_nist.c so that it will properly handle input BIGNUMs
9449 outside the expected range.
9451 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9454 *Neel Mehta, Bodo Moeller*
9456 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9457 the load fails. Useful for distros.
9459 *Ben Laurie and the FreeBSD team*
9461 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9465 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9469 * Expand ENGINE to support engine supplied SSL client certificate functions.
9471 This work was sponsored by Logica.
9475 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9476 keystores. Support for SSL/TLS client authentication too.
9477 Not compiled unless enable-capieng specified to Configure.
9479 This work was sponsored by Logica.
9483 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9484 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9485 attribute creation routines such as certificate requests and PKCS#12
9490 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9492 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9493 handshake which could lead to a client crash as found using the
9494 Codenomicon TLS test suite ([CVE-2008-1672])
9496 *Steve Henson, Mark Cox*
9498 * Fix double free in TLS server name extensions which could lead to
9499 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9503 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9505 Clear the error queue to ensure that error entries left from
9506 older function calls do not interfere with the correct operation.
9508 *Lutz Jaenicke, Erik de Castro Lopo*
9510 * Remove root CA certificates of commercial CAs:
9512 The OpenSSL project does not recommend any specific CA and does not
9513 have any policy with respect to including or excluding any CA.
9514 Therefore it does not make any sense to ship an arbitrary selection
9515 of root CA certificates with the OpenSSL software.
9519 * RSA OAEP patches to fix two separate invalid memory reads.
9520 The first one involves inputs when 'lzero' is greater than
9521 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9522 before the beginning of from). The second one involves inputs where
9523 the 'db' section contains nothing but zeroes (there is a one-byte
9524 invalid read after the end of 'db').
9526 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9528 * Partial backport from 0.9.9-dev:
9530 Introduce bn_mul_mont (dedicated Montgomery multiplication
9531 procedure) as a candidate for BIGNUM assembler implementation.
9532 While 0.9.9-dev uses assembler for various architectures, only
9533 x86_64 is available by default here in the 0.9.8 branch, and
9534 32-bit x86 is available through a compile-time setting.
9536 To try the 32-bit x86 assembler implementation, use Configure
9537 option "enable-montasm" (which exists only for this backport).
9539 As "enable-montasm" for 32-bit x86 disclaims code stability
9540 anyway, in this constellation we activate additional code
9541 backported from 0.9.9-dev for further performance improvements,
9542 namely BN_from_montgomery_word. (To enable this otherwise,
9543 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9545 *Andy Polyakov (backport partially by Bodo Moeller)*
9547 * Add TLS session ticket callback. This allows an application to set
9548 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9549 values. This is useful for key rollover for example where several key
9550 sets may exist with different names.
9554 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9555 This was broken until now in 0.9.8 releases, such that the only way
9556 a registered ENGINE could be used (assuming it initialises
9557 successfully on the host) was to explicitly set it as the default
9558 for the relevant algorithms. This is in contradiction with 0.9.7
9559 behaviour and the documentation. With this fix, when an ENGINE is
9560 registered into a given algorithm's table of implementations, the
9561 'uptodate' flag is reset so that auto-discovery will be used next
9562 time a new context for that algorithm attempts to select an
9565 *Ian Lister (tweaked by Geoff Thorpe)*
9567 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9568 implementation in the following ways:
9570 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9573 Lack of BER streaming support means one pass streaming processing is
9574 only supported if data is detached: setting the streaming flag is
9575 ignored for embedded content.
9577 CMS support is disabled by default and must be explicitly enabled
9578 with the enable-cms configuration option.
9582 * Update the GMP engine glue to do direct copies between BIGNUM and
9583 mpz_t when openssl and GMP use the same limb size. Otherwise the
9584 existing "conversion via a text string export" trick is still used.
9586 *Paul Sheer <paulsheer@gmail.com>*
9588 * Zlib compression BIO. This is a filter BIO which compressed and
9589 uncompresses any data passed through it.
9593 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9594 RFC3394 compatible AES key wrapping.
9598 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9599 sets string data without copying. X509_ALGOR_set0() and
9600 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9601 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9602 from an X509_ATTRIBUTE structure optionally checking it occurs only
9603 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9608 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9609 to get the expected BN_FLG_CONSTTIME behavior.
9611 *Bodo Moeller (Google)*
9615 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9616 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9617 - added some more tests to do_tests.pl
9618 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9619 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9620 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9621 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9622 - various changes to netware.pl to enable gcc-cross builds on Win32
9624 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9625 - various changes to fix missing prototype warnings
9626 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9627 - added AES, WHIRLPOOL and CPUID assembler code to build files
9628 - added missing AES assembler make rules to mk1mf.pl
9629 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9631 *Guenter Knauf <eflash@gmx.net>*
9633 * Implement certificate status request TLS extension defined in RFC3546.
9634 A client can set the appropriate parameters and receive the encoded
9635 OCSP response via a callback. A server can query the supplied parameters
9636 and set the encoded OCSP response in the callback. Add simplified examples
9637 to s_client and s_server.
9641 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9644 + Binary incompatibility of ssl_ctx_st structure
9645 + DTLS interoperation with non-compliant servers
9646 + Don't call get_session_cb() without proposed session
9647 + Fix ia64 assembler code
9649 *Andy Polyakov, Steve Henson*
9651 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9653 * DTLS Handshake overhaul. There were longstanding issues with
9654 OpenSSL DTLS implementation, which were making it impossible for
9655 RFC 4347 compliant client to communicate with OpenSSL server.
9656 Unfortunately just fixing these incompatibilities would "cut off"
9657 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9658 server keeps tolerating non RFC compliant syntax. The opposite is
9659 not true, 0.9.8f client can not communicate with earlier server.
9660 This update even addresses CVE-2007-4995.
9664 * Changes to avoid need for function casts in OpenSSL: some compilers
9665 (gcc 4.2 and later) reject their use.
9666 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9669 * Add RFC4507 support to OpenSSL. This includes the corrections in
9670 RFC4507bis. The encrypted ticket format is an encrypted encoded
9671 SSL_SESSION structure, that way new session features are automatically
9674 If a client application caches session in an SSL_SESSION structure
9675 support is transparent because tickets are now stored in the encoded
9678 The SSL_CTX structure automatically generates keys for ticket
9679 protection in servers so again support should be possible
9680 with no application modification.
9682 If a client or server wishes to disable RFC4507 support then the option
9683 SSL_OP_NO_TICKET can be set.
9685 Add a TLS extension debugging callback to allow the contents of any client
9686 or server extensions to be examined.
9688 This work was sponsored by Google.
9692 * Add initial support for TLS extensions, specifically for the server_name
9693 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9694 have new members for a host name. The SSL data structure has an
9695 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9696 stored in that context to allow for session resumption, even after the
9697 SSL has been switched to a new SSL_CTX in reaction to a client's
9698 server_name extension.
9700 New functions (subject to change):
9702 SSL_get_servername()
9703 SSL_get_servername_type()
9706 New CTRL codes and macros (subject to change):
9708 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9709 - SSL_CTX_set_tlsext_servername_callback()
9710 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9711 - SSL_CTX_set_tlsext_servername_arg()
9712 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9714 openssl s_client has a new '-servername ...' option.
9716 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9717 '-key2 ...', '-servername_fatal' (subject to change). This allows
9718 testing the HostName extension for a specific single host name ('-cert'
9719 and '-key' remain fallbacks for handshakes without HostName
9720 negotiation). If the unrecognized_name alert has to be sent, this by
9721 default is a warning; it becomes fatal with the '-servername_fatal'
9724 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9726 * Add AES and SSE2 assembly language support to VC++ build.
9730 * Mitigate attack on final subtraction in Montgomery reduction.
9734 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9735 (which previously caused an internal error).
9739 * Squeeze another 10% out of IGE mode when in != out.
9743 * AES IGE mode speedup.
9745 *Dean Gaudet (Google)*
9747 * Add the Korean symmetric 128-bit cipher SEED (see
9748 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9749 add SEED ciphersuites from RFC 4162:
9751 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9752 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9753 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9754 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9756 To minimize changes between patchlevels in the OpenSSL 0.9.8
9757 series, SEED remains excluded from compilation unless OpenSSL
9758 is configured with 'enable-seed'.
9760 *KISA, Bodo Moeller*
9762 * Mitigate branch prediction attacks, which can be practical if a
9763 single processor is shared, allowing a spy process to extract
9764 information. For detailed background information, see
9765 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9766 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9767 and Necessary Software Countermeasures"). The core of the change
9768 are new versions BN_div_no_branch() and
9769 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9770 respectively, which are slower, but avoid the security-relevant
9771 conditional branches. These are automatically called by BN_div()
9772 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9773 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9774 remove a conditional branch.
9776 BN_FLG_CONSTTIME is the new name for the previous
9777 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9778 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9779 in the exponent causes BN_mod_exp_mont() to use the alternative
9780 implementation in BN_mod_exp_mont_consttime().) The old name
9781 remains as a deprecated alias.
9783 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9784 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9785 constant-time implementations for more than just exponentiation.
9786 Here too the old name is kept as a deprecated alias.
9788 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9789 the BN_BLINDING structure gets an independent copy of the
9790 modulus. This means that the previous `BIGNUM *m` argument to
9791 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9792 essentially becomes `const BIGNUM *m`, although we can't actually
9793 change this in the header file before 0.9.9. It allows
9794 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9795 enable BN_FLG_CONSTTIME.
9797 *Matthew D Wood (Intel Corp)*
9799 * In the SSL/TLS server implementation, be strict about session ID
9800 context matching (which matters if an application uses a single
9801 external cache for different purposes). Previously,
9802 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9803 set. This did ensure strict client verification, but meant that,
9804 with applications using a single external cache for quite
9805 different requirements, clients could circumvent ciphersuite
9806 restrictions for a given session ID context by starting a session
9807 in a different context.
9811 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9812 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9813 authentication-only ciphersuites.
9817 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9818 not complete and could lead to a possible single byte overflow
9819 ([CVE-2007-5135]) [Ben Laurie]
9821 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9823 * Since AES128 and AES256 (and similarly Camellia128 and
9824 Camellia256) share a single mask bit in the logic of
9825 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9826 kludge to work properly if AES128 is available and AES256 isn't
9827 (or if Camellia128 is available and Camellia256 isn't).
9831 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9832 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9833 When a point or a seed is encoded in a BIT STRING, we need to
9834 prevent the removal of trailing zero bits to get the proper DER
9835 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9836 of a NamedBitList, for which trailing 0 bits need to be removed.)
9840 * Have SSL/TLS server implementation tolerate "mismatched" record
9841 protocol version while receiving ClientHello even if the
9842 ClientHello is fragmented. (The server can't insist on the
9843 particular protocol version it has chosen before the ServerHello
9844 message has informed the client about his choice.)
9848 * Add RFC 3779 support.
9850 *Rob Austein for ARIN, Ben Laurie*
9852 * Load error codes if they are not already present instead of using a
9853 static variable. This allows them to be cleanly unloaded and reloaded.
9854 Improve header file function name parsing.
9858 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9859 or CAPABILITY handshake as required by RFCs.
9863 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9865 * Introduce limits to prevent malicious keys being able to
9866 cause a denial of service. ([CVE-2006-2940])
9868 *Steve Henson, Bodo Moeller*
9870 * Fix ASN.1 parsing of certain invalid structures that can result
9871 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9873 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9874 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9876 * Fix SSL client code which could crash if connecting to a
9877 malicious SSLv2 server. ([CVE-2006-4343])
9879 *Tavis Ormandy and Will Drewry, Google Security Team*
9881 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9882 match only those. Before that, "AES256-SHA" would be interpreted
9883 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9884 the same strength classification in 0.9.7h) as we currently only
9885 have a single AES bit in the ciphersuite description bitmap.
9886 That change, however, also applied to ciphersuite strings such as
9887 "RC4-MD5" that intentionally matched multiple ciphersuites --
9888 namely, SSL 2.0 ciphersuites in addition to the more common ones
9889 from SSL 3.0/TLS 1.0.
9891 So we change the selection algorithm again: Naming an explicit
9892 ciphersuite selects this one ciphersuite, and any other similar
9893 ciphersuite (same bitmap) from *other* protocol versions.
9894 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9895 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9897 Since SSL 2.0 does not have any ciphersuites for which the
9898 128/256 bit distinction would be relevant, this works for now.
9899 The proper fix will be to use different bits for AES128 and
9900 AES256, which would have avoided the problems from the beginning;
9901 however, bits are scarce, so we can only do this in a new release
9902 (not just a patchlevel) when we can change the SSL_CIPHER
9903 definition to split the single 'unsigned long mask' bitmap into
9904 multiple values to extend the available space.
9908 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9910 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9911 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9913 * Add AES IGE and biIGE modes.
9917 * Change the Unix randomness entropy gathering to use poll() when
9918 possible instead of select(), since the latter has some
9919 undesirable limitations.
9921 *Darryl Miles via Richard Levitte and Bodo Moeller*
9923 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9924 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9925 cannot be implicitly activated as part of, e.g., the "AES" alias.
9926 However, please upgrade to OpenSSL 0.9.9[-dev] for
9927 non-experimental use of the ECC ciphersuites to get TLS extension
9928 support, which is required for curve and point format negotiation
9929 to avoid potential handshake problems.
9933 * Disable rogue ciphersuites:
9935 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9936 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9937 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9939 The latter two were purportedly from
9940 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9943 Also deactivate the remaining ciphersuites from
9944 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9945 unofficial, and the ID has long expired.
9949 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9950 dual-core machines) and other potential thread-safety issues.
9954 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9955 versions), which is now available for royalty-free use
9956 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9957 Also, add Camellia TLS ciphersuites from RFC 4132.
9959 To minimize changes between patchlevels in the OpenSSL 0.9.8
9960 series, Camellia remains excluded from compilation unless OpenSSL
9961 is configured with 'enable-camellia'.
9965 * Disable the padding bug check when compression is in use. The padding
9966 bug check assumes the first packet is of even length, this is not
9967 necessarily true if compression is enabled and can result in false
9968 positives causing handshake failure. The actual bug test is ancient
9969 code so it is hoped that implementations will either have fixed it by
9970 now or any which still have the bug do not support compression.
9974 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9976 * When applying a cipher rule check to see if string match is an explicit
9977 cipher suite and only match that one cipher suite if it is.
9981 * Link in manifests for VC++ if needed.
9983 *Austin Ziegler <halostatue@gmail.com>*
9985 * Update support for ECC-based TLS ciphersuites according to
9986 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9987 TLS extensions, which are supported starting with the 0.9.9
9988 branch, not in the OpenSSL 0.9.8 branch).
9992 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9993 opaque EVP_CIPHER_CTX handling.
9997 * Fixes and enhancements to zlib compression code. We now only use
9998 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9999 to conform with the standards mentioned here:
10000 <http://www.zlib.net/DLL_FAQ.txt>
10001 Static zlib linking now works on Windows and the new --with-zlib-include
10002 --with-zlib-lib options to Configure can be used to supply the location
10003 of the headers and library. Gracefully handle case where zlib library
10008 * Several fixes and enhancements to the OID generation code. The old code
10009 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10010 handle numbers larger than ULONG_MAX, truncated printing and had a
10011 non standard OBJ_obj2txt() behaviour.
10015 * Add support for building of engines under engine/ as shared libraries
10016 under VC++ build system.
10020 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10021 Hopefully, we will not see any false combination of paths any more.
10025 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10027 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10028 (part of SSL_OP_ALL). This option used to disable the
10029 countermeasure against man-in-the-middle protocol-version
10030 rollback in the SSL 2.0 server implementation, which is a bad
10031 idea. ([CVE-2005-2969])
10033 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10034 for Information Security, National Institute of Advanced Industrial
10035 Science and Technology [AIST], Japan)*
10037 * Add two function to clear and return the verify parameter flags.
10041 * Keep cipherlists sorted in the source instead of sorting them at
10042 runtime, thus removing the need for a lock.
10046 * Avoid some small subgroup attacks in Diffie-Hellman.
10048 *Nick Mathewson and Ben Laurie*
10050 * Add functions for well-known primes.
10054 * Extended Windows CE support.
10056 *Satoshi Nakamura and Andy Polyakov*
10058 * Initialize SSL_METHOD structures at compile time instead of during
10059 runtime, thus removing the need for a lock.
10063 * Make PKCS7_decrypt() work even if no certificate is supplied by
10064 attempting to decrypt each encrypted key in turn. Add support to
10069 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10071 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10074 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10078 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10079 key into the same file any more.
10083 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10087 * Add -utf8 command line and config file option to 'ca'.
10089 *Stefan <stf@udoma.org*
10091 * Removed the macro des_crypt(), as it seems to conflict with some
10092 libraries. Use DES_crypt().
10096 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10097 involves renaming the source and generated shared-libs for
10098 both. The engines will accept the corrected or legacy ids
10099 ('ncipher' and '4758_cca' respectively) when binding. NB,
10100 this only applies when building 'shared'.
10102 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10104 * Add attribute functions to EVP_PKEY structure. Modify
10105 PKCS12_create() to recognize a CSP name attribute and
10106 use it. Make -CSP option work again in pkcs12 utility.
10110 * Add new functionality to the bn blinding code:
10111 - automatic re-creation of the BN_BLINDING parameters after
10112 a fixed number of uses (currently 32)
10113 - add new function for parameter creation
10114 - introduce flags to control the update behaviour of the
10115 BN_BLINDING parameters
10116 - hide BN_BLINDING structure
10117 Add a second BN_BLINDING slot to the RSA structure to improve
10118 performance when a single RSA object is shared among several
10123 * Add support for DTLS.
10125 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10127 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10128 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10132 * Remove buggy and incomplete DH cert support from
10133 ssl/ssl_rsa.c and ssl/s3_both.c
10137 * Use SHA-1 instead of MD5 as the default digest algorithm for
10138 the `apps/openssl` commands.
10142 * Compile clean with "-Wall -Wmissing-prototypes
10143 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10144 DEBUG_SAFESTACK must also be set.
10148 * Change ./Configure so that certain algorithms can be disabled by default.
10149 The new counterpiece to "no-xxx" is "enable-xxx".
10151 The patented RC5 and MDC2 algorithms will now be disabled unless
10152 "enable-rc5" and "enable-mdc2", respectively, are specified.
10154 (IDEA remains enabled despite being patented. This is because IDEA
10155 is frequently required for interoperability, and there is no license
10156 fee for non-commercial use. As before, "no-idea" can be used to
10157 avoid this algorithm.)
10161 * Add processing of proxy certificates (see RFC 3820). This work was
10162 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10163 EGEE (Enabling Grids for E-science in Europe).
10167 * RC4 performance overhaul on modern architectures/implementations, such
10168 as Intel P4, IA-64 and AMD64.
10172 * New utility extract-section.pl. This can be used specify an alternative
10173 section number in a pod file instead of having to treat each file as
10174 a separate case in Makefile. This can be done by adding two lines to the
10177 =for comment openssl_section:XXX
10179 The blank line is mandatory.
10183 * New arguments -certform, -keyform and -pass for s_client and s_server
10184 to allow alternative format key and certificate files and passphrase
10189 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10190 update associated structures and add various utility functions.
10192 Add new policy related verify parameters, include policy checking in
10193 standard verify code. Enhance 'smime' application with extra parameters
10194 to support policy checking and print out.
10198 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10199 Nehemiah processors. These extensions support AES encryption in hardware
10200 as well as RNG (though RNG support is currently disabled).
10202 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10204 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10208 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10210 *Andy Polyakov and a number of other people*
10212 * Improved PowerPC platform support. Most notably BIGNUM assembler
10213 implementation contributed by IBM.
10215 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10217 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10218 exponent rather than 'unsigned long'. There is a corresponding change to
10219 the new 'rsa_keygen' element of the RSA_METHOD structure.
10221 *Jelte Jansen, Geoff Thorpe*
10223 * Functionality for creating the initial serial number file is now
10224 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10226 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10227 number file to 1, which is bound to cause problems. To avoid
10228 the problems while respecting compatibility between different 0.9.7
10229 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10230 CA.pl for serial number initialization. With the new release 0.9.8,
10231 we can fix the problem directly in the 'ca' utility.)
10235 * Reduced header interdependencies by declaring more opaque objects in
10236 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10237 give fewer recursive includes, which could break lazy source code - so
10238 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10239 developers should define this symbol when building and using openssl to
10240 ensure they track the recommended behaviour, interfaces, [etc], but
10241 backwards-compatible behaviour prevails when this isn't defined.
10245 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10249 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10250 This will generate a random key of the appropriate length based on the
10251 cipher context. The EVP_CIPHER can provide its own random key generation
10252 routine to support keys of a specific form. This is used in the des and
10253 3des routines to generate a key of the correct parity. Update S/MIME
10254 code to use new functions and hence generate correct parity DES keys.
10255 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10256 valid (weak or incorrect parity).
10260 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10261 as looking them up. This is useful when the verified structure may contain
10262 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10263 present unless the new PKCS7_NO_CRL flag is asserted.
10267 * Extend ASN1 oid configuration module. It now additionally accepts the
10270 shortName = some long name, 1.2.3.4
10274 * Reimplemented the BN_CTX implementation. There is now no more static
10275 limitation on the number of variables it can handle nor the depth of the
10276 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10277 information can now expand as required, and rather than having a single
10278 static array of bignums, BN_CTX now uses a linked-list of such arrays
10279 allowing it to expand on demand whilst maintaining the usefulness of
10280 BN_CTX's "bundling".
10284 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10285 to allow all RSA operations to function using a single BN_CTX.
10289 * Preliminary support for certificate policy evaluation and checking. This
10290 is initially intended to pass the tests outlined in "Conformance Testing
10291 of Relying Party Client Certificate Path Processing Logic" v1.07.
10295 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10296 remained unused and not that useful. A variety of other little bignum
10297 tweaks and fixes have also been made continuing on from the audit (see
10302 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10303 associated ASN1, EVP and SSL functions and old ASN1 macros.
10307 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10308 and this should never fail. So the return value from the use of
10309 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10310 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10314 * BN_CTX_get() should return zero-valued bignums, providing the same
10315 initialised value as BN_new().
10317 *Geoff Thorpe, suggested by Ulf Möller*
10319 * Support for inhibitAnyPolicy certificate extension.
10323 * An audit of the BIGNUM code is underway, for which debugging code is
10324 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10325 is considered valid when processing BIGNUMs, and causes execution to
10326 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10327 further steps are taken to deliberately pollute unused data in BIGNUM
10328 structures to try and expose faulty code further on. For now, openssl will
10329 (in its default mode of operation) continue to tolerate the inconsistent
10330 forms that it has tolerated in the past, but authors and packagers should
10331 consider trying openssl and their own applications when compiled with
10332 these debugging symbols defined. It will help highlight potential bugs in
10333 their own code, and will improve the test coverage for OpenSSL itself. At
10334 some point, these tighter rules will become openssl's default to improve
10335 maintainability, though the assert()s and other overheads will remain only
10336 in debugging configurations. See bn.h for more details.
10338 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10340 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10341 that can only be obtained through BN_CTX_new() (which implicitly
10342 initialises it). The presence of this function only made it possible
10343 to overwrite an existing structure (and cause memory leaks).
10347 * Because of the callback-based approach for implementing LHASH as a
10348 template type, lh_insert() adds opaque objects to hash-tables and
10349 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10350 to clean up those corresponding objects before destroying the hash table
10351 (and losing the object pointers). So some over-zealous constifications in
10352 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10353 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10354 prototyped to have "const" restrictions on the object pointers they are
10355 given (and so aren't required to cast them away any more).
10359 * The tmdiff.h API was so ugly and minimal that our own timing utility
10360 (speed) prefers to use its own implementation. The two implementations
10361 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10362 its object type properly exposed (MS_TM) instead of casting to/from
10363 `char *`. This may still change yet if someone realises MS_TM and
10365 aren't necessarily the greatest nomenclatures - but this is what was used
10366 internally to the implementation so I've used that for now.
10370 * Ensure that deprecated functions do not get compiled when
10371 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10372 the self-tests were still using deprecated key-generation functions so
10373 these have been updated also.
10377 * Reorganise PKCS#7 code to separate the digest location functionality
10378 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10379 New function PKCS7_set_digest() to set the digest type for PKCS#7
10380 digestedData type. Add additional code to correctly generate the
10381 digestedData type and add support for this type in PKCS7 initialization
10386 * New function PKCS7_set0_type_other() this initializes a PKCS7
10387 structure of type "other".
10391 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10392 sure the loop does correctly stop and breaking ("division by zero")
10393 modulus operations are not performed. The (pre-generated) prime
10394 table crypto/bn/bn_prime.h was already correct, but it could not be
10395 re-generated on some platforms because of the "division by zero"
10396 situation in the script.
10398 *Ralf S. Engelschall*
10400 * Update support for ECC-based TLS ciphersuites according to
10401 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10402 SHA-1 now is only used for "small" curves (where the
10403 representation of a field element takes up to 24 bytes); for
10404 larger curves, the field element resulting from ECDH is directly
10405 used as premaster secret.
10407 *Douglas Stebila (Sun Microsystems Laboratories)*
10409 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10410 curve secp160r1 to the tests.
10412 *Douglas Stebila (Sun Microsystems Laboratories)*
10414 * Add the possibility to load symbols globally with DSO.
10416 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10418 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10419 control of the error stack.
10423 * Add support for STORE in ENGINE.
10427 * Add the STORE type. The intention is to provide a common interface
10428 to certificate and key stores, be they simple file-based stores, or
10429 HSM-type store, or LDAP stores, or...
10430 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10434 * Add a generic structure called OPENSSL_ITEM. This can be used to
10435 pass a list of arguments to any function as well as provide a way
10436 for a function to pass data back to the caller.
10440 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10441 works like BUF_strdup() but can be used to duplicate a portion of
10442 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10447 * Add the function sk_find_ex() which works like sk_find(), but will
10448 return an index to an element even if an exact match couldn't be
10449 found. The index is guaranteed to point at the element where the
10450 searched-for key would be inserted to preserve sorting order.
10454 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10455 takes an extra flags argument for optional functionality. Currently,
10456 the following flags are defined:
10458 OBJ_BSEARCH_VALUE_ON_NOMATCH
10459 This one gets OBJ_bsearch_ex() to return a pointer to the first
10460 element where the comparing function returns a negative or zero
10463 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10464 This one gets OBJ_bsearch_ex() to return a pointer to the first
10465 element where the comparing function returns zero. This is useful
10466 if there are more than one element where the comparing function
10471 * Make it possible to create self-signed certificates with 'openssl ca'
10472 in such a way that the self-signed certificate becomes part of the
10473 CA database and uses the same mechanisms for serial number generation
10474 as all other certificate signing. The new flag '-selfsign' enables
10475 this functionality. Adapt CA.sh and CA.pl.in.
10479 * Add functionality to check the public key of a certificate request
10480 against a given private. This is useful to check that a certificate
10481 request can be signed by that key (self-signing).
10485 * Make it possible to have multiple active certificates with the same
10486 subject in the CA index file. This is done only if the keyword
10487 'unique_subject' is set to 'no' in the main CA section (default
10488 if 'CA_default') of the configuration file. The value is saved
10489 with the database itself in a separate index attribute file,
10490 named like the index file with '.attr' appended to the name.
10494 * Generate multi-valued AVAs using '+' notation in config files for
10499 * Support for nameConstraints certificate extension.
10503 * Support for policyConstraints certificate extension.
10507 * Support for policyMappings certificate extension.
10511 * Make sure the default DSA_METHOD implementation only uses its
10512 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10513 and change its own handlers to be NULL so as to remove unnecessary
10514 indirection. This lets alternative implementations fallback to the
10515 default implementation more easily.
10519 * Support for directoryName in GeneralName related extensions
10524 * Make it possible to link applications using Makefile.shared.
10525 Make that possible even when linking against static libraries!
10529 * Support for single pass processing for S/MIME signing. This now
10530 means that S/MIME signing can be done from a pipe, in addition
10531 cleartext signing (multipart/signed type) is effectively streaming
10532 and the signed data does not need to be all held in memory.
10534 This is done with a new flag PKCS7_STREAM. When this flag is set
10535 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10536 is done after the data is output (and digests calculated) in
10537 SMIME_write_PKCS7().
10541 * Add full support for -rpath/-R, both in shared libraries and
10542 applications, at least on the platforms where it's known how
10547 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10548 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10549 will now compute a table of multiples of the generator that
10550 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10551 faster (notably in the case of a single point multiplication,
10552 scalar * generator).
10554 *Nils Larsch, Bodo Moeller*
10556 * IPv6 support for certificate extensions. The various extensions
10557 which use the IP:a.b.c.d can now take IPv6 addresses using the
10558 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10563 * Added an ENGINE that implements RSA by performing private key
10564 exponentiations with the GMP library. The conversions to and from
10565 GMP's mpz_t format aren't optimised nor are any montgomery forms
10566 cached, and on x86 it appears OpenSSL's own performance has caught up.
10567 However there are likely to be other architectures where GMP could
10568 provide a boost. This ENGINE is not built in by default, but it can be
10569 specified at Configure time and should be accompanied by the necessary
10570 linker additions, eg;
10571 ./config -DOPENSSL_USE_GMP -lgmp
10575 * "openssl engine" will not display ENGINE/DSO load failure errors when
10576 testing availability of engines with "-t" - the old behaviour is
10577 produced by increasing the feature's verbosity with "-tt".
10581 * ECDSA routines: under certain error conditions uninitialized BN objects
10582 could be freed. Solution: make sure initialization is performed early
10583 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10588 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10589 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10590 software implementations. For DSA and DH, parameter generation can
10591 also be overridden by providing the appropriate method callbacks.
10595 * Change the "progress" mechanism used in key-generation and
10596 primality testing to functions that take a new BN_GENCB pointer in
10597 place of callback/argument pairs. The new API functions have `_ex`
10598 postfixes and the older functions are reimplemented as wrappers for
10599 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10600 declarations of the old functions to help (graceful) attempts to
10601 migrate to the new functions. Also, the new key-generation API
10602 functions operate on a caller-supplied key-structure and return
10603 success/failure rather than returning a key or NULL - this is to
10604 help make "keygen" another member function of RSA_METHOD etc.
10606 Example for using the new callback interface:
10608 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10609 void *my_arg = ...;
10612 BN_GENCB_set(&my_cb, my_callback, my_arg);
10614 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10615 /* For the meaning of a, b in calls to my_callback(), see the
10616 * documentation of the function that calls the callback.
10617 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10618 * my_callback should return 1 if it wants BN_is_prime_ex()
10619 * to continue, or 0 to stop.
10624 * Change the ZLIB compression method to be stateful, and make it
10625 available to TLS with the number defined in
10626 draft-ietf-tls-compression-04.txt.
10630 * Add the ASN.1 structures and functions for CertificatePair, which
10631 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10633 CertificatePair ::= SEQUENCE {
10634 forward [0] Certificate OPTIONAL,
10635 reverse [1] Certificate OPTIONAL,
10636 -- at least one of the pair shall be present -- }
10638 Also implement the PEM functions to read and write certificate
10639 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10641 This needed to be defined, mostly for the sake of the LDAP
10642 attribute crossCertificatePair, but may prove useful elsewhere as
10647 * Make it possible to inhibit symlinking of shared libraries in
10648 Makefile.shared, for Cygwin's sake.
10652 * Extend the BIGNUM API by creating a function
10653 void BN_set_negative(BIGNUM *a, int neg);
10654 and a macro that behave like
10655 int BN_is_negative(const BIGNUM *a);
10657 to avoid the need to access 'a->neg' directly in applications.
10661 * Implement fast modular reduction for pseudo-Mersenne primes
10662 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10663 EC_GROUP_new_curve_GFp() will now automatically use this
10666 *Nils Larsch <nla@trustcenter.de>*
10668 * Add new lock type (CRYPTO_LOCK_BN).
10672 * Change the ENGINE framework to automatically load engines
10673 dynamically from specific directories unless they could be
10674 found to already be built in or loaded. Move all the
10675 current engines except for the cryptodev one to a new
10676 directory engines/.
10677 The engines in engines/ are built as shared libraries if
10678 the "shared" options was given to ./Configure or ./config.
10679 Otherwise, they are inserted in libcrypto.a.
10680 /usr/local/ssl/engines is the default directory for dynamic
10681 engines, but that can be overridden at configure time through
10682 the usual use of --prefix and/or --openssldir, and at run
10683 time with the environment variable OPENSSL_ENGINES.
10685 *Geoff Thorpe and Richard Levitte*
10687 * Add Makefile.shared, a helper makefile to build shared
10688 libraries. Adapt Makefile.org.
10692 * Add version info to Win32 DLLs.
10694 *Peter 'Luna' Runestig" <peter@runestig.com>*
10696 * Add new 'medium level' PKCS#12 API. Certificates and keys
10697 can be added using this API to created arbitrary PKCS#12
10698 files while avoiding the low level API.
10700 New options to PKCS12_create(), key or cert can be NULL and
10701 will then be omitted from the output file. The encryption
10702 algorithm NIDs can be set to -1 for no encryption, the mac
10703 iteration count can be set to 0 to omit the mac.
10705 Enhance pkcs12 utility by making the -nokeys and -nocerts
10706 options work when creating a PKCS#12 file. New option -nomac
10707 to omit the mac, NONE can be set for an encryption algorithm.
10708 New code is modified to use the enhanced PKCS12_create()
10709 instead of the low level API.
10713 * Extend ASN1 encoder to support indefinite length constructed
10714 encoding. This can output sequences tags and octet strings in
10715 this form. Modify pk7_asn1.c to support indefinite length
10716 encoding. This is experimental and needs additional code to
10717 be useful, such as an ASN1 bio and some enhanced streaming
10720 Extend template encode functionality so that tagging is passed
10721 down to the template encoder.
10725 * Let 'openssl req' fail if an argument to '-newkey' is not
10726 recognized instead of using RSA as a default.
10730 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10731 As these are not official, they are not included in "ALL";
10732 the "ECCdraft" ciphersuite group alias can be used to select them.
10734 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10736 * Add ECDH engine support.
10738 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10740 * Add ECDH in new directory crypto/ecdh/.
10742 *Douglas Stebila (Sun Microsystems Laboratories)*
10744 * Let BN_rand_range() abort with an error after 100 iterations
10745 without success (which indicates a broken PRNG).
10749 * Change BN_mod_sqrt() so that it verifies that the input value
10750 is really the square of the return value. (Previously,
10751 BN_mod_sqrt would show GIGO behaviour.)
10755 * Add named elliptic curves over binary fields from X9.62, SECG,
10756 and WAP/WTLS; add OIDs that were still missing.
10758 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10760 * Extend the EC library for elliptic curves over binary fields
10761 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10764 EC_GF2m_simple_method
10768 EC_GROUP_new_curve_GF2m
10769 EC_GROUP_set_curve_GF2m
10770 EC_GROUP_get_curve_GF2m
10771 EC_POINT_set_affine_coordinates_GF2m
10772 EC_POINT_get_affine_coordinates_GF2m
10773 EC_POINT_set_compressed_coordinates_GF2m
10775 Point compression for binary fields is disabled by default for
10776 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10779 As binary polynomials are represented as BIGNUMs, various members
10780 of the EC_GROUP and EC_POINT data structures can be shared
10781 between the implementations for prime fields and binary fields;
10782 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10783 are essentially identical to their `..._GFp` counterparts.
10784 (For simplicity, the `..._GFp` prefix has been dropped from
10785 various internal method names.)
10787 An internal 'field_div' method (similar to 'field_mul' and
10788 'field_sqr') has been added; this is used only for binary fields.
10790 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10792 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10793 through methods ('mul', 'precompute_mult').
10795 The generic implementations (now internally called 'ec_wNAF_mul'
10796 and 'ec_wNAF_precomputed_mult') remain the default if these
10797 methods are undefined.
10799 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10801 * New function EC_GROUP_get_degree, which is defined through
10802 EC_METHOD. For curves over prime fields, this returns the bit
10803 length of the modulus.
10805 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10807 * New functions EC_GROUP_dup, EC_POINT_dup.
10808 (These simply call ..._new and ..._copy).
10810 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10812 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10813 Polynomials are represented as BIGNUMs (where the sign bit is not
10814 used) in the following functions [macros]:
10817 BN_GF2m_sub [= BN_GF2m_add]
10818 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10819 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10820 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10822 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10823 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10824 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10825 BN_GF2m_cmp [= BN_ucmp]
10827 (Note that only the 'mod' functions are actually for fields GF(2^m).
10828 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10830 For some functions, an the irreducible polynomial defining a
10831 field can be given as an 'unsigned int[]' with strictly
10832 decreasing elements giving the indices of those bits that are set;
10833 i.e., p[] represents the polynomial
10834 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10836 p[0] > p[1] > ... > p[k] = 0.
10837 This applies to the following functions:
10840 BN_GF2m_mod_mul_arr
10841 BN_GF2m_mod_sqr_arr
10842 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10843 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10844 BN_GF2m_mod_exp_arr
10845 BN_GF2m_mod_sqrt_arr
10846 BN_GF2m_mod_solve_quad_arr
10850 Conversion can be performed by the following functions:
10855 bntest.c has additional tests for binary polynomial arithmetic.
10857 Two implementations for BN_GF2m_mod_div() are available.
10858 The default algorithm simply uses BN_GF2m_mod_inv() and
10859 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10860 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10861 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10863 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10865 * Add new error code 'ERR_R_DISABLED' that can be used when some
10866 functionality is disabled at compile-time.
10868 *Douglas Stebila <douglas.stebila@sun.com>*
10870 * Change default behaviour of 'openssl asn1parse' so that more
10871 information is visible when viewing, e.g., a certificate:
10873 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10874 mode the content of non-printable OCTET STRINGs is output in a
10875 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10876 avoid the appearance of a printable string.
10878 *Nils Larsch <nla@trustcenter.de>*
10880 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10882 EC_GROUP_set_asn1_flag()
10883 EC_GROUP_get_asn1_flag()
10884 EC_GROUP_set_point_conversion_form()
10885 EC_GROUP_get_point_conversion_form()
10886 These control ASN1 encoding details:
10887 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10888 has been set to OPENSSL_EC_NAMED_CURVE.
10889 - Points are encoded in uncompressed form by default; options for
10890 asn1_for are as for point2oct, namely
10891 POINT_CONVERSION_COMPRESSED
10892 POINT_CONVERSION_UNCOMPRESSED
10893 POINT_CONVERSION_HYBRID
10895 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10897 EC_GROUP_set_seed()
10898 EC_GROUP_get0_seed()
10899 EC_GROUP_get_seed_len()
10900 This is used only for ASN1 purposes (so far).
10902 *Nils Larsch <nla@trustcenter.de>*
10904 * Add 'field_type' member to EC_METHOD, which holds the NID
10905 of the appropriate field type OID. The new function
10906 EC_METHOD_get_field_type() returns this value.
10908 *Nils Larsch <nla@trustcenter.de>*
10911 EC_POINT_point2bn()
10912 EC_POINT_bn2point()
10913 EC_POINT_point2hex()
10914 EC_POINT_hex2point()
10915 providing useful interfaces to EC_POINT_point2oct() and
10916 EC_POINT_oct2point().
10918 *Nils Larsch <nla@trustcenter.de>*
10920 * Change internals of the EC library so that the functions
10921 EC_GROUP_set_generator()
10922 EC_GROUP_get_generator()
10923 EC_GROUP_get_order()
10924 EC_GROUP_get_cofactor()
10925 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10926 to methods, which would lead to unnecessary code duplication when
10927 adding different types of curves.
10929 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10931 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10932 arithmetic, and such that modified wNAFs are generated
10933 (which avoid length expansion in many cases).
10937 * Add a function EC_GROUP_check_discriminant() (defined via
10938 EC_METHOD) that verifies that the curve discriminant is non-zero.
10940 Add a function EC_GROUP_check() that makes some sanity tests
10941 on a EC_GROUP, its generator and order. This includes
10942 EC_GROUP_check_discriminant().
10944 *Nils Larsch <nla@trustcenter.de>*
10946 * Add ECDSA in new directory crypto/ecdsa/.
10948 Add applications 'openssl ecparam' and 'openssl ecdsa'
10949 (these are based on 'openssl dsaparam' and 'openssl dsa').
10951 ECDSA support is also included in various other files across the
10952 library. Most notably,
10953 - 'openssl req' now has a '-newkey ecdsa:file' option;
10954 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10955 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10956 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10957 them suitable for ECDSA where domain parameters must be
10958 extracted before the specific public key;
10959 - ECDSA engine support has been added.
10961 *Nils Larsch <nla@trustcenter.de>*
10963 * Include some named elliptic curves, and add OIDs from X9.62,
10964 SECG, and WAP/WTLS. Each curve can be obtained from the new
10966 EC_GROUP_new_by_curve_name(),
10967 and the list of available named curves can be obtained with
10968 EC_get_builtin_curves().
10969 Also add a 'curve_name' member to EC_GROUP objects, which can be
10971 EC_GROUP_set_curve_name()
10972 EC_GROUP_get_curve_name()
10974 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10976 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10977 was actually never needed) and in BN_mul(). The removal in BN_mul()
10978 required a small change in bn_mul_part_recursive() and the addition
10979 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10980 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10981 bn_sub_words() and bn_add_words() except they take arrays with
10986 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10988 * Cleanse PEM buffers before freeing them since they may contain
10991 *Benjamin Bennett <ben@psc.edu>*
10993 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10994 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10995 authentication-only ciphersuites.
10999 * Since AES128 and AES256 share a single mask bit in the logic of
11000 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11001 kludge to work properly if AES128 is available and AES256 isn't.
11005 * Expand security boundary to match 1.1.1 module.
11009 * Remove redundant features: hash file source, editing of test vectors
11010 modify fipsld to use external fips_premain.c signature.
11014 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11015 run algorithm test programs.
11019 * Make algorithm test programs more tolerant of whitespace.
11023 * Have SSL/TLS server implementation tolerate "mismatched" record
11024 protocol version while receiving ClientHello even if the
11025 ClientHello is fragmented. (The server can't insist on the
11026 particular protocol version it has chosen before the ServerHello
11027 message has informed the client about his choice.)
11031 * Load error codes if they are not already present instead of using a
11032 static variable. This allows them to be cleanly unloaded and reloaded.
11036 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11038 * Introduce limits to prevent malicious keys being able to
11039 cause a denial of service. ([CVE-2006-2940])
11041 *Steve Henson, Bodo Moeller*
11043 * Fix ASN.1 parsing of certain invalid structures that can result
11044 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11046 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11047 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11049 * Fix SSL client code which could crash if connecting to a
11050 malicious SSLv2 server. ([CVE-2006-4343])
11052 *Tavis Ormandy and Will Drewry, Google Security Team*
11054 * Change ciphersuite string processing so that an explicit
11055 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11056 will no longer include "AES128-SHA"), and any other similar
11057 ciphersuite (same bitmap) from *other* protocol versions (so that
11058 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11059 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11060 changes from 0.9.8b and 0.9.8d.
11064 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11066 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11067 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11069 * Change the Unix randomness entropy gathering to use poll() when
11070 possible instead of select(), since the latter has some
11071 undesirable limitations.
11073 *Darryl Miles via Richard Levitte and Bodo Moeller*
11075 * Disable rogue ciphersuites:
11077 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11078 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11079 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11081 The latter two were purportedly from
11082 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11085 Also deactivate the remaining ciphersuites from
11086 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11087 unofficial, and the ID has long expired.
11091 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11092 dual-core machines) and other potential thread-safety issues.
11096 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11098 * Adapt fipsld and the build system to link against the validated FIPS
11099 module in FIPS mode.
11103 * Fixes for VC++ 2005 build under Windows.
11107 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11108 from a Windows bash shell such as MSYS. It is autodetected from the
11109 "config" script when run from a VC++ environment. Modify standard VC++
11110 build to use fipscanister.o from the GNU make build.
11114 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11116 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11117 The value now differs depending on if you build for FIPS or not.
11118 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11119 safely run with a non-FIPSed libcrypto, as it may crash because of
11120 the difference induced by this change.
11124 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11126 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11127 (part of SSL_OP_ALL). This option used to disable the
11128 countermeasure against man-in-the-middle protocol-version
11129 rollback in the SSL 2.0 server implementation, which is a bad
11130 idea. ([CVE-2005-2969])
11132 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11133 for Information Security, National Institute of Advanced Industrial
11134 Science and Technology [AIST, Japan)]*
11136 * Minimal support for X9.31 signatures and PSS padding modes. This is
11137 mainly for FIPS compliance and not fully integrated at this stage.
11141 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11142 the exponentiation using a fixed-length exponent. (Otherwise,
11143 the information leaked through timing could expose the secret key
11144 after many signatures; cf. Bleichenbacher's attack on DSA with
11149 * Make a new fixed-window mod_exp implementation the default for
11150 RSA, DSA, and DH private-key operations so that the sequence of
11151 squares and multiplies and the memory access pattern are
11152 independent of the particular secret key. This will mitigate
11153 cache-timing and potential related attacks.
11155 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11156 and this is automatically used by BN_mod_exp_mont() if the new flag
11157 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11158 will use this BN flag for private exponents unless the flag
11159 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11160 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11162 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11164 * Change the client implementation for SSLv23_method() and
11165 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11166 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11167 (Previously, the SSL 2.0 backwards compatible Client Hello
11168 message format would be used even with SSL_OP_NO_SSLv2.)
11172 * Add support for smime-type MIME parameter in S/MIME messages which some
11177 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11178 a threadsafe manner. Modify rsa code to use new function and add calls
11179 to dsa and dh code (which had race conditions before).
11183 * Include the fixed error library code in the C error file definitions
11184 instead of fixing them up at runtime. This keeps the error code
11185 structures constant.
11189 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11191 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11194 * Fixes for newer kerberos headers. NB: the casts are needed because
11195 the 'length' field is signed on one version and unsigned on another
11196 with no (?) obvious way to tell the difference, without these VC++
11197 complains. Also the "definition" of FAR (blank) is no longer included
11198 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11199 some needed definitions.
11203 * Undo Cygwin change.
11207 * Added support for proxy certificates according to RFC 3820.
11208 Because they may be a security thread to unaware applications,
11209 they must be explicitly allowed in run-time. See
11210 docs/HOWTO/proxy_certificates.txt for further information.
11214 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11216 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11217 server and client random values. Previously
11218 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11219 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11221 This change has negligible security impact because:
11223 1. Server and client random values still have 24 bytes of pseudo random
11226 2. Server and client random values are sent in the clear in the initial
11229 3. The master secret is derived using the premaster secret (48 bytes in
11230 size for static RSA ciphersuites) as well as client server and random
11233 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11236 *Stephen Henson, reported by UK NISCC*
11238 * Use Windows randomness collection on Cygwin.
11242 * Fix hang in EGD/PRNGD query when communication socket is closed
11243 prematurely by EGD/PRNGD.
11245 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11247 * Prompt for pass phrases when appropriate for PKCS12 input format.
11251 * Back-port of selected performance improvements from development
11252 branch, as well as improved support for PowerPC platforms.
11256 * Add lots of checks for memory allocation failure, error codes to indicate
11257 failure and freeing up memory if a failure occurs.
11259 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11261 * Add new -passin argument to dgst.
11265 * Perform some character comparisons of different types in X509_NAME_cmp:
11266 this is needed for some certificates that re-encode DNs into UTF8Strings
11267 (in violation of RFC3280) and can't or won't issue name rollover
11272 * Make an explicit check during certificate validation to see that
11273 the CA setting in each certificate on the chain is correct. As a
11274 side effect always do the following basic checks on extensions,
11275 not just when there's an associated purpose to the check:
11277 - if there is an unhandled critical extension (unless the user
11278 has chosen to ignore this fault)
11279 - if the path length has been exceeded (if one is set at all)
11280 - that certain extensions fit the associated purpose (if one has
11285 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11287 * Avoid a race condition when CRLs are checked in a multi threaded
11288 environment. This would happen due to the reordering of the revoked
11289 entries during signature checking and serial number lookup. Now the
11290 encoding is cached and the serial number sort performed under a lock.
11291 Add new STACK function sk_is_sorted().
11295 * Add Delta CRL to the extension code.
11299 * Various fixes to s3_pkt.c so alerts are sent properly.
11301 *David Holmes <d.holmes@f5.com>*
11303 * Reduce the chances of duplicate issuer name and serial numbers (in
11304 violation of RFC3280) using the OpenSSL certificate creation utilities.
11305 This is done by creating a random 64 bit value for the initial serial
11306 number when a serial number file is created or when a self signed
11307 certificate is created using 'openssl req -x509'. The initial serial
11308 number file is created using 'openssl x509 -next_serial' in CA.pl
11309 rather than being initialized to 1.
11313 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11315 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11316 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11318 *Joe Orton, Steve Henson*
11320 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11323 *Joe Orton, Steve Henson*
11325 * Make it possible to have multiple active certificates with the same
11326 subject in the CA index file. This is done only if the keyword
11327 'unique_subject' is set to 'no' in the main CA section (default
11328 if 'CA_default') of the configuration file. The value is saved
11329 with the database itself in a separate index attribute file,
11330 named like the index file with '.attr' appended to the name.
11334 * X509 verify fixes. Disable broken certificate workarounds when
11335 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11336 keyUsage extension present. Don't accept CRLs with unhandled critical
11337 extensions: since verify currently doesn't process CRL extensions this
11338 rejects a CRL with *any* critical extensions. Add new verify error codes
11343 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11344 A clarification of RFC2560 will require the use of OCTET STRINGs and
11345 some implementations cannot handle the current raw format. Since OpenSSL
11346 copies and compares OCSP nonces as opaque blobs without any attempt at
11347 parsing them this should not create any compatibility issues.
11351 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11352 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11353 this HMAC (and other) operations are several times slower than OpenSSL
11358 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11360 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11362 * Use the correct content when signing type "other".
11366 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11368 * Fix various bugs revealed by running the NISCC test suite:
11370 Stop out of bounds reads in the ASN1 code when presented with
11371 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11373 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11375 If verify callback ignores invalid public key errors don't try to check
11376 certificate signature with the NULL public key.
11380 * New -ignore_err option in ocsp application to stop the server
11381 exiting on the first error in a request.
11385 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11386 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11391 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11392 extra data after the compression methods not only for TLS 1.0
11393 but also for SSL 3.0 (as required by the specification).
11395 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11397 * Change X509_certificate_type() to mark the key as exported/exportable
11398 when it's 512 *bits* long, not 512 bytes.
11402 * Change AES_cbc_encrypt() so it outputs exact multiple of
11403 blocks during encryption.
11407 * Various fixes to base64 BIO and non blocking I/O. On write
11408 flushes were not handled properly if the BIO retried. On read
11409 data was not being buffered properly and had various logic bugs.
11410 This also affects blocking I/O when the data being decoded is a
11415 * Various S/MIME bugfixes and compatibility changes:
11416 output correct application/pkcs7 MIME type if
11417 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11418 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11419 of files as .eml work). Correctly handle very long lines in MIME
11424 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11426 * Countermeasure against the Klima-Pokorny-Rosa extension of
11427 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11428 a protocol version number mismatch like a decryption error
11429 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11433 * Turn on RSA blinding by default in the default implementation
11434 to avoid a timing attack. Applications that don't want it can call
11435 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11436 They would be ill-advised to do so in most cases.
11438 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11440 * Change RSA blinding code so that it works when the PRNG is not
11441 seeded (in this case, the secret RSA exponent is abused as
11442 an unpredictable seed -- if it is not unpredictable, there
11443 is no point in blinding anyway). Make RSA blinding thread-safe
11444 by remembering the creator's thread ID in rsa->blinding and
11445 having all other threads use local one-time blinding factors
11446 (this requires more computation than sharing rsa->blinding, but
11447 avoids excessive locking; and if an RSA object is not shared
11448 between threads, blinding will still be very fast).
11452 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11453 ENGINE as defaults for all supported algorithms irrespective of
11454 the 'flags' parameter. 'flags' is now honoured, so applications
11455 should make sure they are passing it correctly.
11459 * Target "mingw" now allows native Windows code to be generated in
11460 the Cygwin environment as well as with the MinGW compiler.
11464 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11466 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11467 via timing by performing a MAC computation even if incorrect
11468 block cipher padding has been found. This is a countermeasure
11469 against active attacks where the attacker has to distinguish
11470 between bad padding and a MAC verification error. ([CVE-2003-0078])
11472 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11473 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11474 Martin Vuagnoux (EPFL, Ilion)*
11476 * Make the no-err option work as intended. The intention with no-err
11477 is not to have the whole error stack handling routines removed from
11478 libcrypto, it's only intended to remove all the function name and
11479 reason texts, thereby removing some of the footprint that may not
11480 be interesting if those errors aren't displayed anyway.
11482 NOTE: it's still possible for any application or module to have its
11483 own set of error texts inserted. The routines are there, just not
11484 used by default when no-err is given.
11488 * Add support for FreeBSD on IA64.
11490 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11492 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11493 Kerberos function mit_des_cbc_cksum(). Before this change,
11494 the value returned by DES_cbc_cksum() was like the one from
11495 mit_des_cbc_cksum(), except the bytes were swapped.
11497 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11499 * Allow an application to disable the automatic SSL chain building.
11500 Before this a rather primitive chain build was always performed in
11501 ssl3_output_cert_chain(): an application had no way to send the
11502 correct chain if the automatic operation produced an incorrect result.
11504 Now the chain builder is disabled if either:
11506 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11508 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11510 The reasoning behind this is that an application would not want the
11511 auto chain building to take place if extra chain certificates are
11512 present and it might also want a means of sending no additional
11513 certificates (for example the chain has two certificates and the
11518 * Add the possibility to build without the ENGINE framework.
11520 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11522 * Under Win32 gmtime() can return NULL: check return value in
11523 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11527 * DSA routines: under certain error conditions uninitialized BN objects
11528 could be freed. Solution: make sure initialization is performed early
11529 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11530 Nils Larsch <nla@trustcenter.de> via PR#459)
11534 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11535 checked on reconnect on the client side, therefore session resumption
11536 could still fail with a "ssl session id is different" error. This
11537 behaviour is masked when SSL_OP_ALL is used due to
11538 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11539 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11540 followup to PR #377.
11544 * IA-32 assembler support enhancements: unified ELF targets, support
11545 for SCO/Caldera platforms, fix for Cygwin shared build.
11549 * Add support for FreeBSD on sparc64. As a consequence, support for
11550 FreeBSD on non-x86 processors is separate from x86 processors on
11551 the config script, much like the NetBSD support.
11553 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11555 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11557 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11560 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11561 code (06) was taken as the first octet of the session ID and the last
11562 octet was ignored consequently. As a result SSLv2 client side session
11563 caching could not have worked due to the session ID mismatch between
11565 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11570 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11571 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11576 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11577 seems that in spite of existing for more than a year, many application
11578 author have done nothing to provide the necessary callbacks, which
11579 means that this particular engine will not work properly anywhere.
11580 This is a very unfortunate situation which forces us, in the name
11581 of usability, to give the hw_ncipher.c a static lock, which is part
11583 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11584 appear in 0.9.8 or later. We EXPECT application authors to have
11585 dealt properly with this when 0.9.8 is released (unless we actually
11586 make such changes in the libcrypto locking code that changes will
11587 have to be made anyway).
11591 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11592 octets have been read, EOF or an error occurs. Without this change
11593 some truncated ASN1 structures will not produce an error.
11597 * Disable Heimdal support, since it hasn't been fully implemented.
11598 Still give the possibility to force the use of Heimdal, but with
11599 warnings and a request that patches get sent to openssl-dev.
11603 * Add the VC-CE target, introduce the WINCE sysname, and add
11604 INSTALL.WCE and appropriate conditionals to make it build.
11606 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11608 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11609 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11610 edit numbers of the version.
11612 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11614 * Introduce safe string copy and catenation functions
11615 (BUF_strlcpy() and BUF_strlcat()).
11617 *Ben Laurie (CHATS) and Richard Levitte*
11619 * Avoid using fixed-size buffers for one-line DNs.
11621 *Ben Laurie (CHATS)*
11623 * Add BUF_MEM_grow_clean() to avoid information leakage when
11624 resizing buffers containing secrets, and use where appropriate.
11626 *Ben Laurie (CHATS)*
11628 * Avoid using fixed size buffers for configuration file location.
11630 *Ben Laurie (CHATS)*
11632 * Avoid filename truncation for various CA files.
11634 *Ben Laurie (CHATS)*
11636 * Use sizeof in preference to magic numbers.
11638 *Ben Laurie (CHATS)*
11640 * Avoid filename truncation in cert requests.
11642 *Ben Laurie (CHATS)*
11644 * Add assertions to check for (supposedly impossible) buffer
11647 *Ben Laurie (CHATS)*
11649 * Don't cache truncated DNS entries in the local cache (this could
11650 potentially lead to a spoofing attack).
11652 *Ben Laurie (CHATS)*
11654 * Fix various buffers to be large enough for hex/decimal
11655 representations in a platform independent manner.
11657 *Ben Laurie (CHATS)*
11659 * Add CRYPTO_realloc_clean() to avoid information leakage when
11660 resizing buffers containing secrets, and use where appropriate.
11662 *Ben Laurie (CHATS)*
11664 * Add BIO_indent() to avoid much slightly worrying code to do
11667 *Ben Laurie (CHATS)*
11669 * Convert sprintf()/BIO_puts() to BIO_printf().
11671 *Ben Laurie (CHATS)*
11673 * buffer_gets() could terminate with the buffer only half
11676 *Ben Laurie (CHATS)*
11678 * Add assertions to prevent user-supplied crypto functions from
11679 overflowing internal buffers by having large block sizes, etc.
11681 *Ben Laurie (CHATS)*
11683 * New OPENSSL_assert() macro (similar to assert(), but enabled
11686 *Ben Laurie (CHATS)*
11688 * Eliminate unused copy of key in RC4.
11690 *Ben Laurie (CHATS)*
11692 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11694 *Ben Laurie (CHATS)*
11696 * Fix off-by-one error in EGD path.
11698 *Ben Laurie (CHATS)*
11700 * If RANDFILE path is too long, ignore instead of truncating.
11702 *Ben Laurie (CHATS)*
11704 * Eliminate unused and incorrectly sized X.509 structure
11707 *Ben Laurie (CHATS)*
11709 * Eliminate unused and dangerous function knumber().
11711 *Ben Laurie (CHATS)*
11713 * Eliminate unused and dangerous structure, KSSL_ERR.
11715 *Ben Laurie (CHATS)*
11717 * Protect against overlong session ID context length in an encoded
11718 session object. Since these are local, this does not appear to be
11721 *Ben Laurie (CHATS)*
11723 * Change from security patch (see 0.9.6e below) that did not affect
11724 the 0.9.6 release series:
11726 Remote buffer overflow in SSL3 protocol - an attacker could
11727 supply an oversized master key in Kerberos-enabled versions.
11730 *Ben Laurie (CHATS)*
11732 * Change the SSL kerb5 codes to match RFC 2712.
11736 * Make -nameopt work fully for req and add -reqopt switch.
11738 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11740 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11742 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11744 * Make sure tests can be performed even if the corresponding algorithms
11745 have been removed entirely. This was also the last step to make
11746 OpenSSL compilable with DJGPP under all reasonable conditions.
11748 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11750 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11751 to allow version independent disabling of normally unselected ciphers,
11752 which may be activated as a side-effect of selecting a single cipher.
11754 (E.g., cipher list string "RSA" enables ciphersuites that are left
11755 out of "ALL" because they do not provide symmetric encryption.
11756 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11758 *Lutz Jaenicke, Bodo Moeller*
11760 * Add appropriate support for separate platform-dependent build
11761 directories. The recommended way to make a platform-dependent
11762 build directory is the following (tested on Linux), maybe with
11765 # Place yourself outside of the OpenSSL source tree. In
11766 # this example, the environment variable OPENSSL_SOURCE
11767 # is assumed to contain the absolute OpenSSL source directory.
11768 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11769 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11770 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11771 mkdir -p `dirname $F`
11772 ln -s $OPENSSL_SOURCE/$F $F
11775 To be absolutely sure not to disturb the source tree, a "make clean"
11776 is a good thing. If it isn't successful, don't worry about it,
11777 it probably means the source directory is very clean.
11781 * Make sure any ENGINE control commands make local copies of string
11782 pointers passed to them whenever necessary. Otherwise it is possible
11783 the caller may have overwritten (or deallocated) the original string
11784 data when a later ENGINE operation tries to use the stored values.
11786 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11788 * Improve diagnostics in file reading and command-line digests.
11790 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11792 * Add AES modes CFB and OFB to the object database. Correct an
11793 error in AES-CFB decryption.
11797 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11798 allows existing EVP_CIPHER_CTX structures to be reused after
11799 calling `EVP_*Final()`. This behaviour is used by encryption
11800 BIOs and some applications. This has the side effect that
11801 applications must explicitly clean up cipher contexts with
11802 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11806 * Check the values of dna and dnb in bn_mul_recursive before calling
11807 bn_mul_comba (a non zero value means the a or b arrays do not contain
11808 n2 elements) and fallback to bn_mul_normal if either is not zero.
11812 * Fix escaping of non-ASCII characters when using the -subj option
11813 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11817 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11818 form for "surname", serialNumber has no short form.
11819 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11820 therefore remove "mail" short name for "internet 7".
11821 The OID for unique identifiers in X509 certificates is
11822 x500UniqueIdentifier, not uniqueIdentifier.
11823 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11827 * Add an "init" command to the ENGINE config module and auto initialize
11828 ENGINEs. Without any "init" command the ENGINE will be initialized
11829 after all ctrl commands have been executed on it. If init=1 the
11830 ENGINE is initialized at that point (ctrls before that point are run
11831 on the uninitialized ENGINE and after on the initialized one). If
11832 init=0 then the ENGINE will not be initialized at all.
11836 * Fix the 'app_verify_callback' interface so that the user-defined
11837 argument is actually passed to the callback: In the
11838 SSL_CTX_set_cert_verify_callback() prototype, the callback
11839 declaration has been changed from
11842 int (*cb)(X509_STORE_CTX *,void *);
11843 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11844 i=s->ctx->app_verify_callback(&ctx)
11845 has been changed into
11846 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11848 To update applications using SSL_CTX_set_cert_verify_callback(),
11849 a dummy argument can be added to their callback functions.
11851 *D. K. Smetters <smetters@parc.xerox.com>*
11853 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11855 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11857 * Add and OPENSSL_LOAD_CONF define which will cause
11858 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11859 This allows older applications to transparently support certain
11860 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11861 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11862 load the config file and OPENSSL_add_all_algorithms_conf() which will
11863 always load it have also been added.
11867 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11868 Adjust NIDs and EVP layer.
11870 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11872 * Config modules support in openssl utility.
11874 Most commands now load modules from the config file,
11875 though in a few (such as version) this isn't done
11876 because it couldn't be used for anything.
11878 In the case of ca and req the config file used is
11879 the same as the utility itself: that is the -config
11880 command line option can be used to specify an
11885 * Move default behaviour from OPENSSL_config(). If appname is NULL
11886 use "openssl_conf" if filename is NULL use default openssl config file.
11890 * Add an argument to OPENSSL_config() to allow the use of an alternative
11891 config section name. Add a new flag to tolerate a missing config file
11892 and move code to CONF_modules_load_file().
11896 * Support for crypto accelerator cards from Accelerated Encryption
11897 Processing, www.aep.ie. (Use engine 'aep')
11898 The support was copied from 0.9.6c [engine] and adapted/corrected
11899 to work with the new engine framework.
11901 *AEP Inc. and Richard Levitte*
11903 * Support for SureWare crypto accelerator cards from Baltimore
11904 Technologies. (Use engine 'sureware')
11905 The support was copied from 0.9.6c [engine] and adapted
11906 to work with the new engine framework.
11910 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11911 make the newer ENGINE framework commands for the CHIL engine work.
11913 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11915 * Make it possible to produce shared libraries on ReliantUNIX.
11917 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11919 * Add the configuration target debug-linux-ppro.
11920 Make 'openssl rsa' use the general key loading routines
11921 implemented in `apps.c`, and make those routines able to
11922 handle the key format FORMAT_NETSCAPE and the variant
11925 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11927 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11929 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11931 * Add -keyform to rsautl, and document -engine.
11933 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11935 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11936 BIO_R_NO_SUCH_FILE error code rather than the generic
11937 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11941 * Add new functions
11942 ERR_peek_last_error
11943 ERR_peek_last_error_line
11944 ERR_peek_last_error_line_data.
11945 These are similar to
11947 ERR_peek_error_line
11948 ERR_peek_error_line_data,
11949 but report on the latest error recorded rather than the first one
11950 still in the error queue.
11952 *Ben Laurie, Bodo Moeller*
11954 * default_algorithms option in ENGINE config module. This allows things
11956 default_algorithms = ALL
11957 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11961 * Preliminary ENGINE config module.
11965 * New experimental application configuration code.
11969 * Change the AES code to follow the same name structure as all other
11970 symmetric ciphers, and behave the same way. Move everything to
11971 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11973 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11975 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11977 *Ben Laurie and Theo de Raadt*
11979 * Add option to output public keys in req command.
11981 *Massimiliano Pala madwolf@openca.org*
11983 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11984 (up to about 10% better than before for P-192 and P-224).
11988 * New functions/macros
11990 SSL_CTX_set_msg_callback(ctx, cb)
11991 SSL_CTX_set_msg_callback_arg(ctx, arg)
11992 SSL_set_msg_callback(ssl, cb)
11993 SSL_set_msg_callback_arg(ssl, arg)
11995 to request calling a callback function
11997 void cb(int write_p, int version, int content_type,
11998 const void *buf, size_t len, SSL *ssl, void *arg)
12000 whenever a protocol message has been completely received
12001 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12002 protocol version according to which the SSL library interprets
12003 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12004 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12005 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12006 specification (change_cipher_spec(20), alert(21), handshake(22)).
12007 'buf' and 'len' point to the actual message, 'ssl' to the
12008 SSL object, and 'arg' is the application-defined value set by
12009 SSL[_CTX]_set_msg_callback_arg().
12011 'openssl s_client' and 'openssl s_server' have new '-msg' options
12012 to enable a callback that displays all protocol messages.
12016 * Change the shared library support so shared libraries are built as
12017 soon as the corresponding static library is finished, and thereby get
12018 openssl and the test programs linked against the shared library.
12019 This still only happens when the keyword "shard" has been given to
12020 the configuration scripts.
12022 NOTE: shared library support is still an experimental thing, and
12023 backward binary compatibility is still not guaranteed.
12025 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12027 * Add support for Subject Information Access extension.
12029 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12031 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12032 additional bytes when new memory had to be allocated, not just
12033 when reusing an existing buffer.
12037 * New command line and configuration option 'utf8' for the req command.
12038 This allows field values to be specified as UTF8 strings.
12042 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12043 runs for the former and machine-readable output for the latter.
12047 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12048 of the e-mail address in the DN (i.e., it will go into a certificate
12049 extension only). The new configuration file option 'email_in_dn = no'
12050 has the same effect.
12052 *Massimiliano Pala madwolf@openca.org*
12054 * Change all functions with names starting with `des_` to be starting
12055 with `DES_` instead. Add wrappers that are compatible with libdes,
12056 but are named `_ossl_old_des_*`. Finally, add macros that map the
12057 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12058 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12059 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12062 Since we provide two compatibility mappings, the user needs to
12063 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12064 compatibility is desired. The default (i.e., when that macro
12065 isn't defined) is OpenSSL 0.9.6c compatibility.
12067 There are also macros that enable and disable the support of old
12068 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12069 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12070 are defined, the default will apply: to support the old des routines.
12072 In either case, one must include openssl/des.h to get the correct
12073 definitions. Do not try to just include openssl/des_old.h, that
12076 NOTE: This is a major break of an old API into a new one. Software
12077 authors are encouraged to switch to the `DES_` style functions. Some
12078 time in the future, des_old.h and the libdes compatibility functions
12079 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12080 default), and then completely removed.
12084 * Test for certificates which contain unsupported critical extensions.
12085 If such a certificate is found during a verify operation it is
12086 rejected by default: this behaviour can be overridden by either
12087 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12088 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12089 X509_supported_extension() has also been added which returns 1 if a
12090 particular extension is supported.
12094 * Modify the behaviour of EVP cipher functions in similar way to digests
12095 to retain compatibility with existing code.
12099 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12100 compatibility with existing code. In particular the 'ctx' parameter does
12101 not have to be to be initialized before the call to EVP_DigestInit() and
12102 it is tidied up after a call to EVP_DigestFinal(). New function
12103 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12104 EVP_MD_CTX_copy() changed to not require the destination to be
12105 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12106 requires the destination to be valid.
12108 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12109 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12113 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12114 so that complete 'Handshake' protocol structures are kept in memory
12115 instead of overwriting 'msg_type' and 'length' with 'body' data.
12119 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12121 *Massimo Santin via Richard Levitte*
12123 * Major restructuring to the underlying ENGINE code. This includes
12124 reduction of linker bloat, separation of pure "ENGINE" manipulation
12125 (initialisation, etc) from functionality dealing with implementations
12126 of specific crypto interfaces. This change also introduces integrated
12127 support for symmetric ciphers and digest implementations - so ENGINEs
12128 can now accelerate these by providing EVP_CIPHER and EVP_MD
12129 implementations of their own. This is detailed in
12130 [crypto/engine/README.md](crypto/engine/README.md)
12131 as it couldn't be adequately described here. However, there are a few
12132 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12133 were changed in the original introduction of ENGINE code have now
12134 reverted back - the hooking from this code to ENGINE is now a good
12135 deal more passive and at run-time, operations deal directly with
12136 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12137 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12138 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12139 they were not being used by the framework as there is no concept of a
12140 BIGNUM_METHOD and they could not be generalised to the new
12141 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12142 ENGINE_cpy() has been removed as it cannot be consistently defined in
12147 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12151 * Change mkdef.pl to sort symbols that get the same entry number,
12152 and make sure the automatically generated functions `ERR_load_*`
12153 become part of libeay.num as well.
12157 * New function SSL_renegotiate_pending(). This returns true once
12158 renegotiation has been requested (either SSL_renegotiate() call
12159 or HelloRequest/ClientHello received from the peer) and becomes
12160 false once a handshake has been completed.
12161 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12162 sends a HelloRequest, but does not ensure that a handshake takes
12163 place. SSL_renegotiate_pending() is useful for checking if the
12164 client has followed the request.)
12168 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12169 By default, clients may request session resumption even during
12170 renegotiation (if session ID contexts permit); with this option,
12171 session resumption is possible only in the first handshake.
12173 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12174 more bits available for options that should not be part of
12175 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12179 * Add some demos for certificate and certificate request creation.
12183 * Make maximum certificate chain size accepted from the peer application
12184 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12185 "Douglas E. Engert" <deengert@anl.gov>.
12189 * Add support for shared libraries for Unixware-7
12190 (Boyd Lynn Gerber <gerberb@zenez.com>).
12194 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12195 be done prior to destruction. Use this to unload error strings from
12196 ENGINEs that load their own error strings. NB: This adds two new API
12197 functions to "get" and "set" this destroy handler in an ENGINE.
12201 * Alter all existing ENGINE implementations (except "openssl" and
12202 "openbsd") to dynamically instantiate their own error strings. This
12203 makes them more flexible to be built both as statically-linked ENGINEs
12204 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12205 Also, add stub code to each that makes building them as self-contained
12206 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12210 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12211 implementations into applications that are completely implemented in
12212 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12213 commands that can be used to configure what shared-library to load and
12214 to control aspects of the way it is handled. Also, made an update to
12215 the [README-Engine.md](README-Engine.md) file
12216 that brings its information up-to-date and
12217 provides some information and instructions on the "dynamic" ENGINE
12218 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12222 * Make it possible to unload ranges of ERR strings with a new
12223 "ERR_unload_strings" function.
12227 * Add a copy() function to EVP_MD.
12231 * Make EVP_MD routines take a context pointer instead of just the
12232 md_data void pointer.
12236 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12237 that the digest can only process a single chunk of data
12238 (typically because it is provided by a piece of
12239 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12240 is only going to provide a single chunk of data, and hence the
12241 framework needn't accumulate the data for oneshot drivers.
12245 * As with "ERR", make it possible to replace the underlying "ex_data"
12246 functions. This change also alters the storage and management of global
12247 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12248 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12249 index counters. The API functions that use this state have been changed
12250 to take a "class_index" rather than pointers to the class's local STACK
12251 and counter, and there is now an API function to dynamically create new
12252 classes. This centralisation allows us to (a) plug a lot of the
12253 thread-safety problems that existed, and (b) makes it possible to clean
12254 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12255 such data would previously have always leaked in application code and
12256 workarounds were in place to make the memory debugging turn a blind eye
12257 to it. Application code that doesn't use this new function will still
12258 leak as before, but their memory debugging output will announce it now
12259 rather than letting it slide.
12261 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12262 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12263 has a return value to indicate success or failure.
12267 * Make it possible to replace the underlying "ERR" functions such that the
12268 global state (2 LHASH tables and 2 locks) is only used by the "default"
12269 implementation. This change also adds two functions to "get" and "set"
12270 the implementation prior to it being automatically set the first time
12271 any other ERR function takes place. Ie. an application can call "get",
12272 pass the return value to a module it has just loaded, and that module
12273 can call its own "set" function using that value. This means the
12274 module's "ERR" operations will use (and modify) the error state in the
12275 application and not in its own statically linked copy of OpenSSL code.
12279 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12280 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12281 the operation, and provides a more encapsulated way for external code
12282 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12283 to use these functions rather than manually incrementing the counts.
12285 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12289 * Add EVP test program.
12293 * Add symmetric cipher support to ENGINE. Expect the API to change!
12297 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12298 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12299 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12300 These allow a CRL to be built without having to access X509_CRL fields
12301 directly. Modify 'ca' application to use new functions.
12305 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12306 bug workarounds. Rollback attack detection is a security feature.
12307 The problem will only arise on OpenSSL servers when TLSv1 is not
12308 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12309 Software authors not wanting to support TLSv1 will have special reasons
12310 for their choice and can explicitly enable this option.
12312 *Bodo Moeller, Lutz Jaenicke*
12314 * Rationalise EVP so it can be extended: don't include a union of
12315 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12316 (similar to those existing for EVP_CIPHER_CTX).
12321 EVP_MD_CTX_init(&md); /* new function call */
12322 EVP_DigestInit(&md, EVP_sha1());
12323 EVP_DigestUpdate(&md, in, len);
12324 EVP_DigestFinal(&md, out, NULL);
12325 EVP_MD_CTX_cleanup(&md); /* new function call */
12329 * Make DES key schedule conform to the usual scheme, as well as
12330 correcting its structure. This means that calls to DES functions
12331 now have to pass a pointer to a des_key_schedule instead of a
12332 plain des_key_schedule (which was actually always a pointer
12335 des_key_schedule ks;
12337 des_set_key_checked(..., &ks);
12338 des_ncbc_encrypt(..., &ks, ...);
12340 (Note that a later change renames 'des_...' into 'DES_...'.)
12344 * Initial reduction of linker bloat: the use of some functions, such as
12345 PEM causes large amounts of unused functions to be linked in due to
12346 poor organisation. For example pem_all.c contains every PEM function
12347 which has a knock on effect of linking in large amounts of (unused)
12348 ASN1 code. Grouping together similar functions and splitting unrelated
12349 functions prevents this.
12353 * Cleanup of EVP macros.
12357 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12358 correct `_ecb suffix`.
12362 * Add initial OCSP responder support to ocsp application. The
12363 revocation information is handled using the text based index
12364 use by the ca application. The responder can either handle
12365 requests generated internally, supplied in files (for example
12366 via a CGI script) or using an internal minimal server.
12370 * Add configuration choices to get zlib compression for TLS.
12374 * Changes to Kerberos SSL for RFC 2712 compliance:
12375 1. Implemented real KerberosWrapper, instead of just using
12376 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12377 2. Implemented optional authenticator field of KerberosWrapper.
12379 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12380 and authenticator structs; see crypto/krb5/.
12382 Generalized Kerberos calls to support multiple Kerberos libraries.
12383 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12384 via Richard Levitte*
12386 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12387 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12388 values for each of the key sizes rather than having just
12389 parameters (and 'speed' generating keys each time).
12393 * Speed up EVP routines.
12396 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12397 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12398 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12399 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12401 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12402 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12403 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12406 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12408 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12412 * Added the OS2-EMX target.
12414 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12416 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12417 New functions to support `NCONF `routines in extension code.
12418 New function `CONF_set_nconf()`
12419 to allow functions which take an `NCONF` to also handle the old `LHASH`
12420 structure: this means that the old `CONF` compatible routines can be
12421 retained (in particular w.rt. extensions) without having to duplicate the
12422 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12426 * Enhance the general user interface with mechanisms for inner control
12427 and with possibilities to have yes/no kind of prompts.
12431 * Change all calls to low level digest routines in the library and
12432 applications to use EVP. Add missing calls to HMAC_cleanup() and
12433 don't assume HMAC_CTX can be copied using memcpy().
12435 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12437 * Add the possibility to control engines through control names but with
12438 arbitrary arguments instead of just a string.
12439 Change the key loaders to take a UI_METHOD instead of a callback
12440 function pointer. NOTE: this breaks binary compatibility with earlier
12441 versions of OpenSSL [engine].
12442 Adapt the nCipher code for these new conditions and add a card insertion
12447 * Enhance the general user interface with mechanisms to better support
12448 dialog box interfaces, application-defined prompts, the possibility
12449 to use defaults (for example default passwords from somewhere else)
12450 and interrupts/cancellations.
12454 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12455 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12459 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12460 tidy up some unnecessarily weird code in 'sk_new()').
12462 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12464 * Change the key loading routines for ENGINEs to use the same kind
12465 callback (pem_password_cb) as all other routines that need this
12470 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12471 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12472 than this minimum value is recommended.
12476 * New random seeder for OpenVMS, using the system process statistics
12477 that are easily reachable.
12481 * Windows apparently can't transparently handle global
12482 variables defined in DLLs. Initialisations such as:
12484 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12486 won't compile. This is used by the any applications that need to
12487 declare their own ASN1 modules. This was fixed by adding the option
12488 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12489 needed for static libraries under Win32.
12493 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12494 setting of purpose and trust fields. New X509_STORE trust and
12495 purpose functions and tidy up setting in other SSL functions.
12499 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12500 structure. These are inherited by X509_STORE_CTX when it is
12501 initialised. This allows various defaults to be set in the
12502 X509_STORE structure (such as flags for CRL checking and custom
12503 purpose or trust settings) for functions which only use X509_STORE_CTX
12504 internally such as S/MIME.
12506 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12507 trust settings if they are not set in X509_STORE. This allows X509_STORE
12508 purposes and trust (in S/MIME for example) to override any set by default.
12510 Add command line options for CRL checking to smime, s_client and s_server
12515 * Initial CRL based revocation checking. If the CRL checking flag(s)
12516 are set then the CRL is looked up in the X509_STORE structure and
12517 its validity and signature checked, then if the certificate is found
12518 in the CRL the verify fails with a revoked error.
12520 Various new CRL related callbacks added to X509_STORE_CTX structure.
12522 Command line options added to 'verify' application to support this.
12524 This needs some additional work, such as being able to handle multiple
12525 CRLs with different times, extension based lookup (rather than just
12526 by subject name) and ultimately more complete V2 CRL extension
12531 * Add a general user interface API (crypto/ui/). This is designed
12532 to replace things like des_read_password and friends (backward
12533 compatibility functions using this new API are provided).
12534 The purpose is to remove prompting functions from the DES code
12535 section as well as provide for prompting through dialog boxes in
12536 a window system and the like.
12540 * Add "ex_data" support to ENGINE so implementations can add state at a
12541 per-structure level rather than having to store it globally.
12545 * Make it possible for ENGINE structures to be copied when retrieved by
12546 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12547 This causes the "original" ENGINE structure to act like a template,
12548 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12549 operational state can be localised to each ENGINE structure, despite the
12550 fact they all share the same "methods". New ENGINE structures returned in
12551 this case have no functional references and the return value is the single
12552 structural reference. This matches the single structural reference returned
12553 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12558 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12559 needs to match any other type at all we need to manually clear the
12564 * Changes to the "openssl engine" utility to include;
12565 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12566 about an ENGINE's available control commands.
12567 - executing control commands from command line arguments using the
12568 '-pre' and '-post' switches. '-post' is only used if '-t' is
12569 specified and the ENGINE is successfully initialised. The syntax for
12570 the individual commands are colon-separated, for example;
12571 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12575 * New dynamic control command support for ENGINEs. ENGINEs can now
12576 declare their own commands (numbers), names (strings), descriptions,
12577 and input types for run-time discovery by calling applications. A
12578 subset of these commands are implicitly classed as "executable"
12579 depending on their input type, and only these can be invoked through
12580 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12581 can be based on user input, config files, etc). The distinction is
12582 that "executable" commands cannot return anything other than a boolean
12583 result and can only support numeric or string input, whereas some
12584 discoverable commands may only be for direct use through
12585 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12586 pointers, or other custom uses. The "executable" commands are to
12587 support parameterisations of ENGINE behaviour that can be
12588 unambiguously defined by ENGINEs and used consistently across any
12589 OpenSSL-based application. Commands have been added to all the
12590 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12591 control over shared-library paths without source code alterations.
12595 * Changed all ENGINE implementations to dynamically allocate their
12596 ENGINEs rather than declaring them statically. Apart from this being
12597 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12598 this also allows the implementations to compile without using the
12599 internal engine_int.h header.
12603 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12604 'const' value. Any code that should be able to modify a RAND_METHOD
12605 should already have non-const pointers to it (ie. they should only
12606 modify their own ones).
12610 * Made a variety of little tweaks to the ENGINE code.
12611 - "atalla" and "ubsec" string definitions were moved from header files
12612 to C code. "nuron" string definitions were placed in variables
12613 rather than hard-coded - allowing parameterisation of these values
12614 later on via ctrl() commands.
12615 - Removed unused "#if 0"'d code.
12616 - Fixed engine list iteration code so it uses ENGINE_free() to release
12617 structural references.
12618 - Constified the RAND_METHOD element of ENGINE structures.
12619 - Constified various get/set functions as appropriate and added
12620 missing functions (including a catch-all ENGINE_cpy that duplicates
12621 all ENGINE values onto a new ENGINE except reference counts/state).
12622 - Removed NULL parameter checks in get/set functions. Setting a method
12623 or function to NULL is a way of cancelling out a previously set
12624 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12625 and doesn't justify the extra error symbols and code.
12626 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12627 flags from engine_int.h to engine.h.
12628 - Changed prototypes for ENGINE handler functions (init(), finish(),
12629 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12633 * Implement binary inversion algorithm for BN_mod_inverse in addition
12634 to the algorithm using long division. The binary algorithm can be
12635 used only if the modulus is odd. On 32-bit systems, it is faster
12636 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12637 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12638 up to 450 bits. In 64-bit environments, the binary algorithm
12639 appears to be advantageous for much longer moduli; here we use it
12640 for moduli up to 2048 bits.
12644 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12645 could not support the combine flag in choice fields.
12649 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12650 extensions from a certificate request to the certificate.
12654 * Allow multiple 'certopt' and 'nameopt' options to be separated
12655 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12656 file: this allows the display of the certificate about to be
12657 signed to be customised, to allow certain fields to be included
12658 or excluded and extension details. The old system didn't display
12659 multicharacter strings properly, omitted fields not in the policy
12660 and couldn't display additional details such as extensions.
12664 * Function EC_POINTs_mul for multiple scalar multiplication
12665 of an arbitrary number of elliptic curve points
12666 \sum scalars[i]*points[i],
12667 optionally including the generator defined for the EC_GROUP:
12668 scalar*generator + \sum scalars[i]*points[i].
12670 EC_POINT_mul is a simple wrapper function for the typical case
12671 that the point list has just one item (besides the optional
12676 * First EC_METHODs for curves over GF(p):
12678 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12679 operations and provides various method functions that can also
12680 operate with faster implementations of modular arithmetic.
12682 EC_GFp_mont_method() reuses most functions that are part of
12683 EC_GFp_simple_method, but uses Montgomery arithmetic.
12685 *Bodo Moeller; point addition and point doubling
12686 implementation directly derived from source code provided by
12687 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12689 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12690 crypto/ec/ec_lib.c):
12692 Curves are EC_GROUP objects (with an optional group generator)
12693 based on EC_METHODs that are built into the library.
12695 Points are EC_POINT objects based on EC_GROUP objects.
12697 Most of the framework would be able to handle curves over arbitrary
12698 finite fields, but as there are no obvious types for fields other
12699 than GF(p), some functions are limited to that for now.
12703 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12704 that the file contains a complete HTTP response.
12708 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12709 change the def and num file printf format specifier from "%-40sXXX"
12710 to "%-39s XXX". The latter will always guarantee a space after the
12711 field while the former will cause them to run together if the field
12712 is 40 of more characters long.
12716 * Constify the cipher and digest 'method' functions and structures
12717 and modify related functions to take constant EVP_MD and EVP_CIPHER
12722 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12723 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12727 * Modify `EVP_Digest*()` routines so they now return values. Although the
12728 internal software routines can never fail additional hardware versions
12733 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12735 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12736 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12739 ERR_R_NESTED_ASN1_ERROR
12741 ERR_R_MISSING_ASN1_EOS
12742 were 4 .. 9, conflicting with
12743 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12745 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12746 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12748 Add new error code 'ERR_R_INTERNAL_ERROR'.
12752 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12757 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12758 sets the subject name for a new request or supersedes the
12759 subject name in a given request. Formats that can be parsed are
12760 'CN=Some Name, OU=myOU, C=IT'
12762 'CN=Some Name/OU=myOU/C=IT'.
12764 Add options '-batch' and '-verbose' to 'openssl req'.
12766 *Massimiliano Pala <madwolf@hackmasters.net>*
12768 * Introduce the possibility to access global variables through
12769 functions on platform were that's the best way to handle exporting
12770 global variables in shared libraries. To enable this functionality,
12771 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12772 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12773 is normally done by Configure or something similar).
12775 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12776 in the source file (foo.c) like this:
12778 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12779 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12781 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12782 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12784 OPENSSL_DECLARE_GLOBAL(int,foo);
12785 #define foo OPENSSL_GLOBAL_REF(foo)
12786 OPENSSL_DECLARE_GLOBAL(double,bar);
12787 #define bar OPENSSL_GLOBAL_REF(bar)
12789 The #defines are very important, and therefore so is including the
12790 header file everywhere where the defined globals are used.
12792 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12793 of ASN.1 items, but that structure is a bit different.
12795 The largest change is in util/mkdef.pl which has been enhanced with
12796 better and easier to understand logic to choose which symbols should
12797 go into the Windows .def files as well as a number of fixes and code
12798 cleanup (among others, algorithm keywords are now sorted
12799 lexicographically to avoid constant rewrites).
12803 * In BN_div() keep a copy of the sign of 'num' before writing the
12804 result to 'rm' because if rm==num the value will be overwritten
12805 and produce the wrong result if 'num' is negative: this caused
12806 problems with BN_mod() and BN_nnmod().
12810 * Function OCSP_request_verify(). This checks the signature on an
12811 OCSP request and verifies the signer certificate. The signer
12812 certificate is just checked for a generic purpose and OCSP request
12817 * Add OCSP_check_validity() function to check the validity of OCSP
12818 responses. OCSP responses are prepared in real time and may only
12819 be a few seconds old. Simply checking that the current time lies
12820 between thisUpdate and nextUpdate max reject otherwise valid responses
12821 caused by either OCSP responder or client clock inaccuracy. Instead
12822 we allow thisUpdate and nextUpdate to fall within a certain period of
12823 the current time. The age of the response can also optionally be
12824 checked. Two new options -validity_period and -status_age added to
12829 * If signature or public key algorithm is unrecognized print out its
12830 OID rather that just UNKNOWN.
12834 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12835 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12836 ID to be generated from the issuer certificate alone which can then be
12837 passed to OCSP_id_issuer_cmp().
12841 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12842 ASN1 modules to export functions returning ASN1_ITEM pointers
12843 instead of the ASN1_ITEM structures themselves. This adds several
12844 new macros which allow the underlying ASN1 function/structure to
12845 be accessed transparently. As a result code should not use ASN1_ITEM
12846 references directly (such as &X509_it) but instead use the relevant
12847 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12848 use of the new ASN1 code on platforms where exporting structures
12849 is problematical (for example in shared libraries) but exporting
12850 functions returning pointers to structures is not.
12854 * Add support for overriding the generation of SSL/TLS session IDs.
12855 These callbacks can be registered either in an SSL_CTX or per SSL.
12856 The purpose of this is to allow applications to control, if they wish,
12857 the arbitrary values chosen for use as session IDs, particularly as it
12858 can be useful for session caching in multiple-server environments. A
12859 command-line switch for testing this (and any client code that wishes
12860 to use such a feature) has been added to "s_server".
12862 *Geoff Thorpe, Lutz Jaenicke*
12864 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12865 of the form `#if defined(...) || defined(...) || ...` and
12866 `#if !defined(...) && !defined(...) && ...`. This also avoids
12867 the growing number of special cases it was previously handling.
12871 * Make all configuration macros available for application by making
12872 sure they are available in opensslconf.h, by giving them names starting
12873 with `OPENSSL_` to avoid conflicts with other packages and by making
12874 sure e_os2.h will cover all platform-specific cases together with
12876 Additionally, it is now possible to define configuration/platform-
12877 specific names (called "system identities"). In the C code, these
12878 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12879 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12880 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12885 * New option -set_serial to 'req' and 'x509' this allows the serial
12886 number to use to be specified on the command line. Previously self
12887 signed certificates were hard coded with serial number 0 and the
12888 CA options of 'x509' had to use a serial number in a file which was
12893 * New options to 'ca' utility to support V2 CRL entry extensions.
12894 Currently CRL reason, invalidity date and hold instruction are
12895 supported. Add new CRL extensions to V3 code and some new objects.
12899 * New function EVP_CIPHER_CTX_set_padding() this is used to
12900 disable standard block padding (aka PKCS#5 padding) in the EVP
12901 API, which was previously mandatory. This means that the data is
12902 not padded in any way and so the total length much be a multiple
12903 of the block size, otherwise an error occurs.
12907 * Initial (incomplete) OCSP SSL support.
12911 * New function OCSP_parse_url(). This splits up a URL into its host,
12912 port and path components: primarily to parse OCSP URLs. New -url
12913 option to ocsp utility.
12917 * New nonce behavior. The return value of OCSP_check_nonce() now
12918 reflects the various checks performed. Applications can decide
12919 whether to tolerate certain situations such as an absent nonce
12920 in a response when one was present in a request: the ocsp application
12921 just prints out a warning. New function OCSP_add1_basic_nonce()
12922 this is to allow responders to include a nonce in a response even if
12923 the request is nonce-less.
12927 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12928 skipped when using openssl x509 multiple times on a single input file,
12929 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12933 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12934 set string type: to handle setting ASN1_TIME structures. Fix ca
12935 utility to correctly initialize revocation date of CRLs.
12939 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12940 the clients preferred ciphersuites and rather use its own preferences.
12941 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12942 Internet Explorer by ensuring unchanged hash method during stepup.
12943 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12947 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12948 to aes and add a new 'exist' option to print out symbols that don't
12953 * Additional options to ocsp utility to allow flags to be set and
12954 additional certificates supplied.
12958 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12959 OCSP client a number of certificate to only verify the response
12964 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12965 handle the new API. Currently only ECB, CBC modes supported. Add new
12968 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12969 Encryption Standard (AES) Ciphersuites for Transport Layer
12970 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12971 not enabled by default and were not part of the "ALL" ciphersuite
12972 alias because they were not yet official; they could be
12973 explicitly requested by specifying the "AESdraft" ciphersuite
12974 group alias. In the final release of OpenSSL 0.9.7, the group
12975 alias is called "AES" and is part of "ALL".)
12977 *Ben Laurie, Steve Henson, Bodo Moeller*
12979 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12980 request to response.
12984 * Functions for OCSP responders. OCSP_request_onereq_count(),
12985 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12986 extract information from a certificate request. OCSP_response_create()
12987 creates a response and optionally adds a basic response structure.
12988 OCSP_basic_add1_status() adds a complete single response to a basic
12989 response and returns the OCSP_SINGLERESP structure just added (to allow
12990 extensions to be included for example). OCSP_basic_add1_cert() adds a
12991 certificate to a basic response and OCSP_basic_sign() signs a basic
12992 response with various flags. New helper functions ASN1_TIME_check()
12993 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12994 (converts ASN1_TIME to GeneralizedTime).
12998 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12999 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13000 structure from a certificate. X509_pubkey_digest() digests the public_key
13001 contents: this is used in various key identifiers.
13005 * Make sk_sort() tolerate a NULL argument.
13007 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13009 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13010 passed by the function are trusted implicitly. If any of them signed the
13011 response then it is assumed to be valid and is not verified.
13015 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13016 to data. This was previously part of the PKCS7 ASN1 code. This
13017 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13018 *Steve Henson, reported by Kenneth R. Robinette
13019 <support@securenetterm.com>*
13021 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13022 routines: without these tracing memory leaks is very painful.
13023 Fix leaks in PKCS12 and PKCS7 routines.
13027 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13028 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13029 effectively meant GeneralizedTime would never be used. Now it
13030 is initialised to -1 but X509_time_adj() now has to check the value
13031 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13032 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13033 *Steve Henson, reported by Kenneth R. Robinette
13034 <support@securenetterm.com>*
13036 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13037 result in a zero length in the ASN1_INTEGER structure which was
13038 not consistent with the structure when d2i_ASN1_INTEGER() was used
13039 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13040 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13041 where it did not print out a minus for negative ASN1_INTEGER.
13045 * Add summary printout to ocsp utility. The various functions which
13046 convert status values to strings have been renamed to:
13047 OCSP_response_status_str(), OCSP_cert_status_str() and
13048 OCSP_crl_reason_str() and are no longer static. New options
13049 to verify nonce values and to disable verification. OCSP response
13050 printout format cleaned up.
13054 * Add additional OCSP certificate checks. These are those specified
13055 in RFC2560. This consists of two separate checks: the CA of the
13056 certificate being checked must either be the OCSP signer certificate
13057 or the issuer of the OCSP signer certificate. In the latter case the
13058 OCSP signer certificate must contain the OCSP signing extended key
13059 usage. This check is performed by attempting to match the OCSP
13060 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13061 in the OCSP_CERTID structures of the response.
13065 * Initial OCSP certificate verification added to OCSP_basic_verify()
13066 and related routines. This uses the standard OpenSSL certificate
13067 verify routines to perform initial checks (just CA validity) and
13068 to obtain the certificate chain. Then additional checks will be
13069 performed on the chain. Currently the root CA is checked to see
13070 if it is explicitly trusted for OCSP signing. This is used to set
13071 a root CA as a global signing root: that is any certificate that
13072 chains to that CA is an acceptable OCSP signing certificate.
13076 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13077 extensions from a separate configuration file.
13078 As when reading extensions from the main configuration file,
13079 the '-extensions ...' option may be used for specifying the
13082 *Massimiliano Pala <madwolf@comune.modena.it>*
13084 * New OCSP utility. Allows OCSP requests to be generated or
13085 read. The request can be sent to a responder and the output
13086 parsed, outputted or printed in text form. Not complete yet:
13087 still needs to check the OCSP response validity.
13091 * New subcommands for 'openssl ca':
13092 `openssl ca -status <serial>` prints the status of the cert with
13093 the given serial number (according to the index file).
13094 `openssl ca -updatedb` updates the expiry status of certificates
13097 *Massimiliano Pala <madwolf@comune.modena.it>*
13099 * New '-newreq-nodes' command option to CA.pl. This is like
13100 '-newreq', but calls 'openssl req' with the '-nodes' option
13101 so that the resulting key is not encrypted.
13103 *Damien Miller <djm@mindrot.org>*
13105 * New configuration for the GNU Hurd.
13107 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13109 * Initial code to implement OCSP basic response verify. This
13110 is currently incomplete. Currently just finds the signer's
13111 certificate and verifies the signature on the response.
13115 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13116 value of OPENSSLDIR. This is available via the new '-d' option
13117 to 'openssl version', and is also included in 'openssl version -a'.
13121 * Allowing defining memory allocation callbacks that will be given
13122 file name and line number information in additional arguments
13123 (a `const char*` and an int). The basic functionality remains, as
13124 well as the original possibility to just replace malloc(),
13125 realloc() and free() by functions that do not know about these
13126 additional arguments. To register and find out the current
13127 settings for extended allocation functions, the following
13128 functions are provided:
13130 CRYPTO_set_mem_ex_functions
13131 CRYPTO_set_locked_mem_ex_functions
13132 CRYPTO_get_mem_ex_functions
13133 CRYPTO_get_locked_mem_ex_functions
13135 These work the same way as CRYPTO_set_mem_functions and friends.
13136 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13137 extended allocation function is enabled.
13138 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13139 a conventional allocation function is enabled.
13141 *Richard Levitte, Bodo Moeller*
13143 * Finish off removing the remaining LHASH function pointer casts.
13144 There should no longer be any prototype-casting required when using
13145 the LHASH abstraction, and any casts that remain are "bugs". See
13146 the callback types and macros at the head of lhash.h for details
13147 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13151 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13152 If /dev/[u]random devices are not available or do not return enough
13153 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13155 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13156 /etc/entropy will be queried once each in this sequence, querying stops
13157 when enough entropy was collected without querying more sockets.
13161 * Change the Unix RAND_poll() variant to be able to poll several
13162 random devices, as specified by DEVRANDOM, until a sufficient amount
13163 of data has been collected. We spend at most 10 ms on each file
13164 (select timeout) and read in non-blocking mode. DEVRANDOM now
13165 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13166 (previously it was just the string "/dev/urandom"), so on typical
13167 platforms the 10 ms delay will never occur.
13168 Also separate out the Unix variant to its own file, rand_unix.c.
13169 For VMS, there's a currently-empty rand_vms.c.
13173 * Move OCSP client related routines to ocsp_cl.c. These
13174 provide utility functions which an application needing
13175 to issue a request to an OCSP responder and analyse the
13176 response will typically need: as opposed to those which an
13177 OCSP responder itself would need which will be added later.
13179 OCSP_request_sign() signs an OCSP request with an API similar
13180 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13181 response. OCSP_response_get1_basic() extracts basic response
13182 from response. OCSP_resp_find_status(): finds and extracts status
13183 information from an OCSP_CERTID structure (which will be created
13184 when the request structure is built). These are built from lower
13185 level functions which work on OCSP_SINGLERESP structures but
13186 won't normally be used unless the application wishes to examine
13187 extensions in the OCSP response for example.
13189 Replace nonce routines with a pair of functions.
13190 OCSP_request_add1_nonce() adds a nonce value and optionally
13191 generates a random value. OCSP_check_nonce() checks the
13192 validity of the nonce in an OCSP response.
13196 * Change function OCSP_request_add() to OCSP_request_add0_id().
13197 This doesn't copy the supplied OCSP_CERTID and avoids the
13198 need to free up the newly created id. Change return type
13199 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13200 This can then be used to add extensions to the request.
13201 Deleted OCSP_request_new(), since most of its functionality
13202 is now in OCSP_REQUEST_new() (and the case insensitive name
13203 clash) apart from the ability to set the request name which
13204 will be added elsewhere.
13208 * Update OCSP API. Remove obsolete extensions argument from
13209 various functions. Extensions are now handled using the new
13210 OCSP extension code. New simple OCSP HTTP function which
13211 can be used to send requests and parse the response.
13215 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13216 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13217 uses the special reorder version of SET OF to sort the attributes
13218 and reorder them to match the encoded order. This resolves a long
13219 standing problem: a verify on a PKCS7 structure just after signing
13220 it used to fail because the attribute order did not match the
13221 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13222 it uses the received order. This is necessary to tolerate some broken
13223 software that does not order SET OF. This is handled by encoding
13224 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13225 to produce the required SET OF.
13229 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13230 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13231 files to get correct declarations of the ASN.1 item variables.
13235 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13236 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13237 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13238 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13239 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13240 ASN1_ITEM and no wrapper functions.
13244 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13245 replace the old function pointer based I/O routines. Change most of
13246 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13250 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13251 lines, recognize more "algorithms" that can be deselected, and make
13252 it complain about algorithm deselection that isn't recognised.
13256 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13257 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13258 to use new functions. Add NO_ASN1_OLD which can be set to remove
13259 some old style ASN1 functions: this can be used to determine if old
13260 code will still work when these eventually go away.
13264 * New extension functions for OCSP structures, these follow the
13265 same conventions as certificates and CRLs.
13269 * New function X509V3_add1_i2d(). This automatically encodes and
13270 adds an extension. Its behaviour can be customised with various
13271 flags to append, replace or delete. Various wrappers added for
13272 certificates and CRLs.
13276 * Fix to avoid calling the underlying ASN1 print routine when
13277 an extension cannot be parsed. Correct a typo in the
13278 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13282 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13283 entries for variables.
13287 * Add functionality to `apps/openssl.c` for detecting locking
13288 problems: As the program is single-threaded, all we have
13289 to do is register a locking callback using an array for
13290 storing which locks are currently held by the program.
13294 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13295 SSL_get_ex_data_X509_STORE_idx(), which is used in
13296 ssl_verify_cert_chain() and thus can be called at any time
13297 during TLS/SSL handshakes so that thread-safety is essential.
13298 Unfortunately, the ex_data design is not at all suited
13299 for multi-threaded use, so it probably should be abolished.
13303 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13305 *Broadcom, tweaked and integrated by Geoff Thorpe*
13307 * Move common extension printing code to new function
13308 X509V3_print_extensions(). Reorganise OCSP print routines and
13309 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13313 * New function X509_signature_print() to remove duplication in some
13318 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13319 set (this was treated exactly the same as SET OF previously). This
13320 is used to reorder the STACK representing the structure to match the
13321 encoding. This will be used to get round a problem where a PKCS7
13322 structure which was signed could not be verified because the STACK
13323 order did not reflect the encoded order.
13327 * Reimplement the OCSP ASN1 module using the new code.
13331 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13332 for its ASN1 operations. The old style function pointers still exist
13333 for now but they will eventually go away.
13337 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13338 completely replaces the old ASN1 functionality with a table driven
13339 encoder and decoder which interprets an ASN1_ITEM structure describing
13340 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13341 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13342 has also been converted to the new form.
13346 * Change BN_mod_exp_recp so that negative moduli are tolerated
13347 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13348 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13349 for negative moduli.
13353 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13354 of not touching the result's sign bit.
13358 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13363 * Changed the LHASH code to use prototypes for callbacks, and created
13364 macros to declare and implement thin (optionally static) functions
13365 that provide type-safety and avoid function pointer casting for the
13366 type-specific callbacks.
13370 * Added Kerberos Cipher Suites to be used with TLS, as written in
13372 *Veers Staats <staatsvr@asc.hpc.mil>,
13373 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13375 * Reformat the FAQ so the different questions and answers can be divided
13376 in sections depending on the subject.
13380 * Have the zlib compression code load ZLIB.DLL dynamically under
13385 * New function BN_mod_sqrt for computing square roots modulo a prime
13386 (using the probabilistic Tonelli-Shanks algorithm unless
13387 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13388 be handled deterministically).
13390 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13392 * Make BN_mod_inverse faster by explicitly handling small quotients
13393 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13394 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13398 * New function BN_kronecker.
13402 * Fix BN_gcd so that it works on negative inputs; the result is
13403 positive unless both parameters are zero.
13404 Previously something reasonably close to an infinite loop was
13405 possible because numbers could be growing instead of shrinking
13406 in the implementation of Euclid's algorithm.
13410 * Fix BN_is_word() and BN_is_one() macros to take into account the
13411 sign of the number in question.
13413 Fix BN_is_word(a,w) to work correctly for w == 0.
13415 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13416 because its test if the absolute value of 'a' equals 'w'.
13417 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13418 it exists mostly for use in the implementations of BN_is_zero(),
13419 BN_is_one(), and BN_is_word().
13423 * New function BN_swap.
13427 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13428 the exponentiation functions are more likely to produce reasonable
13429 results on negative inputs.
13433 * Change BN_mod_mul so that the result is always non-negative.
13434 Previously, it could be negative if one of the factors was negative;
13435 I don't think anyone really wanted that behaviour.
13439 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13440 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13441 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13442 and add new functions:
13451 BN_mod_lshift1_quick
13453 BN_mod_lshift_quick
13455 These functions always generate non-negative results.
13457 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13458 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13460 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13461 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13462 be reduced modulo `m`.
13464 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13467 The following entry accidentally appeared in the CHANGES file
13468 distributed with OpenSSL 0.9.7. The modifications described in
13469 it do *not* apply to OpenSSL 0.9.7.
13471 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13472 was actually never needed) and in BN_mul(). The removal in BN_mul()
13473 required a small change in bn_mul_part_recursive() and the addition
13474 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13475 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13476 bn_sub_words() and bn_add_words() except they take arrays with
13482 * In 'openssl passwd', verify passwords read from the terminal
13483 unless the '-salt' option is used (which usually means that
13484 verification would just waste user's time since the resulting
13485 hash is going to be compared with some given password hash)
13486 or the new '-noverify' option is used.
13488 This is an incompatible change, but it does not affect
13489 non-interactive use of 'openssl passwd' (passwords on the command
13490 line, '-stdin' option, '-in ...' option) and thus should not
13491 cause any problems.
13495 * Remove all references to RSAref, since there's no more need for it.
13499 * Make DSO load along a path given through an environment variable
13500 (SHLIB_PATH) with shl_load().
13504 * Constify the ENGINE code as a result of BIGNUM constification.
13505 Also constify the RSA code and most things related to it. In a
13506 few places, most notable in the depth of the ASN.1 code, ugly
13507 casts back to non-const were required (to be solved at a later
13512 * Make it so the openssl application has all engines loaded by default.
13516 * Constify the BIGNUM routines a little more.
13520 * Add the following functions:
13522 ENGINE_load_cswift()
13524 ENGINE_load_atalla()
13525 ENGINE_load_nuron()
13526 ENGINE_load_builtin_engines()
13528 That way, an application can itself choose if external engines that
13529 are built-in in OpenSSL shall ever be used or not. The benefit is
13530 that applications won't have to be linked with libdl or other dso
13531 libraries unless it's really needed.
13533 Changed 'openssl engine' to load all engines on demand.
13534 Changed the engine header files to avoid the duplication of some
13535 declarations (they differed!).
13539 * 'openssl engine' can now list capabilities.
13543 * Better error reporting in 'openssl engine'.
13547 * Never call load_dh_param(NULL) in s_server.
13551 * Add engine application. It can currently list engines by name and
13552 identity, and test if they are actually available.
13556 * Improve RPM specification file by forcing symbolic linking and making
13557 sure the installed documentation is also owned by root.root.
13559 *Damien Miller <djm@mindrot.org>*
13561 * Give the OpenSSL applications more possibilities to make use of
13562 keys (public as well as private) handled by engines.
13566 * Add OCSP code that comes from CertCo.
13570 * Add VMS support for the Rijndael code.
13574 * Added untested support for Nuron crypto accelerator.
13578 * Add support for external cryptographic devices. This code was
13579 previously distributed separately as the "engine" branch.
13581 *Geoff Thorpe, Richard Levitte*
13583 * Rework the filename-translation in the DSO code. It is now possible to
13584 have far greater control over how a "name" is turned into a filename
13585 depending on the operating environment and any oddities about the
13586 different shared library filenames on each system.
13590 * Support threads on FreeBSD-elf in Configure.
13594 * Fix for SHA1 assembly problem with MASM: it produces
13595 warnings about corrupt line number information when assembling
13596 with debugging information. This is caused by the overlapping
13599 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13602 NCONF_get_number() has no error checking at all. As a replacement,
13603 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13604 promoted strongly. The old NCONF_get_number is kept around for
13605 binary backward compatibility.
13606 Make it possible for methods to load from something other than a BIO,
13607 by providing a function pointer that is given a name instead of a BIO.
13608 For example, this could be used to load configuration data from an
13613 * Fix for non blocking accept BIOs. Added new I/O special reason
13614 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13615 with non blocking I/O was not possible because no retry code was
13616 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13621 * Added the beginnings of Rijndael support.
13625 * Fix for bug in DirectoryString mask setting. Add support for
13626 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13627 to allow certificate printing to more controllable, additional
13628 'certopt' option to 'x509' to allow new printing options to be
13633 * Clean old EAY MD5 hack from e_os.h.
13637 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13639 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13640 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13642 *Joe Orton, Steve Henson*
13644 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13646 * Fix additional bug revealed by the NISCC test suite:
13648 Stop bug triggering large recursion when presented with
13649 certain ASN.1 tags ([CVE-2003-0851])
13653 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13655 * Fix various bugs revealed by running the NISCC test suite:
13657 Stop out of bounds reads in the ASN1 code when presented with
13658 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13660 If verify callback ignores invalid public key errors don't try to check
13661 certificate signature with the NULL public key.
13665 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13666 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13671 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13672 extra data after the compression methods not only for TLS 1.0
13673 but also for SSL 3.0 (as required by the specification).
13675 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13677 * Change X509_certificate_type() to mark the key as exported/exportable
13678 when it's 512 *bits* long, not 512 bytes.
13682 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13684 * Countermeasure against the Klima-Pokorny-Rosa extension of
13685 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13686 a protocol version number mismatch like a decryption error
13687 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13691 * Turn on RSA blinding by default in the default implementation
13692 to avoid a timing attack. Applications that don't want it can call
13693 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13694 They would be ill-advised to do so in most cases.
13696 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13698 * Change RSA blinding code so that it works when the PRNG is not
13699 seeded (in this case, the secret RSA exponent is abused as
13700 an unpredictable seed -- if it is not unpredictable, there
13701 is no point in blinding anyway). Make RSA blinding thread-safe
13702 by remembering the creator's thread ID in rsa->blinding and
13703 having all other threads use local one-time blinding factors
13704 (this requires more computation than sharing rsa->blinding, but
13705 avoids excessive locking; and if an RSA object is not shared
13706 between threads, blinding will still be very fast).
13710 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13712 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13713 via timing by performing a MAC computation even if incorrect
13714 block cipher padding has been found. This is a countermeasure
13715 against active attacks where the attacker has to distinguish
13716 between bad padding and a MAC verification error. ([CVE-2003-0078])
13718 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13719 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13720 Martin Vuagnoux (EPFL, Ilion)*
13722 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13724 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13725 memory from its contents. This is done with a counter that will
13726 place alternating values in each byte. This can be used to solve
13727 two issues: 1) the removal of calls to memset() by highly optimizing
13728 compilers, and 2) cleansing with other values than 0, since those can
13729 be read through on certain media, for example a swap space on disk.
13733 * Bugfix: client side session caching did not work with external caching,
13734 because the session->cipher setting was not restored when reloading
13735 from the external cache. This problem was masked, when
13736 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13737 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13741 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13742 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13744 *Zeev Lieber <zeev-l@yahoo.com>*
13746 * Undo an undocumented change introduced in 0.9.6e which caused
13747 repeated calls to OpenSSL_add_all_ciphers() and
13748 OpenSSL_add_all_digests() to be ignored, even after calling
13753 * Change the default configuration reader to deal with last line not
13754 being properly terminated.
13758 * Change X509_NAME_cmp() so it applies the special rules on handling
13759 DN values that are of type PrintableString, as well as RDNs of type
13760 emailAddress where the value has the type ia5String.
13762 *stefank@valicert.com via Richard Levitte*
13764 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13765 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13766 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13767 the bitwise-OR of the two for use by the majority of applications
13768 wanting this behaviour, and update the docs. The documented
13769 behaviour and actual behaviour were inconsistent and had been
13770 changing anyway, so this is more a bug-fix than a behavioural
13773 *Geoff Thorpe, diagnosed by Nadav Har'El*
13775 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13776 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13780 * Fix initialization code race conditions in
13781 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13782 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13783 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13784 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13785 ssl2_get_cipher_by_char(),
13786 ssl3_get_cipher_by_char().
13788 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13790 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13791 the cached sessions are flushed, as the remove_cb() might use ex_data
13792 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13793 (see [openssl.org #212]).
13795 *Geoff Thorpe, Lutz Jaenicke*
13797 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13798 length, instead of the encoding length to d2i_ASN1_OBJECT.
13802 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13804 * [In 0.9.6g-engine release:]
13805 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13807 *Lynn Gazis <lgazis@rainbow.com>*
13809 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13811 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13812 and get fix the header length calculation.
13813 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13814 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13816 * Use proper error handling instead of 'assertions' in buffer
13817 overflow checks added in 0.9.6e. This prevents DoS (the
13818 assertions could call abort()).
13820 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13822 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13824 * Add various sanity checks to asn1_get_length() to reject
13825 the ASN1 length bytes if they exceed sizeof(long), will appear
13826 negative or the content length exceeds the length of the
13829 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13831 * Fix cipher selection routines: ciphers without encryption had no flags
13832 for the cipher strength set and where therefore not handled correctly
13833 by the selection routines (PR #130).
13837 * Fix EVP_dsa_sha macro.
13842 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13843 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13844 that was added in OpenSSL 0.9.6d.
13846 As the countermeasure turned out to be incompatible with some
13847 broken SSL implementations, the new option is part of SSL_OP_ALL.
13848 SSL_OP_ALL is usually employed when compatibility with weird SSL
13849 implementations is desired (e.g. '-bugs' option to 's_client' and
13850 's_server'), so the new option is automatically set in many
13855 * Changes in security patch:
13857 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13858 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13859 Air Force Materiel Command, USAF, under agreement number
13862 * Add various sanity checks to asn1_get_length() to reject
13863 the ASN1 length bytes if they exceed sizeof(long), will appear
13864 negative or the content length exceeds the length of the
13865 supplied buffer. ([CVE-2002-0659])
13867 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13869 * Assertions for various potential buffer overflows, not known to
13870 happen in practice.
13872 *Ben Laurie (CHATS)*
13874 * Various temporary buffers to hold ASCII versions of integers were
13875 too small for 64 bit platforms. ([CVE-2002-0655])
13876 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13878 * Remote buffer overflow in SSL3 protocol - an attacker could
13879 supply an oversized session ID to a client. ([CVE-2002-0656])
13881 *Ben Laurie (CHATS)*
13883 * Remote buffer overflow in SSL2 protocol - an attacker could
13884 supply an oversized client master key. ([CVE-2002-0656])
13886 *Ben Laurie (CHATS)*
13888 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13890 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13891 encoded as NULL) with id-dsa-with-sha1.
13893 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13895 * Check various `X509_...()` return values in `apps/req.c`.
13897 *Nils Larsch <nla@trustcenter.de>*
13899 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13900 an end-of-file condition would erroneously be flagged, when the CRLF
13901 was just at the end of a processed block. The bug was discovered when
13902 processing data through a buffering memory BIO handing the data to a
13903 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13904 <ptsekov@syntrex.com> and Nedelcho Stanev.
13908 * Implement a countermeasure against a vulnerability recently found
13909 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13910 before application data chunks to avoid the use of known IVs
13911 with data potentially chosen by the attacker.
13915 * Fix length checks in ssl3_get_client_hello().
13919 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13920 to prevent ssl3_read_internal() from incorrectly assuming that
13921 ssl3_read_bytes() found application data while handshake
13922 processing was enabled when in fact s->s3->in_read_app_data was
13923 merely automatically cleared during the initial handshake.
13925 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13927 * Fix object definitions for Private and Enterprise: they were not
13928 recognized in their shortname (=lowercase) representation. Extend
13929 obj_dat.pl to issue an error when using undefined keywords instead
13930 of silently ignoring the problem (Svenning Sorensen
13931 <sss@sss.dnsalias.net>).
13935 * Fix DH_generate_parameters() so that it works for 'non-standard'
13936 generators, i.e. generators other than 2 and 5. (Previously, the
13937 code did not properly initialise the 'add' and 'rem' values to
13938 BN_generate_prime().)
13940 In the new general case, we do not insist that 'generator' is
13941 actually a primitive root: This requirement is rather pointless;
13942 a generator of the order-q subgroup is just as good, if not
13947 * Map new X509 verification errors to alerts. Discovered and submitted by
13948 Tom Wu <tom@arcot.com>.
13952 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13953 returning non-zero before the data has been completely received
13954 when using non-blocking I/O.
13956 *Bodo Moeller; problem pointed out by John Hughes*
13958 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13960 *Ben Laurie, Lutz Jaenicke*
13962 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13963 Yoram Zahavi <YoramZ@gilian.com>).
13967 * Add information about CygWin 1.3 and on, and preserve proper
13968 configuration for the versions before that.
13970 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13972 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13973 check whether we deal with a copy of a session and do not delete from
13974 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13975 <izhar@checkpoint.com>.
13979 * Do not store session data into the internal session cache, if it
13980 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13981 flag is set). Proposed by Aslam <aslam@funk.com>.
13985 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13990 * [In 0.9.6d-engine release:]
13991 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13993 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13995 * Add the configuration target linux-s390x.
13997 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13999 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14000 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14001 variable as an indication that a ClientHello message has been
14002 received. As the flag value will be lost between multiple
14003 invocations of ssl3_accept when using non-blocking I/O, the
14004 function may not be aware that a handshake has actually taken
14005 place, thus preventing a new session from being added to the
14008 To avoid this problem, we now set s->new_session to 2 instead of
14009 using a local variable.
14011 *Lutz Jaenicke, Bodo Moeller*
14013 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14014 if the SSL_R_LENGTH_MISMATCH error is detected.
14016 *Geoff Thorpe, Bodo Moeller*
14018 * New 'shared_ldflag' column in Configure platform table.
14022 * Fix EVP_CIPHER_mode macro.
14024 *"Dan S. Camper" <dan@bti.net>*
14026 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14027 type, we must throw them away by setting rr->length to 0.
14029 *D P Chang <dpc@qualys.com>*
14031 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14033 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14034 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14035 worked incorrectly for those cases where range = `10..._2` and
14036 `3*range` is two bits longer than range.)
14040 * Only add signing time to PKCS7 structures if it is not already
14045 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14046 OBJ_ld_ce should be OBJ_id_ce.
14047 Also some ip-pda OIDs in crypto/objects/objects.txt were
14048 incorrect (cf. RFC 3039).
14050 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14052 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14053 returns early because it has nothing to do.
14055 *Andy Schneider <andy.schneider@bjss.co.uk>*
14057 * [In 0.9.6c-engine release:]
14058 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14060 *Andy Schneider <andy.schneider@bjss.co.uk>*
14062 * [In 0.9.6c-engine release:]
14063 Add support for Cryptographic Appliance's keyserver technology.
14064 (Use engine 'keyclient')
14066 *Cryptographic Appliances and Geoff Thorpe*
14068 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14069 is called via tools/c89.sh because arguments have to be
14070 rearranged (all '-L' options must appear before the first object
14073 *Richard Shapiro <rshapiro@abinitio.com>*
14075 * [In 0.9.6c-engine release:]
14076 Add support for Broadcom crypto accelerator cards, backported
14079 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14081 * [In 0.9.6c-engine release:]
14082 Add support for SureWare crypto accelerator cards from
14083 Baltimore Technologies. (Use engine 'sureware')
14085 *Baltimore Technologies and Mark Cox*
14087 * [In 0.9.6c-engine release:]
14088 Add support for crypto accelerator cards from Accelerated
14089 Encryption Processing, www.aep.ie. (Use engine 'aep')
14091 *AEP Inc. and Mark Cox*
14093 * Add a configuration entry for gcc on UnixWare.
14095 *Gary Benson <gbenson@redhat.com>*
14097 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14098 messages are stored in a single piece (fixed-length part and
14099 variable-length part combined) and fix various bugs found on the way.
14103 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14104 instead. BIO_gethostbyname() does not know what timeouts are
14105 appropriate, so entries would stay in cache even when they have
14107 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14109 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14110 faced with a pathologically small ClientHello fragment that does
14111 not contain client_version: Instead of aborting with an error,
14112 simply choose the highest available protocol version (i.e.,
14113 TLS 1.0 unless it is disabled). In practice, ClientHello
14114 messages are never sent like this, but this change gives us
14115 strictly correct behaviour at least for TLS.
14119 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14120 never resets s->method to s->ctx->method when called from within
14121 one of the SSL handshake functions.
14123 *Bodo Moeller; problem pointed out by Niko Baric*
14125 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14126 (sent using the client's version number) if client_version is
14127 smaller than the protocol version in use. Also change
14128 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14129 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14130 the client will at least see that alert.
14134 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14139 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14140 client receives HelloRequest while in a handshake.
14142 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14144 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14145 should end in 'break', not 'goto end' which circumvents various
14146 cleanups done in state SSL_ST_OK. But session related stuff
14147 must be disabled for SSL_ST_OK in the case that we just sent a
14150 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14151 before just sending a HelloRequest.
14153 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14155 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14156 reveal whether illegal block cipher padding was found or a MAC
14157 verification error occurred. (Neither SSLerr() codes nor alerts
14158 are directly visible to potential attackers, but the information
14159 may leak via logfiles.)
14161 Similar changes are not required for the SSL 2.0 implementation
14162 because the number of padding bytes is sent in clear for SSL 2.0,
14163 and the extra bytes are just ignored. However ssl/s2_pkt.c
14164 failed to verify that the purported number of padding bytes is in
14169 * Add OpenUNIX-8 support including shared libraries
14170 (Boyd Lynn Gerber <gerberb@zenez.com>).
14174 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14175 'wristwatch attack' using huge encoding parameters (cf.
14176 James H. Manger's CRYPTO 2001 paper). Note that the
14177 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14178 encoding parameters and hence was not vulnerable.
14182 * BN_sqr() bug fix.
14184 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14186 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14187 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14188 followed by modular reduction.
14190 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14192 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14193 equivalent based on BN_pseudo_rand() instead of BN_rand().
14197 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14198 This function was broken, as the check for a new client hello message
14199 to handle SGC did not allow these large messages.
14200 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14204 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14208 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14209 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14213 * Rework the configuration and shared library support for Tru64 Unix.
14214 The configuration part makes use of modern compiler features and
14215 still retains old compiler behavior for those that run older versions
14216 of the OS. The shared library support part includes a variant that
14217 uses the RPATH feature, and is available through the special
14218 configuration target "alpha-cc-rpath", which will never be selected
14221 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14223 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14224 with the same message size as in ssl3_get_certificate_request().
14225 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14226 messages might inadvertently be reject as too long.
14228 *Petr Lampa <lampa@fee.vutbr.cz>*
14230 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14234 * Modified SSL library such that the verify_callback that has been set
14235 specifically for an SSL object with SSL_set_verify() is actually being
14236 used. Before the change, a verify_callback set with this function was
14237 ignored and the verify_callback() set in the SSL_CTX at the time of
14238 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14239 to allow the necessary settings.
14243 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14244 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14245 done automatically (in contradiction to the requirements of the C
14246 standard). This made problems when used from OpenSSH.
14250 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14251 dh->length and always used
14253 BN_rand_range(priv_key, dh->p).
14255 BN_rand_range() is not necessary for Diffie-Hellman, and this
14256 specific range makes Diffie-Hellman unnecessarily inefficient if
14257 dh->length (recommended exponent length) is much smaller than the
14258 length of dh->p. We could use BN_rand_range() if the order of
14259 the subgroup was stored in the DH structure, but we only have
14264 BN_rand(priv_key, l, ...)
14266 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14273 RSA_eay_public_encrypt
14274 RSA_eay_private_decrypt
14275 RSA_eay_private_encrypt (signing)
14276 RSA_eay_public_decrypt (signature verification)
14278 (default implementations for RSA_public_encrypt,
14279 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14280 always reject numbers >= n.
14284 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14285 to synchronize access to 'locking_thread'. This is necessary on
14286 systems where access to 'locking_thread' (an 'unsigned long'
14287 variable) is not atomic.
14291 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14292 *before* setting the 'crypto_lock_rand' flag. The previous code had
14293 a race condition if 0 is a valid thread ID.
14295 *Travis Vitek <vitek@roguewave.com>*
14297 * Add support for shared libraries under Irix.
14299 *Albert Chin-A-Young <china@thewrittenword.com>*
14301 * Add configuration option to build on Linux on both big-endian and
14302 little-endian MIPS.
14304 *Ralf Baechle <ralf@uni-koblenz.de>*
14306 * Add the possibility to create shared libraries on HP-UX.
14310 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14312 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14313 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14314 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14315 PRNG state recovery was possible based on the output of
14316 one PRNG request appropriately sized to gain knowledge on
14317 'md' followed by enough consecutive 1-byte PRNG requests
14318 to traverse all of 'state'.
14320 1. When updating 'md_local' (the current thread's copy of 'md')
14321 during PRNG output generation, hash all of the previous
14322 'md_local' value, not just the half used for PRNG output.
14324 2. Make the number of bytes from 'state' included into the hash
14325 independent from the number of PRNG bytes requested.
14327 The first measure alone would be sufficient to avoid
14328 Markku-Juhani's attack. (Actually it had never occurred
14329 to me that the half of 'md_local' used for chaining was the
14330 half from which PRNG output bytes were taken -- I had always
14331 assumed that the secret half would be used.) The second
14332 measure makes sure that additional data from 'state' is never
14333 mixed into 'md_local' in small portions; this heuristically
14334 further strengthens the PRNG.
14338 * Fix crypto/bn/asm/mips3.s.
14342 * When only the key is given to "enc", the IV is undefined. Print out
14343 an error message in this case.
14347 * Handle special case when X509_NAME is empty in X509 printing routines.
14351 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14352 positive and less than q.
14356 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14357 used: it isn't thread safe and the add_lock_callback should handle
14360 *Paul Rose <Paul.Rose@bridge.com>*
14362 * Verify that incoming data obeys the block size in
14363 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14369 *Ulf Möller, Bodo Möller*
14371 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14372 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14373 when fixing the server behaviour for backwards-compatible 'client
14374 hello' messages. (Note that the attack is impractical against
14375 SSL 3.0 and TLS 1.0 anyway because length and version checking
14376 means that the probability of guessing a valid ciphertext is
14377 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14380 Before 0.9.5, the countermeasure (hide the error by generating a
14381 random 'decryption result') did not work properly because
14382 ERR_clear_error() was missing, meaning that SSL_get_error() would
14383 detect the supposedly ignored error.
14385 Both problems are now fixed.
14389 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14390 (previously it was 1024).
14394 * Fix for compatibility mode trust settings: ignore trust settings
14395 unless some valid trust or reject settings are present.
14399 * Fix for blowfish EVP: its a variable length cipher.
14403 * Fix various bugs related to DSA S/MIME verification. Handle missing
14404 parameters in DSA public key structures and return an error in the
14405 DSA routines if parameters are absent.
14409 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14410 in the current directory if neither $RANDFILE nor $HOME was set.
14411 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14412 caused some confusion to Windows users who haven't defined $HOME.
14413 Thus RAND_file_name() is changed again: e_os.h can define a
14414 DEFAULT_HOME, which will be used if $HOME is not set.
14415 For Windows, we use "C:"; on other platforms, we still require
14416 environment variables.
14418 * Move 'if (!initialized) RAND_poll()' into regions protected by
14419 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14420 having multiple threads call RAND_poll() concurrently.
14424 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14425 combination of a flag and a thread ID variable.
14426 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14427 flag), *other* threads can enter ssleay_add_bytes without obeying
14428 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14429 that they do not hold after the first thread unsets add_do_not_lock).
14433 * Change bctest again: '-x' expressions are not available in all
14434 versions of 'test'.
14438 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14440 * Fix a couple of memory leaks in PKCS7_dataDecode()
14442 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14444 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14445 the default extension for executables, if any. Also, make the perl
14446 scripts that use symlink() to test if it really exists and use "cp"
14447 if it doesn't. All this made OpenSSL compilable and installable in
14452 * Fix for asn1_GetSequence() for indefinite length constructed data.
14453 If SEQUENCE is length is indefinite just set c->slen to the total
14454 amount of data available.
14456 *Steve Henson, reported by shige@FreeBSD.org*
14458 *This change does not apply to 0.9.7.*
14460 * Change bctest to avoid here-documents inside command substitution
14461 (workaround for FreeBSD /bin/sh bug).
14462 For compatibility with Ultrix, avoid shell functions (introduced
14463 in the bctest version that searches along $PATH).
14467 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14468 with des_encrypt() defined on some operating systems, like Solaris
14473 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14474 On the Importance of Eliminating Errors in Cryptographic
14475 Computations, J. Cryptology 14 (2001) 2, 101-119,
14476 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14480 * MIPS assembler BIGNUM division bug fix.
14484 * Disabled incorrect Alpha assembler code.
14488 * Fix PKCS#7 decode routines so they correctly update the length
14489 after reading an EOC for the EXPLICIT tag.
14493 *This change does not apply to 0.9.7.*
14495 * Fix bug in PKCS#12 key generation routines. This was triggered
14496 if a 3DES key was generated with a 0 initial byte. Include
14497 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14498 (but broken) behaviour.
14502 * Enhance bctest to search for a working bc along $PATH and print
14505 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14507 * Fix memory leaks in err.c: free err_data string if necessary;
14508 don't write to the wrong index in ERR_set_error_data.
14512 * Implement ssl23_peek (analogous to ssl23_read), which previously
14517 * Replace rdtsc with `_emit` statements for VC++ version 5.
14519 *Jeremy Cooper <jeremy@baymoo.org>*
14521 * Make it possible to reuse SSLv2 sessions.
14525 * In copy_email() check for >= 0 as a return value for
14526 X509_NAME_get_index_by_NID() since 0 is a valid index.
14528 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14530 * Avoid coredump with unsupported or invalid public keys by checking if
14531 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14532 PKCS7_verify() fails with non detached data.
14536 * Don't use getenv in library functions when run as setuid/setgid.
14537 New function OPENSSL_issetugid().
14541 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14542 due to incorrect handling of multi-threading:
14544 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14546 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14548 3. Count how many times MemCheck_off() has been called so that
14549 nested use can be treated correctly. This also avoids
14550 inband-signalling in the previous code (which relied on the
14551 assumption that thread ID 0 is impossible).
14555 * Add "-rand" option also to s_client and s_server.
14559 * Fix CPU detection on Irix 6.x.
14560 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14561 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14563 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14568 *This change does not apply to 0.9.7.*
14570 * Use the cached encoding of an X509_NAME structure rather than
14571 copying it. This is apparently the reason for the libsafe "errors"
14572 but the code is actually correct.
14576 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14577 Bleichenbacher's DSA attack.
14578 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14579 to be set and top=0 forces the highest bit to be set; top=-1 is new
14580 and leaves the highest bit random.
14582 *Ulf Moeller, Bodo Moeller*
14584 * In the `NCONF_...`-based implementations for `CONF_...` queries
14585 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14586 a temporary CONF structure with the data component set to NULL
14587 (which gives segmentation faults in lh_retrieve).
14588 Instead, use NULL for the CONF pointer in CONF_get_string and
14589 CONF_get_number (which may use environment variables) and directly
14590 return NULL from CONF_get_section.
14594 * Fix potential buffer overrun for EBCDIC.
14598 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14599 keyUsage if basicConstraints absent for a CA.
14603 * Make SMIME_write_PKCS7() write mail header values with a format that
14604 is more generally accepted (no spaces before the semicolon), since
14605 some programs can't parse those values properly otherwise. Also make
14606 sure BIO's that break lines after each write do not create invalid
14611 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14612 macros previously used would not encode an empty SEQUENCE OF
14613 and break the signature.
14617 *This change does not apply to 0.9.7.*
14619 * Zero the premaster secret after deriving the master secret in
14624 * Add some EVP_add_digest_alias registrations (as found in
14625 OpenSSL_add_all_digests()) to SSL_library_init()
14626 aka OpenSSL_add_ssl_algorithms(). This provides improved
14627 compatibility with peers using X.509 certificates
14628 with unconventional AlgorithmIdentifier OIDs.
14632 * Fix for Irix with NO_ASM.
14634 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14636 * ./config script fixes.
14638 *Ulf Moeller, Richard Levitte*
14640 * Fix 'openssl passwd -1'.
14644 * Change PKCS12_key_gen_asc() so it can cope with non null
14645 terminated strings whose length is passed in the passlen
14646 parameter, for example from PEM callbacks. This was done
14647 by adding an extra length parameter to asc2uni().
14649 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14651 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14652 call failed, free the DSA structure.
14656 * Fix to uni2asc() to cope with zero length Unicode strings.
14657 These are present in some PKCS#12 files.
14661 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14662 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14663 when writing a 32767 byte record.
14665 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14667 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14668 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14670 (RSA objects have a reference count access to which is protected
14671 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14672 so they are meant to be shared between threads.)
14673 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14674 "Reddie, Steven" <Steven.Reddie@ca.com>*
14676 * Fix a deadlock in CRYPTO_mem_leaks().
14680 * Use better test patterns in bntest.
14684 * rand_win.c fix for Borland C.
14688 * BN_rshift bugfix for n == 0.
14692 * Add a 'bctest' script that checks for some known 'bc' bugs
14693 so that 'make test' does not abort just because 'bc' is broken.
14697 * Store verify_result within SSL_SESSION also for client side to
14698 avoid potential security hole. (Re-used sessions on the client side
14699 always resulted in verify_result==X509_V_OK, not using the original
14700 result of the server certificate verification.)
14704 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14705 SSL3_RT_APPLICATION_DATA, return 0.
14706 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14711 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14712 releases, have been re-implemented by renaming the previous
14713 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14714 and ssl3_read_internal, respectively, and adding 'peek' parameters
14715 to them. The new ssl[23]_{read,peek} functions are calls to
14716 ssl[23]_read_internal with the 'peek' flag set appropriately.
14717 A 'peek' parameter has also been added to ssl3_read_bytes, which
14718 does the actual work for ssl3_read_internal.
14722 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14723 the method-specific "init()" handler. Also clean up ex_data after
14724 calling the method-specific "finish()" handler. Previously, this was
14725 happening the other way round.
14729 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14730 The previous value, 12, was not always sufficient for BN_mod_exp().
14734 * Make sure that shared libraries get the internal name engine with
14735 the full version number and not just 0. This should mark the
14736 shared libraries as not backward compatible. Of course, this should
14737 be changed again when we can guarantee backward binary compatibility.
14741 * Fix typo in get_cert_by_subject() in by_dir.c
14743 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14745 * Rework the system to generate shared libraries:
14747 - Make note of the expected extension for the shared libraries and
14748 if there is a need for symbolic links from for example libcrypto.so.0
14749 to libcrypto.so.0.9.7. There is extended info in Configure for
14752 - Make as few rebuilds of the shared libraries as possible.
14754 - Still avoid linking the OpenSSL programs with the shared libraries.
14756 - When installing, install the shared libraries separately from the
14761 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14763 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14764 and not in SSL_clear because the latter is also used by the
14765 accept/connect functions; previously, the settings made by
14766 SSL_set_read_ahead would be lost during the handshake.
14768 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14770 * Correct util/mkdef.pl to be selective about disabled algorithms.
14771 Previously, it would create entries for disabled algorithms no
14776 * Added several new manual pages for SSL_* function.
14780 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14782 * In ssl23_get_client_hello, generate an error message when faced
14783 with an initial SSL 3.0/TLS record that is too small to contain the
14784 first two bytes of the ClientHello message, i.e. client_version.
14785 (Note that this is a pathologic case that probably has never happened
14786 in real life.) The previous approach was to use the version number
14787 from the record header as a substitute; but our protocol choice
14788 should not depend on that one because it is not authenticated
14789 by the Finished messages.
14793 * More robust randomness gathering functions for Windows.
14795 *Jeffrey Altman <jaltman@columbia.edu>*
14797 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14798 not set then we don't setup the error code for issuer check errors
14799 to avoid possibly overwriting other errors which the callback does
14800 handle. If an application does set the flag then we assume it knows
14801 what it is doing and can handle the new informational codes
14806 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14807 a general "ANY" type, as such it should be able to decode anything
14808 including tagged types. However it didn't check the class so it would
14809 wrongly interpret tagged types in the same way as their universal
14810 counterpart and unknown types were just rejected. Changed so that the
14811 tagged and unknown types are handled in the same way as a SEQUENCE:
14812 that is the encoding is stored intact. There is also a new type
14813 "V_ASN1_OTHER" which is used when the class is not universal, in this
14814 case we have no idea what the actual type is so we just lump them all
14819 * On VMS, stdout may very well lead to a file that is written to
14820 in a record-oriented fashion. That means that every write() will
14821 write a separate record, which will be read separately by the
14822 programs trying to read from it. This can be very confusing.
14824 The solution is to put a BIO filter in the way that will buffer
14825 text until a linefeed is reached, and then write everything a
14826 line at a time, so every record written will be an actual line,
14827 not chunks of lines and not (usually doesn't happen, but I've
14828 seen it once) several lines in one record. BIO_f_linebuffer() is
14831 Currently, it's a VMS-only method, because that's where it has
14832 been tested well enough.
14836 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14837 it can return incorrect results.
14838 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14839 but it was in 0.9.6-beta[12].)
14843 * Disable the check for content being present when verifying detached
14844 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14845 include zero length content when signing messages.
14849 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14850 BIO_ctrl (for BIO pairs).
14854 * Add DSO method for VMS.
14858 * Bug fix: Montgomery multiplication could produce results with the
14863 * Add RPM specification openssl.spec and modify it to build three
14864 packages. The default package contains applications, application
14865 documentation and run-time libraries. The devel package contains
14866 include files, static libraries and function documentation. The
14867 doc package contains the contents of the doc directory. The original
14868 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14872 * Add a large number of documentation files for many SSL routines.
14874 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14876 * Add a configuration entry for Sony News 4.
14878 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14880 * Don't set the two most significant bits to one when generating a
14881 random number < q in the DSA library.
14885 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14886 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14887 the underlying transport is blocking) if a handshake took place.
14888 (The default behaviour is needed by applications such as s_client
14889 and s_server that use select() to determine when to use SSL_read;
14890 but for applications that know in advance when to expect data, it
14891 just makes things more complicated.)
14895 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14900 * Add a few more EBCDIC conditionals that make `req` and `x509`
14901 work better on such systems.
14903 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14905 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14906 Update PKCS12_parse() so it copies the friendlyName and the
14907 keyid to the certificates aux info.
14911 * Fix bug in PKCS7_verify() which caused an infinite loop
14912 if there was more than one signature.
14914 *Sven Uszpelkat <su@celocom.de>*
14916 * Major change in util/mkdef.pl to include extra information
14917 about each symbol, as well as presenting variables as well
14918 as functions. This change means that there's n more need
14919 to rebuild the .num files when some algorithms are excluded.
14923 * Allow the verify time to be set by an application,
14924 rather than always using the current time.
14928 * Phase 2 verify code reorganisation. The certificate
14929 verify code now looks up an issuer certificate by a
14930 number of criteria: subject name, authority key id
14931 and key usage. It also verifies self signed certificates
14932 by the same criteria. The main comparison function is
14933 X509_check_issued() which performs these checks.
14935 Lot of changes were necessary in order to support this
14936 without completely rewriting the lookup code.
14938 Authority and subject key identifier are now cached.
14940 The LHASH 'certs' is X509_STORE has now been replaced
14941 by a STACK_OF(X509_OBJECT). This is mainly because an
14942 LHASH can't store or retrieve multiple objects with
14943 the same hash value.
14945 As a result various functions (which were all internal
14946 use only) have changed to handle the new X509_STORE
14947 structure. This will break anything that messed round
14948 with X509_STORE internally.
14950 The functions X509_STORE_add_cert() now checks for an
14951 exact match, rather than just subject name.
14953 The X509_STORE API doesn't directly support the retrieval
14954 of multiple certificates matching a given criteria, however
14955 this can be worked round by performing a lookup first
14956 (which will fill the cache with candidate certificates)
14957 and then examining the cache for matches. This is probably
14958 the best we can do without throwing out X509_LOOKUP
14959 entirely (maybe later...).
14961 The X509_VERIFY_CTX structure has been enhanced considerably.
14963 All certificate lookup operations now go via a get_issuer()
14964 callback. Although this currently uses an X509_STORE it
14965 can be replaced by custom lookups. This is a simple way
14966 to bypass the X509_STORE hackery necessary to make this
14967 work and makes it possible to use more efficient techniques
14968 in future. A very simple version which uses a simple
14969 STACK for its trusted certificate store is also provided
14970 using X509_STORE_CTX_trusted_stack().
14972 The verify_cb() and verify() callbacks now have equivalents
14973 in the X509_STORE_CTX structure.
14975 X509_STORE_CTX also has a 'flags' field which can be used
14976 to customise the verify behaviour.
14980 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14981 excludes S/MIME capabilities.
14985 * When a certificate request is read in keep a copy of the
14986 original encoding of the signed data and use it when outputting
14987 again. Signatures then use the original encoding rather than
14988 a decoded, encoded version which may cause problems if the
14989 request is improperly encoded.
14993 * For consistency with other BIO_puts implementations, call
14994 buffer_write(b, ...) directly in buffer_puts instead of calling
14997 In BIO_puts, increment b->num_write as in BIO_write.
14999 *Peter.Sylvester@EdelWeb.fr*
15001 * Fix BN_mul_word for the case where the word is 0. (We have to use
15002 BN_zero, we may not return a BIGNUM with an array consisting of
15003 words set to zero.)
15007 * Avoid calling abort() from within the library when problems are
15008 detected, except if preprocessor symbols have been defined
15009 (such as REF_CHECK, BN_DEBUG etc.).
15013 * New openssl application 'rsautl'. This utility can be
15014 used for low level RSA operations. DER public key
15015 BIO/fp routines also added.
15019 * New Configure entry and patches for compiling on QNX 4.
15021 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15023 * A demo state-machine implementation was sponsored by
15024 Nuron (<http://www.nuron.com/>) and is now available in
15025 demos/state_machine.
15029 * New options added to the 'dgst' utility for signature
15030 generation and verification.
15034 * Unrecognized PKCS#7 content types are now handled via a
15035 catch all ASN1_TYPE structure. This allows unsupported
15036 types to be stored as a "blob" and an application can
15037 encode and decode it manually.
15041 * Fix various signed/unsigned issues to make a_strex.c
15042 compile under VC++.
15044 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15046 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15047 length if passed a buffer. ASN1_INTEGER_to_BN failed
15048 if passed a NULL BN and its argument was negative.
15050 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15052 * Modification to PKCS#7 encoding routines to output definite
15053 length encoding. Since currently the whole structures are in
15054 memory there's not real point in using indefinite length
15055 constructed encoding. However if OpenSSL is compiled with
15056 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15060 * Added BIO_vprintf() and BIO_vsnprintf().
15064 * Added more prefixes to parse for in the strings written
15065 through a logging bio, to cover all the levels that are available
15066 through syslog. The prefixes are now:
15068 PANIC, EMERG, EMR => LOG_EMERG
15069 ALERT, ALR => LOG_ALERT
15070 CRIT, CRI => LOG_CRIT
15071 ERROR, ERR => LOG_ERR
15072 WARNING, WARN, WAR => LOG_WARNING
15073 NOTICE, NOTE, NOT => LOG_NOTICE
15074 INFO, INF => LOG_INFO
15075 DEBUG, DBG => LOG_DEBUG
15077 and as before, if none of those prefixes are present at the
15078 beginning of the string, LOG_ERR is chosen.
15080 On Win32, the `LOG_*` levels are mapped according to this:
15082 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15083 LOG_WARNING => EVENTLOG_WARNING_TYPE
15084 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15088 * Made it possible to reconfigure with just the configuration
15089 argument "reconf" or "reconfigure". The command line arguments
15090 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15091 and are retrieved from there when reconfiguring.
15097 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15099 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15103 * The obj_dat.pl script was messing up the sorting of object
15104 names. The reason was that it compared the quoted version
15105 of strings as a result "OCSP" > "OCSP Signing" because
15106 " > SPACE. Changed script to store unquoted versions of
15107 names and add quotes on output. It was also omitting some
15108 names from the lookup table if they were given a default
15109 value (that is if SN is missing it is given the same
15110 value as LN and vice versa), these are now added on the
15111 grounds that if an object has a name we should be able to
15112 look it up. Finally added warning output when duplicate
15113 short or long names are found.
15117 * Changes needed for Tandem NSK.
15119 *Scott Uroff <scott@xypro.com>*
15121 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15122 RSA_padding_check_SSLv23(), special padding was never detected
15123 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15124 version rollback attacks was not effective.
15126 In s23_clnt.c, don't use special rollback-attack detection padding
15127 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15128 client; similarly, in s23_srvr.c, don't do the rollback check if
15129 SSL 2.0 is the only protocol enabled in the server.
15133 * Make it possible to get hexdumps of unprintable data with 'openssl
15134 asn1parse'. By implication, the functions ASN1_parse_dump() and
15135 BIO_dump_indent() are added.
15139 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15140 these print out strings and name structures based on various
15141 flags including RFC2253 support and proper handling of
15142 multibyte characters. Added options to the 'x509' utility
15143 to allow the various flags to be set.
15147 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15148 Also change the functions X509_cmp_current_time() and
15149 X509_gmtime_adj() work with an ASN1_TIME structure,
15150 this will enable certificates using GeneralizedTime in validity
15151 dates to be checked.
15155 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15156 negative public key encodings) on by default,
15157 NO_NEG_PUBKEY_BUG can be set to disable it.
15161 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15162 content octets. An i2c_ASN1_OBJECT is unnecessary because
15163 the encoding can be trivially obtained from the structure.
15167 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15168 not read locks (`CRYPTO_r_[un]lock`).
15172 * A first attempt at creating official support for shared
15173 libraries through configuration. I've kept it so the
15174 default is static libraries only, and the OpenSSL programs
15175 are always statically linked for now, but there are
15176 preparations for dynamic linking in place.
15177 This has been tested on Linux and Tru64.
15181 * Randomness polling function for Win9x, as described in:
15182 Peter Gutmann, Software Generation of Practically Strong
15187 * Fix so PRNG is seeded in req if using an already existing
15192 * New options to smime application. -inform and -outform
15193 allow alternative formats for the S/MIME message including
15194 PEM and DER. The -content option allows the content to be
15195 specified separately. This should allow things like Netscape
15196 form signing output easier to verify.
15200 * Fix the ASN1 encoding of tags using the 'long form'.
15204 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15205 STRING types. These convert content octets to and from the
15206 underlying type. The actual tag and length octets are
15207 already assumed to have been read in and checked. These
15208 are needed because all other string types have virtually
15209 identical handling apart from the tag. By having versions
15210 of the ASN1 functions that just operate on content octets
15211 IMPLICIT tagging can be handled properly. It also allows
15212 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15213 and ASN1_INTEGER are identical apart from the tag.
15217 * Change the handling of OID objects as follows:
15219 - New object identifiers are inserted in objects.txt, following
15220 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15221 - objects.pl is used to process obj_mac.num and create a new
15223 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15226 This is currently kind of a hack, and the perl code in objects.pl
15227 isn't very elegant, but it works as I intended. The simplest way
15228 to check that it worked correctly is to look in obj_dat.h and
15229 check the array nid_objs and make sure the objects haven't moved
15230 around (this is important!). Additions are OK, as well as
15231 consistent name changes.
15235 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15239 * Addition of the command line parameter '-rand file' to 'openssl req'.
15240 The given file adds to whatever has already been seeded into the
15241 random pool through the RANDFILE configuration file option or
15242 environment variable, or the default random state file.
15246 * mkstack.pl now sorts each macro group into lexical order.
15247 Previously the output order depended on the order the files
15248 appeared in the directory, resulting in needless rewriting
15253 * Patches to make OpenSSL compile under Win32 again. Mostly
15254 work arounds for the VC++ problem that it treats func() as
15255 func(void). Also stripped out the parts of mkdef.pl that
15256 added extra typesafe functions: these no longer exist.
15260 * Reorganisation of the stack code. The macros are now all
15261 collected in safestack.h . Each macro is defined in terms of
15262 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15263 DEBUG_SAFESTACK is now handled in terms of function casts,
15264 this has the advantage of retaining type safety without the
15265 use of additional functions. If DEBUG_SAFESTACK is not defined
15266 then the non typesafe macros are used instead. Also modified the
15267 mkstack.pl script to handle the new form. Needs testing to see
15268 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15269 the default if no major problems. Similar behaviour for ASN1_SET_OF
15270 and PKCS12_STACK_OF.
15274 * When some versions of IIS use the 'NET' form of private key the
15275 key derivation algorithm is different. Normally MD5(password) is
15276 used as a 128 bit RC4 key. In the modified case
15277 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15278 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15279 as the old Netscape_RSA functions except they have an additional
15280 'sgckey' parameter which uses the modified algorithm. Also added
15281 an -sgckey command line option to the rsa utility. Thanks to
15282 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15283 algorithm to openssl-dev.
15287 * The evp_local.h macros were using 'c.##kname' which resulted in
15288 invalid expansion on some systems (SCO 5.0.5 for example).
15289 Corrected to 'c.kname'.
15291 *Phillip Porch <root@theporch.com>*
15293 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15294 a STACK of email addresses from a certificate or request, these look
15295 in the subject name and the subject alternative name extensions and
15296 omit any duplicate addresses.
15300 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15301 This makes DSA verification about 2 % faster.
15305 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15306 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15307 plus overhead for 1024 bit moduli).
15308 This makes exponentiations about 0.5 % faster for 1024 bit
15309 exponents (as measured by "openssl speed rsa2048").
15313 * Rename memory handling macros to avoid conflicts with other
15315 Malloc => OPENSSL_malloc
15316 Malloc_locked => OPENSSL_malloc_locked
15317 Realloc => OPENSSL_realloc
15318 Free => OPENSSL_free
15322 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15323 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15327 * CygWin32 support.
15329 *John Jarvie <jjarvie@newsguy.com>*
15331 * The type-safe stack code has been rejigged. It is now only compiled
15332 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15333 by default all type-specific stack functions are "#define"d back to
15334 standard stack functions. This results in more streamlined output
15335 but retains the type-safety checking possibilities of the original
15340 * The STACK code has been cleaned up, and certain type declarations
15341 that didn't make a lot of sense have been brought in line. This has
15342 also involved a cleanup of sorts in safestack.h to more correctly
15343 map type-safe stack functions onto their plain stack counterparts.
15344 This work has also resulted in a variety of "const"ifications of
15345 lots of the code, especially `_cmp` operations which should normally
15346 be prototyped with "const" parameters anyway.
15350 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15351 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15352 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15353 where all of 'md' is used each time the PRNG is used, but 'state'
15354 is used only indexed by a cyclic counter. As entropy may not be
15355 well distributed from the beginning, 'md' is important as a
15356 chaining variable. However, the output function chains only half
15357 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15358 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15359 in all of 'state' being rewritten, with the new values depending
15360 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15364 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15365 the handshake is continued after ssl_verify_cert_chain();
15366 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15367 can lead to 'unexplainable' connection aborts later.
15369 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15371 * Major EVP API cipher revision.
15372 Add hooks for extra EVP features. This allows various cipher
15373 parameters to be set in the EVP interface. Support added for variable
15374 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15375 setting of RC2 and RC5 parameters.
15377 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15380 Remove lots of duplicated code from the EVP library. For example *every*
15381 cipher init() function handles the 'iv' in the same way according to the
15382 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15383 for CFB and OFB modes they zero ctx->num.
15385 New functionality allows removal of S/MIME code RC2 hack.
15387 Most of the routines have the same form and so can be declared in terms
15390 By shifting this to the top level EVP_CipherInit() it can be removed from
15391 all individual ciphers. If the cipher wants to handle IVs or keys
15392 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15395 Change lots of functions like EVP_EncryptUpdate() to now return a
15396 value: although software versions of the algorithms cannot fail
15397 any installed hardware versions can.
15401 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15402 this option is set, tolerate broken clients that send the negotiated
15403 protocol version number instead of the requested protocol version
15408 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15409 i.e. non-zero for export ciphersuites, zero otherwise.
15410 Previous versions had this flag inverted, inconsistent with
15411 rsa_tmp_cb (..._TMP_RSA_CB).
15413 *Bodo Moeller; problem reported by Amit Chopra*
15415 * Add missing DSA library text string. Work around for some IIS
15416 key files with invalid SEQUENCE encoding.
15420 * Add a document (doc/standards.txt) that list all kinds of standards
15421 and so on that are implemented in OpenSSL.
15425 * Enhance c_rehash script. Old version would mishandle certificates
15426 with the same subject name hash and wouldn't handle CRLs at all.
15427 Added -fingerprint option to crl utility, to support new c_rehash
15432 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15436 * Fix for SSL server purpose checking. Server checking was
15437 rejecting certificates which had extended key usage present
15438 but no ssl client purpose.
15440 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15442 * Make PKCS#12 code work with no password. The PKCS#12 spec
15443 is a little unclear about how a blank password is handled.
15444 Since the password in encoded as a BMPString with terminating
15445 double NULL a zero length password would end up as just the
15446 double NULL. However no password at all is different and is
15447 handled differently in the PKCS#12 key generation code. NS
15448 treats a blank password as zero length. MSIE treats it as no
15449 password on export: but it will try both on import. We now do
15450 the same: PKCS12_parse() tries zero length and no password if
15451 the password is set to "" or NULL (NULL is now a valid password:
15452 it wasn't before) as does the pkcs12 application.
15456 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15457 perror when PEM_read_bio_X509_REQ fails, the error message must
15458 be obtained from the error queue.
15462 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15463 it in ERR_remove_state if appropriate, and change ERR_get_state
15464 accordingly to avoid race conditions (this is necessary because
15465 thread_hash is no longer constant once set).
15469 * Bugfix for linux-elf makefile.one.
15473 * RSA_get_default_method() will now cause a default
15474 RSA_METHOD to be chosen if one doesn't exist already.
15475 Previously this was only set during a call to RSA_new()
15476 or RSA_new_method(NULL) meaning it was possible for
15477 RSA_get_default_method() to return NULL.
15481 * Added native name translation to the existing DSO code
15482 that will convert (if the flag to do so is set) filenames
15483 that are sufficiently small and have no path information
15484 into a canonical native form. Eg. "blah" converted to
15485 "libblah.so" or "blah.dll" etc.
15489 * New function ERR_error_string_n(e, buf, len) which is like
15490 ERR_error_string(e, buf), but writes at most 'len' bytes
15491 including the 0 terminator. For ERR_error_string_n, 'buf'
15494 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15496 * CONF library reworked to become more general. A new CONF
15497 configuration file reader "class" is implemented as well as a
15498 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15499 old `CONF_*` functions are still there, but are reimplemented to
15500 work in terms of the new functions. Also, a set of functions
15501 to handle the internal storage of the configuration data is
15502 provided to make it easier to write new configuration file
15503 reader "classes" (I can definitely see something reading a
15504 configuration file in XML format, for example), called `_CONF_*`,
15505 or "the configuration storage API"...
15507 The new configuration file reading functions are:
15509 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15510 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15512 NCONF_default, NCONF_WIN32
15514 NCONF_dump_fp, NCONF_dump_bio
15516 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15517 NCONF_new creates a new CONF object. This works in the same way
15518 as other interfaces in OpenSSL, like the BIO interface.
15519 `NCONF_dump_*` dump the internal storage of the configuration file,
15520 which is useful for debugging. All other functions take the same
15521 arguments as the old `CONF_*` functions with the exception of the
15522 first that must be a `CONF *` instead of a `LHASH *`.
15524 To make it easier to use the new classes with the old `CONF_*` functions,
15525 the function CONF_set_default_method is provided.
15529 * Add '-tls1' option to 'openssl ciphers', which was already
15530 mentioned in the documentation but had not been implemented.
15531 (This option is not yet really useful because even the additional
15532 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15536 * Initial DSO code added into libcrypto for letting OpenSSL (and
15537 OpenSSL-based applications) load shared libraries and bind to
15538 them in a portable way.
15540 *Geoff Thorpe, with contributions from Richard Levitte*
15542 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15544 * Make sure _lrotl and _lrotr are only used with MSVC.
15546 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15547 (the default implementation of RAND_status).
15549 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15550 to '-clrext' (= clear extensions), as intended and documented.
15551 *Bodo Moeller; inconsistency pointed out by Michael Attili
15552 <attili@amaxo.com>*
15554 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15555 was larger than the MD block size.
15557 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15559 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15560 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15561 using the passed key: if the passed key was a private key the result
15562 of X509_print(), for example, would be to print out all the private key
15567 * des_quad_cksum() byte order bug fix.
15568 *Ulf Möller, using the problem description in krb4-0.9.7, where
15569 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15571 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15574 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15576 * For easily testing in shell scripts whether some command
15577 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15578 returns with exit code 0 iff no command of the given name is available.
15579 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15580 the output goes to stdout and nothing is printed to stderr.
15581 Additional arguments are always ignored.
15583 Since for each cipher there is a command of the same name,
15584 the 'no-cipher' compilation switches can be tested this way.
15586 ('openssl no-XXX' is not able to detect pseudo-commands such
15587 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15591 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15595 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15596 is set; it will be thrown away anyway because each handshake creates
15598 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15599 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15600 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15601 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15605 * New s_client option -ign_eof: EOF at stdin is ignored, and
15606 'Q' and 'R' lose their special meanings (quit/renegotiate).
15607 This is part of what -quiet does; unlike -quiet, -ign_eof
15608 does not suppress any output.
15612 * Add compatibility options to the purpose and trust code. The
15613 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15614 accepts a certificate or CA, this was the previous behaviour,
15615 with all the associated security issues.
15617 X509_TRUST_COMPAT is the old trust behaviour: only and
15618 automatically trust self signed roots in certificate store. A
15619 new trust setting X509_TRUST_DEFAULT is used to specify that
15620 a purpose has no associated trust setting and it should instead
15621 use the value in the default purpose.
15625 * Fix the PKCS#8 DSA private key code so it decodes keys again
15626 and fix a memory leak.
15630 * In util/mkerr.pl (which implements 'make errors'), preserve
15631 reason strings from the previous version of the .c file, as
15632 the default to have only downcase letters (and digits) in
15633 automatically generated reasons codes is not always appropriate.
15637 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15638 using strerror. Previously, ERR_reason_error_string() returned
15639 library names as reason strings for SYSerr; but SYSerr is a special
15640 case where small numbers are errno values, not library numbers.
15644 * Add '-dsaparam' option to 'openssl dhparam' application. This
15645 converts DSA parameters into DH parameters. (When creating parameters,
15646 DSA_generate_parameters is used.)
15650 * Include 'length' (recommended exponent length) in C code generated
15651 by 'openssl dhparam -C'.
15655 * The second argument to set_label in perlasm was already being used
15656 so couldn't be used as a "file scope" flag. Moved to third argument
15661 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15662 instead of RAND_bytes for encryption IVs and salts.
15666 * Include RAND_status() into RAND_METHOD instead of implementing
15667 it only for md_rand.c Otherwise replacing the PRNG by calling
15668 RAND_set_rand_method would be impossible.
15672 * Don't let DSA_generate_key() enter an infinite loop if the random
15673 number generation fails.
15677 * New 'rand' application for creating pseudo-random output.
15681 * Added configuration support for Linux/IA64
15683 *Rolf Haberrecker <rolf@suse.de>*
15685 * Assembler module support for Mingw32.
15689 * Shared library support for HPUX (in shlib/).
15691 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15693 * Shared library support for Solaris gcc.
15695 *Lutz Behnke <behnke@trustcenter.de>*
15697 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15699 * PKCS7_encrypt() was adding text MIME headers twice because they
15700 were added manually and by SMIME_crlf_copy().
15704 * In bntest.c don't call BN_rand with zero bits argument.
15706 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15708 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15709 case was implemented. This caused BN_div_recp() to fail occasionally.
15713 * Add an optional second argument to the set_label() in the perl
15714 assembly language builder. If this argument exists and is set
15715 to 1 it signals that the assembler should use a symbol whose
15716 scope is the entire file, not just the current function. This
15717 is needed with MASM which uses the format label:: for this scope.
15719 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15721 * Change the ASN1 types so they are typedefs by default. Before
15722 almost all types were #define'd to ASN1_STRING which was causing
15723 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15728 * Change names of new functions to the new get1/get0 naming
15729 convention: After 'get1', the caller owns a reference count
15730 and has to call `..._free`; 'get0' returns a pointer to some
15731 data structure without incrementing reference counters.
15732 (Some of the existing 'get' functions increment a reference
15733 counter, some don't.)
15734 Similarly, 'set1' and 'add1' functions increase reference
15735 counters or duplicate objects.
15739 * Allow for the possibility of temp RSA key generation failure:
15740 the code used to assume it always worked and crashed on failure.
15744 * Fix potential buffer overrun problem in BIO_printf().
15745 *Ulf Möller, using public domain code by Patrick Powell; problem
15746 pointed out by David Sacerdote <das33@cornell.edu>*
15748 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15749 RAND_egd() and RAND_status(). In the command line application,
15750 the EGD socket can be specified like a seed file using RANDFILE
15755 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15756 Some CAs (e.g. Verisign) distribute certificates in this form.
15760 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15761 list to exclude them. This means that no special compilation option
15762 is needed to use anonymous DH: it just needs to be included in the
15767 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15768 EVP_MD_type. The old functionality is available in a new macro called
15769 EVP_MD_md(). Change code that uses it and update docs.
15773 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15774 where the `void *` argument is replaced by a function pointer argument.
15775 Previously `void *` was abused to point to functions, which works on
15776 many platforms, but is not correct. As these functions are usually
15777 called by macros defined in OpenSSL header files, most source code
15778 should work without changes.
15782 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15783 sections with information on -D... compiler switches used for
15784 compiling the library so that applications can see them. To enable
15785 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15786 must be defined. E.g.,
15787 #define OPENSSL_ALGORITHM_DEFINES
15788 #include <openssl/opensslconf.h>
15789 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15791 *Richard Levitte, Ulf and Bodo Möller*
15793 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15798 * Change the 'other' type in certificate aux info to a STACK_OF
15799 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15800 the required ASN1 format: arbitrary types determined by an OID.
15804 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15805 argument to 'req'. This is not because the function is newer or
15806 better than others it just uses the work 'NEW' in the certificate
15807 request header lines. Some software needs this.
15811 * Reorganise password command line arguments: now passwords can be
15812 obtained from various sources. Delete the PEM_cb function and make
15813 it the default behaviour: i.e. if the callback is NULL and the
15814 usrdata argument is not NULL interpret it as a null terminated pass
15815 phrase. If usrdata and the callback are NULL then the pass phrase
15816 is prompted for as usual.
15820 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15821 the support is automatically enabled. The resulting binaries will
15822 autodetect the card and use it if present.
15824 *Ben Laurie and Compaq Inc.*
15826 * Work around for Netscape hang bug. This sends certificate request
15827 and server done in one record. Since this is perfectly legal in the
15828 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15829 the bugs/SSLv3 entry for more info.
15833 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15837 * Add -rand argument to smime and pkcs12 applications and read/write
15842 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15846 * Add command line password options to the remaining applications.
15850 * Bug fix for BN_div_recp() for numerators with an even number of
15855 * More tests in bntest.c, and changed test_bn output.
15859 * ./config recognizes MacOS X now.
15863 * Bug fix for BN_div() when the first words of num and divisor are
15864 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15868 * Add support for various broken PKCS#8 formats, and command line
15869 options to produce them.
15873 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15874 get temporary BIGNUMs from a BN_CTX.
15878 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15883 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15884 include a #define from the old name to the new. The original intent
15885 was that statically linked binaries could for example just call
15886 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15887 link with digests. This never worked because SSLeay_add_all_digests()
15888 and SSLeay_add_all_ciphers() were in the same source file so calling
15889 one would link with the other. They are now in separate source files.
15893 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15897 * Use a less unusual form of the Miller-Rabin primality test (it used
15898 a binary algorithm for exponentiation integrated into the Miller-Rabin
15899 loop, our standard modexp algorithms are faster).
15903 * Support for the EBCDIC character set completed.
15905 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15907 * Source code cleanups: use const where appropriate, eliminate casts,
15908 use `void *` instead of `char *` in lhash.
15912 * Bugfix: ssl3_send_server_key_exchange was not restartable
15913 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15914 this the server could overwrite ephemeral keys that the client
15919 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15920 using 50 iterations of the Rabin-Miller test.
15922 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15923 iterations of the Rabin-Miller test as required by the appendix
15924 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15925 As BN_is_prime_fasttest includes trial division, DSA parameter
15926 generation becomes much faster.
15928 This implies a change for the callback functions in DSA_is_prime
15929 and DSA_generate_parameters: The callback function is called once
15930 for each positive witness in the Rabin-Miller test, not just
15931 occasionally in the inner loop; and the parameters to the
15932 callback function now provide an iteration count for the outer
15933 loop rather than for the current invocation of the inner loop.
15934 DSA_generate_parameters additionally can call the callback
15935 function with an 'iteration count' of -1, meaning that a
15936 candidate has passed the trial division test (when q is generated
15937 from an application-provided seed, trial division is skipped).
15941 * New function BN_is_prime_fasttest that optionally does trial
15942 division before starting the Rabin-Miller test and has
15943 an additional BN_CTX * argument (whereas BN_is_prime always
15944 has to allocate at least one BN_CTX).
15945 'callback(1, -1, cb_arg)' is called when a number has passed the
15946 trial division stage.
15950 * Fix for bug in CRL encoding. The validity dates weren't being handled
15955 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15959 * New function BN_pseudo_rand().
15963 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15964 bignum version of BN_from_montgomery() with the working code from
15965 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15970 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15971 made it impossible to use the same SSL_SESSION data structure in
15972 SSL2 clients in multiple threads.
15976 * The return value of RAND_load_file() no longer counts bytes obtained
15977 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15978 to seed the PRNG (previously an explicit byte count was required).
15980 *Ulf Möller, Bodo Möller*
15982 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15983 used `char *` instead of `void *` and had casts all over the place.
15987 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15991 * Retain source code compatibility for BN_prime_checks macro:
15992 BN_is_prime(..., BN_prime_checks, ...) now uses
15993 BN_prime_checks_for_size to determine the appropriate number of
15994 Rabin-Miller iterations.
15998 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15999 DH_CHECK_P_NOT_SAFE_PRIME.
16000 (Check if this is true? OpenPGP calls them "strong".)
16004 * Merge the functionality of "dh" and "gendh" programs into a new program
16005 "dhparam". The old programs are retained for now but will handle DH keys
16006 (instead of parameters) in future.
16010 * Make the ciphers, s_server and s_client programs check the return values
16011 when a new cipher list is set.
16015 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16016 ciphers. Before when the 56bit ciphers were enabled the sorting was
16019 The syntax for the cipher sorting has been extended to support sorting by
16020 cipher-strength (using the strength_bits hard coded in the tables).
16021 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16023 Fix a bug in the cipher-command parser: when supplying a cipher command
16024 string with an "undefined" symbol (neither command nor alphanumeric
16025 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16026 an error is flagged.
16028 Due to the strength-sorting extension, the code of the
16029 ssl_create_cipher_list() function was completely rearranged. I hope that
16030 the readability was also increased :-)
16032 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16034 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16035 for the first serial number and places 2 in the serial number file. This
16036 avoids problems when the root CA is created with serial number zero and
16037 the first user certificate has the same issuer name and serial number
16042 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16043 the new code. Add documentation for this stuff.
16047 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16048 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16049 structures and behave in an analogous way to the X509v3 functions:
16050 they shouldn't be called directly but wrapper functions should be used
16053 So we also now have some wrapper functions that call the X509at functions
16054 when passed certificate requests. (TO DO: similar things can be done with
16055 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16056 things. Some of these need some d2i or i2d and print functionality
16057 because they handle more complex structures.)
16061 * Add missing #ifndefs that caused missing symbols when building libssl
16062 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16063 NO_RSA in `ssl/s2*.c`.
16065 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16067 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16068 has a return value which indicates the quality of the random data
16069 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16070 error queue. New function RAND_pseudo_bytes() generates output that is
16071 guaranteed to be unique but not unpredictable. RAND_add is like
16072 RAND_seed, but takes an extra argument for an entropy estimate
16073 (RAND_seed always assumes full entropy).
16077 * Do more iterations of Rabin-Miller probable prime test (specifically,
16078 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16079 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16080 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16081 false-positive rate of at most 2^-80 for random input.
16085 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16089 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16090 in the 0.9.5 release), this returns the chain
16091 from an X509_CTX structure with a dup of the stack and all
16092 the X509 reference counts upped: so the stack will exist
16093 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16096 Also make SSL_SESSION_print() print out the verify return
16101 * Add manpage for the pkcs12 command. Also change the default
16102 behaviour so MAC iteration counts are used unless the new
16103 -nomaciter option is used. This improves file security and
16104 only older versions of MSIE (4.0 for example) need it.
16108 * Honor the no-xxx Configure options when creating .DEF files.
16112 * Add PKCS#10 attributes to field table: challengePassword,
16113 unstructuredName and unstructuredAddress. These are taken from
16114 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16115 international characters are used.
16117 More changes to X509_ATTRIBUTE code: allow the setting of types
16118 based on strings. Remove the 'loc' parameter when adding
16119 attributes because these will be a SET OF encoding which is sorted
16124 * Initial changes to the 'req' utility to allow request generation
16125 automation. This will allow an application to just generate a template
16126 file containing all the field values and have req construct the
16129 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16130 used all over the place including certificate requests and PKCS#7
16131 structures. They are currently handled manually where necessary with
16132 some primitive wrappers for PKCS#7. The new functions behave in a
16133 manner analogous to the X509 extension functions: they allow
16134 attributes to be looked up by NID and added.
16136 Later something similar to the X509V3 code would be desirable to
16137 automatically handle the encoding, decoding and printing of the
16138 more complex types. The string types like challengePassword can
16139 be handled by the string table functions.
16141 Also modified the multi byte string table handling. Now there is
16142 a 'global mask' which masks out certain types. The table itself
16143 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16144 is useful when for example there is only one permissible type
16145 (as in countryName) and using the mask might result in no valid
16150 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16151 SSL_get_peer_finished to allow applications to obtain the latest
16152 Finished messages sent to the peer or expected from the peer,
16153 respectively. (SSL_get_peer_finished is usually the Finished message
16154 actually received from the peer, otherwise the protocol will be aborted.)
16156 As the Finished message are message digests of the complete handshake
16157 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16158 be used for external authentication procedures when the authentication
16159 provided by SSL/TLS is not desired or is not enough.
16163 * Enhanced support for Alpha Linux is added. Now ./config checks if
16164 the host supports BWX extension and if Compaq C is present on the
16165 $PATH. Just exploiting of the BWX extension results in 20-30%
16166 performance kick for some algorithms, e.g. DES and RC4 to mention
16167 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16172 * Add support for MS "fast SGC". This is arguably a violation of the
16173 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16174 weak crypto and after checking the certificate is SGC a second one
16175 with strong crypto. MS SGC stops the first handshake after receiving
16176 the server certificate message and sends a second client hello. Since
16177 a server will typically do all the time consuming operations before
16178 expecting any further messages from the client (server key exchange
16179 is the most expensive) there is little difference between the two.
16181 To get OpenSSL to support MS SGC we have to permit a second client
16182 hello message after we have sent server done. In addition we have to
16183 reset the MAC if we do get this second client hello.
16187 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16188 if a DER encoded private key is RSA or DSA traditional format. Changed
16189 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16190 format DER encoded private key. Newer code should use PKCS#8 format which
16191 has the key type encoded in the ASN1 structure. Added DER private key
16192 support to pkcs8 application.
16196 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16197 ciphersuites has been selected (as required by the SSL 3/TLS 1
16198 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16199 is set, we interpret this as a request to violate the specification
16200 (the worst that can happen is a handshake failure, and 'correct'
16201 behaviour would result in a handshake failure anyway).
16205 * In SSL_CTX_add_session, take into account that there might be multiple
16206 SSL_SESSION structures with the same session ID (e.g. when two threads
16207 concurrently obtain them from an external cache).
16208 The internal cache can handle only one SSL_SESSION with a given ID,
16209 so if there's a conflict, we now throw out the old one to achieve
16214 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16215 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16216 some routines that use cipher OIDs: some ciphers do not have OIDs
16217 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16222 * Simplify the trust setting structure and code. Now we just have
16223 two sequences of OIDs for trusted and rejected settings. These will
16224 typically have values the same as the extended key usage extension
16225 and any application specific purposes.
16227 The trust checking code now has a default behaviour: it will just
16228 check for an object with the same NID as the passed id. Functions can
16229 be provided to override either the default behaviour or the behaviour
16230 for a given id. SSL client, server and email already have functions
16231 in place for compatibility: they check the NID and also return "trusted"
16232 if the certificate is self signed.
16236 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16237 traditional format into an EVP_PKEY structure.
16241 * Add a password callback function PEM_cb() which either prompts for
16242 a password if usr_data is NULL or otherwise assumes it is a null
16243 terminated password. Allow passwords to be passed on command line
16244 environment or config files in a few more utilities.
16248 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16249 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16250 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16251 Update documentation.
16255 * Support for ASN1 "NULL" type. This could be handled before by using
16256 ASN1_TYPE but there wasn't any function that would try to read a NULL
16257 and produce an error if it couldn't. For compatibility we also have
16258 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16259 don't allocate anything because they don't need to.
16263 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16266 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16268 * Rebuild of the memory allocation routines used by OpenSSL code and
16269 possibly others as well. The purpose is to make an interface that
16270 provide hooks so anyone can build a separate set of allocation and
16271 deallocation routines to be used by OpenSSL, for example memory
16272 pool implementations, or something else, which was previously hard
16273 since Malloc(), Realloc() and Free() were defined as macros having
16274 the values malloc, realloc and free, respectively (except for Win32
16275 compilations). The same is provided for memory debugging code.
16276 OpenSSL already comes with functionality to find memory leaks, but
16277 this gives people a chance to debug other memory problems.
16279 With these changes, a new set of functions and macros have appeared:
16281 CRYPTO_set_mem_debug_functions() [F]
16282 CRYPTO_get_mem_debug_functions() [F]
16283 CRYPTO_dbg_set_options() [F]
16284 CRYPTO_dbg_get_options() [F]
16285 CRYPTO_malloc_debug_init() [M]
16287 The memory debug functions are NULL by default, unless the library
16288 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16289 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16290 gives the standard debugging functions that come with OpenSSL) or
16291 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16292 provided by the library user) must be used. When the standard
16293 debugging functions are used, CRYPTO_dbg_set_options can be used to
16294 request additional information:
16295 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16296 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16298 Also, things like CRYPTO_set_mem_functions will always give the
16299 expected result (the new set of functions is used for allocation
16300 and deallocation) at all times, regardless of platform and compiler
16303 To finish it up, some functions that were never use in any other
16304 way than through macros have a new API and new semantic:
16306 CRYPTO_dbg_malloc()
16307 CRYPTO_dbg_realloc()
16310 All macros of value have retained their old syntax.
16312 *Richard Levitte and Bodo Moeller*
16314 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16315 ordering of SMIMECapabilities wasn't in "strength order" and there
16316 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16321 * Some ASN1 types with illegal zero length encoding (INTEGER,
16322 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16324 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16326 * Merge in my S/MIME library for OpenSSL. This provides a simple
16327 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16328 functionality to handle multipart/signed properly) and a utility
16329 called 'smime' to call all this stuff. This is based on code I
16330 originally wrote for Celo who have kindly allowed it to be
16331 included in OpenSSL.
16335 * Add variants des_set_key_checked and des_set_key_unchecked of
16336 des_set_key (aka des_key_sched). Global variable des_check_key
16337 decides which of these is called by des_set_key; this way
16338 des_check_key behaves as it always did, but applications and
16339 the library itself, which was buggy for des_check_key == 1,
16340 have a cleaner way to pick the version they need.
16344 * New function PKCS12_newpass() which changes the password of a
16349 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16350 dynamic mix. In both cases the ids can be used as an index into the
16351 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16352 functions so they accept a list of the field values and the
16353 application doesn't need to directly manipulate the X509_TRUST
16358 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16363 * Modify the way the V3 extension code looks up extensions. This now
16364 works in a similar way to the object code: we have some "standard"
16365 extensions in a static table which is searched with OBJ_bsearch()
16366 and the application can add dynamic ones if needed. The file
16367 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16368 updated whenever a new extension is added to the core code and kept
16369 in ext_nid order. There is a simple program 'tabtest.c' which checks
16370 this. New extensions are not added too often so this file can readily
16371 be maintained manually.
16373 There are two big advantages in doing things this way. The extensions
16374 can be looked up immediately and no longer need to be "added" using
16375 X509V3_add_standard_extensions(): this function now does nothing.
16376 Side note: I get *lots* of email saying the extension code doesn't
16377 work because people forget to call this function.
16378 Also no dynamic allocation is done unless new extensions are added:
16379 so if we don't add custom extensions there is no need to call
16380 X509V3_EXT_cleanup().
16384 * Modify enc utility's salting as follows: make salting the default. Add a
16385 magic header, so unsalted files fail gracefully instead of just decrypting
16386 to garbage. This is because not salting is a big security hole, so people
16387 should be discouraged from doing it.
16391 * Fixes and enhancements to the 'x509' utility. It allowed a message
16392 digest to be passed on the command line but it only used this
16393 parameter when signing a certificate. Modified so all relevant
16394 operations are affected by the digest parameter including the
16395 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16396 DSA key was used because it didn't fix the digest.
16400 * Initial certificate chain verify code. Currently tests the untrusted
16401 certificates for consistency with the verify purpose (which is set
16402 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16404 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16405 this is because it will reject chains with invalid extensions whereas
16406 every previous version of OpenSSL and SSLeay made no checks at all.
16408 Trust code: checks the root CA for the relevant trust settings. Trust
16409 settings have an initial value consistent with the verify purpose: e.g.
16410 if the verify purpose is for SSL client use it expects the CA to be
16411 trusted for SSL client use. However the default value can be changed to
16412 permit custom trust settings: one example of this would be to only trust
16413 certificates from a specific "secure" set of CAs.
16415 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16416 which should be used for version portability: especially since the
16417 verify structure is likely to change more often now.
16419 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16420 to set them. If not set then assume SSL clients will verify SSL servers
16423 Two new options to the verify program: -untrusted allows a set of
16424 untrusted certificates to be passed in and -purpose which sets the
16425 intended purpose of the certificate. If a purpose is set then the
16426 new chain verify code is used to check extension consistency.
16430 * Support for the authority information access extension.
16434 * Modify RSA and DSA PEM read routines to transparently handle
16435 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16436 public keys in a format compatible with certificate
16437 SubjectPublicKeyInfo structures. Unfortunately there were already
16438 functions called *_PublicKey_* which used various odd formats so
16439 these are retained for compatibility: however the DSA variants were
16440 never in a public release so they have been deleted. Changed dsa/rsa
16441 utilities to handle the new format: note no releases ever handled public
16442 keys so we should be OK.
16444 The primary motivation for this change is to avoid the same fiasco
16445 that dogs private keys: there are several incompatible private key
16446 formats some of which are standard and some OpenSSL specific and
16447 require various evil hacks to allow partial transparent handling and
16448 even then it doesn't work with DER formats. Given the option anything
16449 other than PKCS#8 should be dumped: but the other formats have to
16450 stay in the name of compatibility.
16452 With public keys and the benefit of hindsight one standard format
16453 is used which works with EVP_PKEY, RSA or DSA structures: though
16454 it clearly returns an error if you try to read the wrong kind of key.
16456 Added a -pubkey option to the 'x509' utility to output the public key.
16457 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16458 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16459 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16460 that do the same as the `EVP_PKEY_assign_*()` except they up the
16461 reference count of the added key (they don't "swallow" the
16466 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16467 CRLs would fail if the file contained no certificates or no CRLs:
16468 added a new function to read in both types and return the number
16469 read: this means that if none are read it will be an error. The
16470 DER versions of the certificate and CRL reader would always fail
16471 because it isn't possible to mix certificates and CRLs in DER format
16472 without choking one or the other routine. Changed this to just read
16473 a certificate: this is the best we can do. Also modified the code
16474 in `apps/verify.c` to take notice of return codes: it was previously
16475 attempting to read in certificates from NULL pointers and ignoring
16476 any errors: this is one reason why the cert and CRL reader seemed
16477 to work. It doesn't check return codes from the default certificate
16478 routines: these may well fail if the certificates aren't installed.
16482 * Code to support otherName option in GeneralName.
16486 * First update to verify code. Change the verify utility
16487 so it warns if it is passed a self signed certificate:
16488 for consistency with the normal behaviour. X509_verify
16489 has been modified to it will now verify a self signed
16490 certificate if *exactly* the same certificate appears
16491 in the store: it was previously impossible to trust a
16492 single self signed certificate. This means that:
16493 openssl verify ss.pem
16494 now gives a warning about a self signed certificate but
16495 openssl verify -CAfile ss.pem ss.pem
16500 * For servers, store verify_result in SSL_SESSION data structure
16501 (and add it to external session representation).
16502 This is needed when client certificate verifications fails,
16503 but an application-provided verification callback (set by
16504 SSL_CTX_set_cert_verify_callback) allows accepting the session
16505 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16506 but returns 1): When the session is reused, we have to set
16507 ssl->verify_result to the appropriate error code to avoid
16510 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16512 * Fix a bug in the new PKCS#7 code: it didn't consider the
16513 case in PKCS7_dataInit() where the signed PKCS7 structure
16514 didn't contain any existing data because it was being created.
16516 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16518 * Add a salt to the key derivation routines in enc.c. This
16519 forms the first 8 bytes of the encrypted file. Also add a
16520 -S option to allow a salt to be input on the command line.
16524 * New function X509_cmp(). Oddly enough there wasn't a function
16525 to compare two certificates. We do this by working out the SHA1
16526 hash and comparing that. X509_cmp() will be needed by the trust
16531 * SSL_get1_session() is like SSL_get_session(), but increments
16532 the reference count in the SSL_SESSION returned.
16534 *Geoff Thorpe <geoff@eu.c2.net>*
16536 * Fix for 'req': it was adding a null to request attributes.
16537 Also change the X509_LOOKUP and X509_INFO code to handle
16538 certificate auxiliary information.
16542 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16547 * Add the possibility to add extra information to the memory leak
16548 detecting output, to form tracebacks, showing from where each
16549 allocation was originated: CRYPTO_push_info("constant string") adds
16550 the string plus current file name and line number to a per-thread
16551 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16552 is like calling CYRPTO_pop_info() until the stack is empty.
16553 Also updated memory leak detection code to be multi-thread-safe.
16557 * Add options -text and -noout to pkcs7 utility and delete the
16558 encryption options which never did anything. Update docs.
16562 * Add options to some of the utilities to allow the pass phrase
16563 to be included on either the command line (not recommended on
16564 OSes like Unix) or read from the environment. Update the
16565 manpages and fix a few bugs.
16569 * Add a few manpages for some of the openssl commands.
16573 * Fix the -revoke option in ca. It was freeing up memory twice,
16574 leaking and not finding already revoked certificates.
16578 * Extensive changes to support certificate auxiliary information.
16579 This involves the use of X509_CERT_AUX structure and X509_AUX
16580 functions. An X509_AUX function such as PEM_read_X509_AUX()
16581 can still read in a certificate file in the usual way but it
16582 will also read in any additional "auxiliary information". By
16583 doing things this way a fair degree of compatibility can be
16584 retained: existing certificates can have this information added
16585 using the new 'x509' options.
16587 Current auxiliary information includes an "alias" and some trust
16588 settings. The trust settings will ultimately be used in enhanced
16589 certificate chain verification routines: currently a certificate
16590 can only be trusted if it is self signed and then it is trusted
16595 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16596 The problem was that one of the replacement routines had not been working
16597 since SSLeay releases. For now the offending routine has been replaced
16598 with non-optimised assembler. Even so, this now gives around 95%
16599 performance improvement for 1024 bit RSA signs.
16603 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16604 handling. Most clients have the effective key size in bits equal to
16605 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16606 A few however don't do this and instead use the size of the decrypted key
16607 to determine the RC2 key length and the AlgorithmIdentifier to determine
16608 the effective key length. In this case the effective key length can still
16609 be 40 bits but the key length can be 168 bits for example. This is fixed
16610 by manually forcing an RC2 key into the EVP_PKEY structure because the
16611 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16612 the key length and effective key length are equal.
16616 * Add a bunch of functions that should simplify the creation of
16617 X509_NAME structures. Now you should be able to do:
16618 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16619 and have it automatically work out the correct field type and fill in
16620 the structures. The more adventurous can try:
16621 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16622 and it will (hopefully) work out the correct multibyte encoding.
16626 * Change the 'req' utility to use the new field handling and multibyte
16627 copy routines. Before the DN field creation was handled in an ad hoc
16628 way in req, ca, and x509 which was rather broken and didn't support
16629 BMPStrings or UTF8Strings. Since some software doesn't implement
16630 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16631 using the dirstring_type option. See the new comment in the default
16632 openssl.cnf for more info.
16636 * Make crypto/rand/md_rand.c more robust:
16637 - Assure unique random numbers after fork().
16638 - Make sure that concurrent threads access the global counter and
16639 md serializably so that we never lose entropy in them
16640 or use exactly the same state in multiple threads.
16641 Access to the large state is not always serializable because
16642 the additional locking could be a performance killer, and
16643 md should be large enough anyway.
16647 * New file `apps/app_rand.c` with commonly needed functionality
16648 for handling the random seed file.
16650 Use the random seed file in some applications that previously did not:
16652 dsaparam -genkey (which also ignored its '-rand' option),
16655 x509 (when signing).
16656 Except on systems with /dev/urandom, it is crucial to have a random
16657 seed file at least for key creation, DSA signing, and for DH exchanges;
16658 for RSA signatures we could do without one.
16660 gendh and gendsa (unlike genrsa) used to read only the first byte
16661 of each file listed in the '-rand' option. The function as previously
16662 found in genrsa is now in app_rand.c and is used by all programs
16663 that support '-rand'.
16667 * In RAND_write_file, use mode 0600 for creating files;
16668 don't just chmod when it may be too late.
16672 * Report an error from X509_STORE_load_locations
16673 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16677 * New function ASN1_mbstring_copy() this copies a string in either
16678 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16679 into an ASN1_STRING type. A mask of permissible types is passed
16680 and it chooses the "minimal" type to use or an error if not type
16685 * Add function equivalents to the various macros in asn1.h. The old
16686 macros are retained with an `M_` prefix. Code inside the library can
16687 use the `M_` macros. External code (including the openssl utility)
16688 should *NOT* in order to be "shared library friendly".
16692 * Add various functions that can check a certificate's extensions
16693 to see if it usable for various purposes such as SSL client,
16694 server or S/MIME and CAs of these types. This is currently
16695 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16696 verification. Also added a -purpose flag to x509 utility to
16697 print out all the purposes.
16701 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16706 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16707 for, obtain and decode and extension and obtain its critical flag.
16708 This allows all the necessary extension code to be handled in a
16709 single function call.
16713 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16714 platforms. See crypto/rc4/rc4_enc.c for further details.
16718 * New -noout option to asn1parse. This causes no output to be produced
16719 its main use is when combined with -strparse and -out to extract data
16720 from a file (which may not be in ASN.1 format).
16724 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16725 when producing the local key id.
16727 *Richard Levitte <levitte@stacken.kth.se>*
16729 * New option -dhparam in s_server. This allows a DH parameter file to be
16730 stated explicitly. If it is not stated then it tries the first server
16731 certificate file. The previous behaviour hard coded the filename
16736 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16737 a public key to be input or output. For example:
16738 openssl rsa -in key.pem -pubout -out pubkey.pem
16739 Also added necessary DSA public key functions to handle this.
16743 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16744 in the message. This was handled by allowing
16745 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16747 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16749 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16750 to the end of the strings whereas this didn't. This would cause problems
16751 if strings read with d2i_ASN1_bytes() were later modified.
16753 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16755 * Fix for base64 decode bug. When a base64 bio reads only one line of
16756 data and it contains EOF it will end up returning an error. This is
16757 caused by input 46 bytes long. The cause is due to the way base64
16758 BIOs find the start of base64 encoded data. They do this by trying a
16759 trial decode on each line until they find one that works. When they
16760 do a flag is set and it starts again knowing it can pass all the
16761 data directly through the decoder. Unfortunately it doesn't reset
16762 the context it uses. This means that if EOF is reached an attempt
16763 is made to pass two EOFs through the context and this causes the
16764 resulting error. This can also cause other problems as well. As is
16765 usual with these problems it takes *ages* to find and the fix is
16766 trivial: move one line.
16768 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16770 * Ugly workaround to get s_client and s_server working under Windows. The
16771 old code wouldn't work because it needed to select() on sockets and the
16772 tty (for keypresses and to see if data could be written). Win32 only
16773 supports select() on sockets so we select() with a 1s timeout on the
16774 sockets and then see if any characters are waiting to be read, if none
16775 are present then we retry, we also assume we can always write data to
16776 the tty. This isn't nice because the code then blocks until we've
16777 received a complete line of data and it is effectively polling the
16778 keyboard at 1s intervals: however it's quite a bit better than not
16779 working at all :-) A dedicated Windows application might handle this
16780 with an event loop for example.
16784 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16785 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16786 will be called when RSA_sign() and RSA_verify() are used. This is useful
16787 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16788 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16789 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16790 This necessitated the support of an extra signature type NID_md5_sha1
16791 for SSL signatures and modifications to the SSL library to use it instead
16792 of calling RSA_public_decrypt() and RSA_private_encrypt().
16796 * Add new -verify -CAfile and -CApath options to the crl program, these
16797 will lookup a CRL issuers certificate and verify the signature in a
16798 similar way to the verify program. Tidy up the crl program so it
16799 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16800 less strict. It will now permit CRL extensions even if it is not
16801 a V2 CRL: this will allow it to tolerate some broken CRLs.
16805 * Initialize all non-automatic variables each time one of the openssl
16806 sub-programs is started (this is necessary as they may be started
16807 multiple times from the "OpenSSL>" prompt).
16809 *Lennart Bang, Bodo Moeller*
16811 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16812 removing all other RSA functionality (this is what NO_RSA does). This
16813 is so (for example) those in the US can disable those operations covered
16814 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16819 * Non-copying interface to BIO pairs.
16820 (still largely untested)
16824 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16825 ASCII string. This was handled independently in various places before.
16829 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16830 UTF8 strings a character at a time.
16834 * Use client_version from client hello to select the protocol
16835 (s23_srvr.c) and for RSA client key exchange verification
16836 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16840 * Add various utility functions to handle SPKACs, these were previously
16841 handled by poking round in the structure internals. Added new function
16842 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16843 print, verify and generate SPKACs. Based on an original idea from
16844 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16848 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16852 * Allow the config file extension section to be overwritten on the
16853 command line. Based on an original idea from Massimiliano Pala
16854 <madwolf@comune.modena.it>. The new option is called -extensions
16855 and can be applied to ca, req and x509. Also -reqexts to override
16856 the request extensions in req and -crlexts to override the crl extensions
16861 * Add new feature to the SPKAC handling in ca. Now you can include
16862 the same field multiple times by preceding it by "XXXX." for example:
16865 this is the same syntax as used in the req config file.
16869 * Allow certificate extensions to be added to certificate requests. These
16870 are specified in a 'req_extensions' option of the req section of the
16871 config file. They can be printed out with the -text option to req but
16872 are otherwise ignored at present.
16876 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16877 data read consists of only the final block it would not decrypted because
16878 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16879 A misplaced 'break' also meant the decrypted final block might not be
16880 copied until the next read.
16884 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16885 a few extra parameters to the DH structure: these will be useful if
16886 for example we want the value of 'q' or implement X9.42 DH.
16890 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16891 provides hooks that allow the default DSA functions or functions on a
16892 "per key" basis to be replaced. This allows hardware acceleration and
16893 hardware key storage to be handled without major modification to the
16894 library. Also added low level modexp hooks and CRYPTO_EX structure and
16895 associated functions.
16899 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16900 as "read only": it can't be written to and the buffer it points to will
16901 not be freed. Reading from a read only BIO is much more efficient than
16902 a normal memory BIO. This was added because there are several times when
16903 an area of memory needs to be read from a BIO. The previous method was
16904 to create a memory BIO and write the data to it, this results in two
16905 copies of the data and an O(n^2) reading algorithm. There is a new
16906 function BIO_new_mem_buf() which creates a read only memory BIO from
16907 an area of memory. Also modified the PKCS#7 routines to use read only
16912 * Bugfix: ssl23_get_client_hello did not work properly when called in
16913 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16914 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16915 but a retry condition occurred while trying to read the rest.
16919 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16920 NID_pkcs7_encrypted by default: this was wrong since this should almost
16921 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16922 the encrypted data type: this is a more sensible place to put it and it
16923 allows the PKCS#12 code to be tidied up that duplicated this
16928 * Changed obj_dat.pl script so it takes its input and output files on
16929 the command line. This should avoid shell escape redirection problems
16934 * Initial support for certificate extension requests, these are included
16935 in things like Xenroll certificate requests. Included functions to allow
16936 extensions to be obtained and added.
16940 * -crlf option to s_client and s_server for sending newlines as
16941 CRLF (as required by many protocols).
16945 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16947 * Install libRSAglue.a when OpenSSL is built with RSAref.
16949 *Ralf S. Engelschall*
16951 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16953 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16955 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16960 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16961 DH parameters/keys (q is lost during that conversion, but the resulting
16962 DH parameters contain its length).
16964 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16965 much faster than DH_generate_parameters (which creates parameters
16966 where `p = 2*q + 1`), and also the smaller q makes DH computations
16967 much more efficient (160-bit exponentiation instead of 1024-bit
16968 exponentiation); so this provides a convenient way to support DHE
16969 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16970 utter importance to use
16971 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16973 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16974 when such DH parameters are used, because otherwise small subgroup
16975 attacks may become possible!
16979 * Avoid memory leak in i2d_DHparams.
16983 * Allow the -k option to be used more than once in the enc program:
16984 this allows the same encrypted message to be read by multiple recipients.
16988 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16989 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16990 it will always use the numerical form of the OID, even if it has a short
16995 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16996 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16997 otherwise bn_mod_exp was called. In the case of hardware keys for example
16998 no private key components need be present and it might store extra data
16999 in the RSA structure, which cannot be accessed from bn_mod_exp.
17000 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17001 private key operations.
17005 * Added support for SPARC Linux.
17009 * pem_password_cb function type incompatibly changed from
17010 typedef int pem_password_cb(char *buf, int size, int rwflag);
17012 ....(char *buf, int size, int rwflag, void *userdata);
17013 so that applications can pass data to their callbacks:
17014 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17015 additional void * argument, which is just handed through whenever
17016 the password callback is called.
17018 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17020 New function SSL_CTX_set_default_passwd_cb_userdata.
17022 Compatibility note: As many C implementations push function arguments
17023 onto the stack in reverse order, the new library version is likely to
17024 interoperate with programs that have been compiled with the old
17025 pem_password_cb definition (PEM_whatever takes some data that
17026 happens to be on the stack as its last argument, and the callback
17027 just ignores this garbage); but there is no guarantee whatsoever that
17030 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17031 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17032 problems not only on Windows, but also on some Unix platforms.
17033 To avoid problematic command lines, these definitions are now in an
17034 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17035 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17039 * MIPS III/IV assembler module is reimplemented.
17043 * More DES library cleanups: remove references to srand/rand and
17044 delete an unused file.
17048 * Add support for the free Netwide assembler (NASM) under Win32,
17049 since not many people have MASM (ml) and it can be hard to obtain.
17050 This is currently experimental but it seems to work OK and pass all
17051 the tests. Check out INSTALL.W32 for info.
17055 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17056 without temporary keys kept an extra copy of the server key,
17057 and connections with temporary keys did not free everything in case
17062 * New function RSA_check_key and new openssl rsa option -check
17063 for verifying the consistency of RSA keys.
17065 *Ulf Moeller, Bodo Moeller*
17067 * Various changes to make Win32 compile work:
17068 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17069 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17070 comparison" warnings.
17071 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17075 * Add a debugging option to PKCS#5 v2 key generation function: when
17076 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17077 derived keys are printed to stderr.
17081 * Copy the flags in ASN1_STRING_dup().
17083 *Roman E. Pavlov <pre@mo.msk.ru>*
17085 * The x509 application mishandled signing requests containing DSA
17086 keys when the signing key was also DSA and the parameters didn't match.
17088 It was supposed to omit the parameters when they matched the signing key:
17089 the verifying software was then supposed to automatically use the CA's
17090 parameters if they were absent from the end user certificate.
17092 Omitting parameters is no longer recommended. The test was also
17093 the wrong way round! This was probably due to unusual behaviour in
17094 EVP_cmp_parameters() which returns 1 if the parameters match.
17095 This meant that parameters were omitted when they *didn't* match and
17096 the certificate was useless. Certificates signed with 'ca' didn't have
17099 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17101 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17102 The interface is as follows:
17103 Applications can use
17104 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17105 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17106 "off" is now the default.
17107 The library internally uses
17108 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17109 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17110 to disable memory-checking temporarily.
17112 Some inconsistent states that previously were possible (and were
17113 even the default) are now avoided.
17115 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17116 with each memory chunk allocated; this is occasionally more helpful
17117 than just having a counter.
17119 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17121 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17126 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17127 which largely parallels "options", but is for changing API behaviour,
17128 whereas "options" are about protocol behaviour.
17129 Initial "mode" flags are:
17131 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17132 a single record has been written.
17133 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17134 retries use the same buffer location.
17135 (But all of the contents must be
17140 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17143 * Fix problems with no-hmac etc.
17145 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17147 * New functions RSA_get_default_method(), RSA_set_method() and
17148 RSA_get_method(). These allows replacement of RSA_METHODs without having
17149 to mess around with the internals of an RSA structure.
17153 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17154 Also really enable memory leak checks in openssl.c and in some
17157 *Chad C. Mulligan, Bodo Moeller*
17159 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17160 up the length of negative integers. This has now been simplified to just
17161 store the length when it is first determined and use it later, rather
17162 than trying to keep track of where data is copied and updating it to
17164 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17166 * Add a new function PKCS7_signatureVerify. This allows the verification
17167 of a PKCS#7 signature but with the signing certificate passed to the
17168 function itself. This contrasts with PKCS7_dataVerify which assumes the
17169 certificate is present in the PKCS#7 structure. This isn't always the
17170 case: certificates can be omitted from a PKCS#7 structure and be
17171 distributed by "out of band" means (such as a certificate database).
17175 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17176 function prototypes in pem.h, also change util/mkdef.pl to add the
17177 necessary function names.
17181 * mk1mf.pl (used by Windows builds) did not properly read the
17182 options set by Configure in the top level Makefile, and Configure
17183 was not even able to write more than one option correctly.
17184 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17188 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17189 file to be loaded from a BIO or FILE pointer. The BIO version will
17190 for example allow memory BIOs to contain config info.
17194 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17195 Whoever hopes to achieve shared-library compatibility across versions
17196 must use this, not the compile-time macro.
17197 (Exercise 0.9.4: Which is the minimum library version required by
17199 Note: All this applies only to multi-threaded programs, others don't
17204 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17205 through a BIO pair triggered the default case, i.e.
17206 SSLerr(...,SSL_R_UNKNOWN_STATE).
17210 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17211 can use the SSL library even if none of the specific BIOs is
17216 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17217 for the encoded length.
17219 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17221 * Add initial documentation of the X509V3 functions.
17225 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17226 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17227 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17228 secure PKCS#8 private key format with a high iteration count.
17232 * Fix determination of Perl interpreter: A perl or perl5
17233 *directory* in $PATH was also accepted as the interpreter.
17235 *Ralf S. Engelschall*
17237 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17238 wrong with it but it was very old and did things like calling
17239 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17240 unusual formatting.
17244 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17245 to use the new extension code.
17249 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17250 with macros. This should make it easier to change their form, add extra
17251 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17256 * Add to configuration table a new entry that can specify an alternative
17257 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17258 according to Mark Crispin <MRC@Panda.COM>.
17262 * DES CBC did not update the IV. Weird.
17266 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17267 Changing the behaviour of the former might break existing programs --
17268 where IV updating is needed, des_ncbc_encrypt can be used.
17271 * When bntest is run from "make test" it drives bc to check its
17272 calculations, as well as internally checking them. If an internal check
17273 fails, it needs to cause bc to give a non-zero result or make test carries
17274 on without noticing the failure. Fixed.
17278 * DES library cleanups.
17282 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17283 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17284 ciphers. NOTE: although the key derivation function has been verified
17285 against some published test vectors it has not been extensively tested
17286 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17291 * Instead of "mkdir -p", which is not fully portable, use new
17292 Perl script "util/mkdir-p.pl".
17296 * Rewrite the way password based encryption (PBE) is handled. It used to
17297 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17298 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17299 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17300 the 'parameter' field of the AlgorithmIdentifier is passed to the
17301 underlying key generation function so it must do its own ASN1 parsing.
17302 This has also changed the EVP_PBE_CipherInit() function which now has a
17303 'parameter' argument instead of literal salt and iteration count values
17304 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17308 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17309 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17310 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17311 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17312 value was just used as a "magic string" and not used directly its
17313 value doesn't matter.
17317 * Introduce some semblance of const correctness to BN. Shame C doesn't
17322 * "linux-sparc64" configuration (ultrapenguin).
17324 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17325 "linux-sparc" configuration.
17327 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17329 * config now generates no-xxx options for missing ciphers.
17333 * Support the EBCDIC character set (work in progress).
17334 File ebcdic.c not yet included because it has a different license.
17336 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17338 * Support BS2000/OSD-POSIX.
17340 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17342 * Make callbacks for key generation use `void *` instead of `char *`.
17346 * Make S/MIME samples compile (not yet tested).
17350 * Additional typesafe stacks.
17354 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17358 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17360 * New configuration variant "sco5-gcc".
17362 * Updated some demos.
17364 *Sean O Riordain, Wade Scholine*
17366 * Add missing BIO_free at exit of pkcs12 application.
17370 * Fix memory leak in conf.c.
17374 * Updates for Win32 to assembler version of MD5.
17378 * Set #! path to perl in `apps/der_chop` to where we found it
17379 instead of using a fixed path.
17383 * SHA library changes for irix64-mips4-cc.
17387 * Improvements for VMS support.
17391 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17393 * Bignum library bug fix. IRIX 6 passes "make test" now!
17394 This also avoids the problems with SC4.2 and unpatched SC5.
17396 *Andy Polyakov <appro@fy.chalmers.se>*
17398 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17399 These are required because of the typesafe stack would otherwise break
17400 existing code. If old code used a structure member which used to be STACK
17401 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17402 sk_num or sk_value it would produce an error because the num, data members
17403 are not present in STACK_OF. Now it just produces a warning. sk_set
17404 replaces the old method of assigning a value to sk_value
17405 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17406 that does this will no longer work (and should use sk_set instead) but
17407 this could be regarded as a "questionable" behaviour anyway.
17411 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17412 correctly handle encrypted S/MIME data.
17416 * Change type of various DES function arguments from des_cblock
17417 (which means, in function argument declarations, pointer to char)
17418 to des_cblock * (meaning pointer to array with 8 char elements),
17419 which allows the compiler to do more typechecking; it was like
17420 that back in SSLeay, but with lots of ugly casts.
17422 Introduce new type const_des_cblock.
17426 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17427 problems: find RecipientInfo structure that matches recipient certificate
17428 and initialise the ASN1 structures properly based on passed cipher.
17432 * Belatedly make the BN tests actually check the results.
17436 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17437 to and from BNs: it was completely broken. New compilation option
17438 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17439 key elements as negative integers.
17443 * Reorganize and speed up MD5.
17445 *Andy Polyakov <appro@fy.chalmers.se>*
17449 *Richard Levitte <richard@levitte.org>*
17451 * New option -out to asn1parse to allow the parsed structure to be
17452 output to a file. This is most useful when combined with the -strparse
17453 option to examine the output of things like OCTET STRINGS.
17457 * Make SSL library a little more fool-proof by not requiring any longer
17458 that `SSL_set_{accept,connect}_state` be called before
17459 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17460 in many applications because usually everything *appeared* to work as
17461 intended anyway -- now it really works as intended).
17465 * Move openssl.cnf out of lib/.
17469 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17470 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17471 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17473 *Ralf S. Engelschall*
17475 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17476 handle PKCS#7 enveloped data properly.
17478 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17480 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17481 copying pointers. The cert_st handling is changed by this in
17482 various ways (and thus what used to be known as ctx->default_cert
17483 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17484 any longer when s->cert does not give us what we need).
17485 ssl_cert_instantiate becomes obsolete by this change.
17486 As soon as we've got the new code right (possibly it already is?),
17487 we have solved a couple of bugs of the earlier code where s->cert
17488 was used as if it could not have been shared with other SSL structures.
17490 Note that using the SSL API in certain dirty ways now will result
17491 in different behaviour than observed with earlier library versions:
17492 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17493 does not influence s as it used to.
17495 In order to clean up things more thoroughly, inside SSL_SESSION
17496 we don't use CERT any longer, but a new structure SESS_CERT
17497 that holds per-session data (if available); currently, this is
17498 the peer's certificate chain and, for clients, the server's certificate
17499 and temporary key. CERT holds only those values that can have
17500 meaningful defaults in an SSL_CTX.
17504 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17505 from the internal representation. Various PKCS#7 fixes: remove some
17506 evil casts and set the enc_dig_alg field properly based on the signing
17511 * Allow PKCS#12 password to be set from the command line or the
17512 environment. Let 'ca' get its config file name from the environment
17513 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17518 * Allow certificate policies extension to use an IA5STRING for the
17519 organization field. This is contrary to the PKIX definition but
17520 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17525 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17526 without disallowing inline assembler and the like for non-pedantic builds.
17530 * Support Borland C++ builder.
17532 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17538 * SHA-1 cleanups and performance enhancements.
17540 *Andy Polyakov <appro@fy.chalmers.se>*
17542 * Sparc v8plus assembler for the bignum library.
17544 *Andy Polyakov <appro@fy.chalmers.se>*
17546 * Accept any -xxx and +xxx compiler options in Configure.
17550 * Update HPUX configuration.
17554 * Add missing `sk_<type>_unshift()` function to safestack.h
17556 *Ralf S. Engelschall*
17558 * New function SSL_CTX_use_certificate_chain_file that sets the
17559 "extra_cert"s in addition to the certificate. (This makes sense
17560 only for "PEM" format files, as chains as a whole are not
17565 * Support verify_depth from the SSL API.
17566 x509_vfy.c had what can be considered an off-by-one-error:
17567 Its depth (which was not part of the external interface)
17568 was actually counting the number of certificates in a chain;
17569 now it really counts the depth.
17573 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17574 instead of X509err, which often resulted in confusing error
17575 messages since the error codes are not globally unique
17576 (e.g. an alleged error in ssl3_accept when a certificate
17577 didn't match the private key).
17579 * New function SSL_CTX_set_session_id_context that allows to set a default
17580 value (so that you don't need SSL_set_session_id_context for each
17581 connection using the SSL_CTX).
17585 * OAEP decoding bug fix.
17589 * Support INSTALL_PREFIX for package builders, as proposed by
17594 * New Configure options "threads" and "no-threads". For systems
17595 where the proper compiler options are known (currently Solaris
17596 and Linux), "threads" is the default.
17600 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17604 * Install various scripts to $(OPENSSLDIR)/misc, not to
17605 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17606 such as /usr/local/bin.
17610 * "make linux-shared" to build shared libraries.
17612 *Niels Poppe <niels@netbox.org>*
17614 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17618 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17619 extension adding in x509 utility.
17623 * Remove NOPROTO sections and error code comments.
17627 * Partial rewrite of the DEF file generator to now parse the ANSI
17632 * New Configure options --prefix=DIR and --openssldir=DIR.
17636 * Complete rewrite of the error code script(s). It is all now handled
17637 by one script at the top level which handles error code gathering,
17638 header rewriting and C source file generation. It should be much better
17639 than the old method: it now uses a modified version of Ulf's parser to
17640 read the ANSI prototypes in all header files (thus the old K&R definitions
17641 aren't needed for error creation any more) and do a better job of
17642 translating function codes into names. The old 'ASN1 error code embedded
17643 in a comment' is no longer necessary and it doesn't use .err files which
17644 have now been deleted. Also the error code call doesn't have to appear all
17645 on one line (which resulted in some large lines...).
17649 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17653 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17654 0 (which usually indicates a closed connection), but continue reading.
17658 * Fix some race conditions.
17662 * Add support for CRL distribution points extension. Add Certificate
17663 Policies and CRL distribution points documentation.
17667 * Move the autogenerated header file parts to crypto/opensslconf.h.
17671 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17672 8 of keying material. Merlin has also confirmed interop with this fix
17673 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17675 *Merlin Hughes <merlin@baltimore.ie>*
17677 * Fix lots of warnings.
17679 *Richard Levitte <levitte@stacken.kth.se>*
17681 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17682 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17684 *Richard Levitte <levitte@stacken.kth.se>*
17686 * Fix problems with sizeof(long) == 8.
17688 *Andy Polyakov <appro@fy.chalmers.se>*
17690 * Change functions to ANSI C.
17694 * Fix typos in error codes.
17696 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17698 * Remove defunct assembler files from Configure.
17702 * SPARC v8 assembler BIGNUM implementation.
17704 *Andy Polyakov <appro@fy.chalmers.se>*
17706 * Support for Certificate Policies extension: both print and set.
17707 Various additions to support the r2i method this uses.
17711 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17712 return a const string when you are expecting an allocated buffer.
17716 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17717 types DirectoryString and DisplayText.
17721 * Add code to allow r2i extensions to access the configuration database,
17722 add an LHASH database driver and add several ctx helper functions.
17726 * Fix an evil bug in bn_expand2() which caused various BN functions to
17727 fail when they extended the size of a BIGNUM.
17731 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17732 support typesafe stack.
17736 * Fix typo in SSL_[gs]et_options().
17738 *Nils Frostberg <nils@medcom.se>*
17740 * Delete various functions and files that belonged to the (now obsolete)
17741 old X509V3 handling code.
17745 * New Configure option "rsaref".
17749 * Don't auto-generate pem.h.
17753 * Introduce type-safe ASN.1 SETs.
17757 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17759 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17761 * Introduce type-safe STACKs. This will almost certainly break lots of code
17762 that links with OpenSSL (well at least cause lots of warnings), but fear
17763 not: the conversion is trivial, and it eliminates loads of evil casts. A
17764 few STACKed things have been converted already. Feel free to convert more.
17765 In the fullness of time, I'll do away with the STACK type altogether.
17769 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17770 specified in `<certfile>` by updating the entry in the index.txt file.
17771 This way one no longer has to edit the index.txt file manually for
17772 revoking a certificate. The -revoke option does the gory details now.
17774 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17776 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17777 `-text` option at all and this way the `-noout -text` combination was
17778 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17780 *Ralf S. Engelschall*
17782 * Make sure a corresponding plain text error message exists for the
17783 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17784 verify callback function determined that a certificate was revoked.
17786 *Ralf S. Engelschall*
17788 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17789 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17790 all available ciphers including rc5, which was forgotten until now.
17791 In order to let the testing shell script know which algorithms
17792 are available, a new (up to now undocumented) command
17793 `openssl list-cipher-commands` is used.
17797 * Bugfix: s_client occasionally would sleep in select() when
17798 it should have checked SSL_pending() first.
17802 * New functions DSA_do_sign and DSA_do_verify to provide access to
17803 the raw DSA values prior to ASN.1 encoding.
17807 * Tweaks to Configure
17809 *Niels Poppe <niels@netbox.org>*
17811 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17816 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17820 * New config option to avoid instructions that are illegal on the 80386.
17821 The default code is faster, but requires at least a 486.
17825 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17826 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17827 same as SSL2_VERSION anyway.
17831 * New "-showcerts" option for s_client.
17835 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17836 application. Various cleanups and fixes.
17840 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17841 modify error routines to work internally. Add error codes and PBE init
17842 to library startup routines.
17846 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17847 packing functions to asn1 and evp. Changed function names and error
17848 codes along the way.
17852 * PKCS12 integration: and so it begins... First of several patches to
17853 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17854 objects to objects.h
17858 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17859 and display support for Thawte strong extranet extension.
17863 * Add LinuxPPC support.
17865 *Jeff Dubrule <igor@pobox.org>*
17867 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17868 bn_div_words in alpha.s.
17870 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17872 * Make sure the RSA OAEP test is skipped under -DRSAref because
17873 OAEP isn't supported when OpenSSL is built with RSAref.
17875 *Ulf Moeller <ulf@fitug.de>*
17877 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17878 so they no longer are missing under -DNOPROTO.
17880 *Soren S. Jorvang <soren@t.dk>*
17882 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17884 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17885 doesn't work when the session is reused. Coming soon!
17889 * Fix a security hole, that allows sessions to be reused in the wrong
17890 context thus bypassing client cert protection! All software that uses
17891 client certs and session caches in multiple contexts NEEDS PATCHING to
17892 allow session reuse! A fuller solution is in the works.
17894 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17896 * Some more source tree cleanups (removed obsolete files
17897 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17898 permission on "config" script to be executable) and a fix for the INSTALL
17901 *Ulf Moeller <ulf@fitug.de>*
17903 * Remove some legacy and erroneous uses of malloc, free instead of
17906 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17908 * Make rsa_oaep_test return non-zero on error.
17910 *Ulf Moeller <ulf@fitug.de>*
17912 * Add support for native Solaris shared libraries. Configure
17913 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17914 if someone would make that last step automatic.
17916 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17918 * ctx_size was not built with the right compiler during "make links". Fixed.
17922 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17923 except NULL ciphers". This means the default cipher list will no longer
17924 enable NULL ciphers. They need to be specifically enabled e.g. with
17925 the string "DEFAULT:eNULL".
17929 * Fix to RSA private encryption routines: if p < q then it would
17930 occasionally produce an invalid result. This will only happen with
17931 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17935 * Be less restrictive and allow also `perl util/perlpath.pl
17936 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17937 because this way one can also use an interpreter named `perl5` (which is
17938 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17939 installed as `perl`).
17941 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17943 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17945 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17947 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17948 advapi32.lib to Win32 build and change the pem test comparison
17949 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17950 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17951 and crypto/des/ede_cbcm_enc.c.
17955 * DES quad checksum was broken on big-endian architectures. Fixed.
17959 * Comment out two functions in bio.h that aren't implemented. Fix up the
17960 Win32 test batch file so it (might) work again. The Win32 test batch file
17961 is horrible: I feel ill....
17965 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17966 in e_os.h. Audit of header files to check ANSI and non ANSI
17967 sections: 10 functions were absent from non ANSI section and not exported
17968 from Windows DLLs. Fixed up libeay.num for new functions.
17972 * Make `openssl version` output lines consistent.
17974 *Ralf S. Engelschall*
17976 * Fix Win32 symbol export lists for BIO functions: Added
17977 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17978 to ms/libeay{16,32}.def.
17980 *Ralf S. Engelschall*
17982 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17983 fine under Unix and passes some trivial tests I've now added. But the
17984 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17985 added to make sure no one expects that this stuff really works in the
17986 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17987 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17990 *Ralf S. Engelschall*
17992 * Fix the generation of two part addresses in perl.
17994 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17996 * Add config entry for Linux on MIPS.
17998 *John Tobey <jtobey@channel1.com>*
18000 * Make links whenever Configure is run, unless we are on Windoze.
18004 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18005 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18010 * Add a useful kludge to allow package maintainers to specify compiler and
18011 other platforms details on the command line without having to patch the
18012 Configure script every time: One now can use
18013 `perl Configure <id>:<details>`,
18014 i.e. platform ids are allowed to have details appended
18015 to them (separated by colons). This is treated as there would be a static
18016 pre-configured entry in Configure's %table under key `<id>` with value
18017 `<details>` and `perl Configure <id>` is called. So, when you want to
18018 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18019 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18020 now, which overrides the FreeBSD-elf entry on-the-fly.
18022 *Ralf S. Engelschall*
18024 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18028 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18029 on the `perl Configure ...` command line. This way one can compile
18030 OpenSSL libraries with Position Independent Code (PIC) which is needed
18031 for linking it into DSOs.
18033 *Ralf S. Engelschall*
18035 * Remarkably, export ciphers were totally broken and no-one had noticed!
18040 * Cleaned up the LICENSE document: The official contact for any license
18041 questions now is the OpenSSL core team under openssl-core@openssl.org.
18042 And add a paragraph about the dual-license situation to make sure people
18043 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18044 to the OpenSSL toolkit.
18046 *Ralf S. Engelschall*
18048 * General source tree makefile cleanups: Made `making xxx in yyy...`
18049 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18050 Additionally cleaned up the `make links` target: Remove unnecessary
18051 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18052 to speed processing and no longer clutter the display with confusing
18053 stuff. Instead only the actually done links are displayed.
18055 *Ralf S. Engelschall*
18057 * Permit null encryption ciphersuites, used for authentication only. It used
18058 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18059 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18064 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18065 signed attributes when verifying signatures (this would break them),
18066 the detached data encoding was wrong and public keys obtained using
18067 X509_get_pubkey() weren't freed.
18071 * Add text documentation for the BUFFER functions. Also added a work around
18072 to a Win95 console bug. This was triggered by the password read stuff: the
18073 last character typed gets carried over to the next fread(). If you were
18074 generating a new cert request using 'req' for example then the last
18075 character of the passphrase would be CR which would then enter the first
18080 * Added the new 'Includes OpenSSL Cryptography Software' button as
18081 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18082 button and can be used by applications based on OpenSSL to show the
18083 relationship to the OpenSSL project.
18085 *Ralf S. Engelschall*
18087 * Remove confusing variables in function signatures in files
18088 ssl/ssl_lib.c and ssl/ssl.h.
18090 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18092 * Don't install bss_file.c under PREFIX/include/
18094 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18096 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18097 functions that return function pointers and has support for NT specific
18098 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18099 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18100 unsigned to signed types: this was killing the Win32 compile.
18104 * Add new certificate file to stack functions,
18105 SSL_add_dir_cert_subjects_to_stack() and
18106 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18107 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18108 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18109 This means that Apache-SSL and similar packages don't have to mess around
18110 to add as many CAs as they want to the preferred list.
18114 * Experiment with doxygen documentation. Currently only partially applied to
18116 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18117 openssl.doxy as the configuration file.
18121 * Get rid of remaining C++-style comments which strict C compilers hate.
18123 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18125 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18126 compiled in by default: it has problems with large keys.
18130 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18131 DH private keys and/or callback functions which directly correspond to
18132 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18133 is needed for applications which have to configure certificates on a
18134 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18136 For the RSA certificate situation is makes no difference, but
18137 for the DSA certificate situation this fixes the "no shared cipher"
18138 problem where the OpenSSL cipher selection procedure failed because the
18139 temporary keys were not overtaken from the context and the API provided
18140 no way to reconfigure them.
18141 The new functions now let applications reconfigure the stuff and they
18142 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18143 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18144 non-public-API function ssl_cert_instantiate() is used as a helper
18145 function and also to reduce code redundancy inside ssl_rsa.c.
18147 *Ralf S. Engelschall*
18149 * Move s_server -dcert and -dkey options out of the undocumented feature
18150 area because they are useful for the DSA situation and should be
18151 recognized by the users.
18153 *Ralf S. Engelschall*
18155 * Fix the cipher decision scheme for export ciphers: the export bits are
18156 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18157 SSL_EXP_MASK. So, the original variable has to be used instead of the
18158 already masked variable.
18160 *Richard Levitte <levitte@stacken.kth.se>*
18162 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18164 *Richard Levitte <levitte@stacken.kth.se>*
18166 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18167 from `int` to `unsigned int` because it is a length and initialized by
18168 EVP_DigestFinal() which expects an `unsigned int *`.
18170 *Richard Levitte <levitte@stacken.kth.se>*
18172 * Don't hard-code path to Perl interpreter on shebang line of Configure
18173 script. Instead use the usual Shell->Perl transition trick.
18175 *Ralf S. Engelschall*
18177 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18178 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18179 -noout -modulus` as it's already the case for `openssl rsa -noout
18180 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18181 currently the public key is printed (a decision which was already done by
18182 `openssl dsa -modulus` in the past) which serves a similar purpose.
18183 Additionally the NO_RSA no longer completely removes the whole -modulus
18184 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18187 *Ralf S. Engelschall*
18189 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18190 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18192 *Arne Ansper <arne@ats.cyber.ee>*
18194 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18195 to be added. Now both 'req' and 'ca' can use new objects defined in the
18200 * Add cool BIO that does syslog (or event log on NT).
18202 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18204 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18205 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18206 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18207 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18211 * Add preliminary config info for new extension code.
18215 * Make RSA_NO_PADDING really use no padding.
18217 *Ulf Moeller <ulf@fitug.de>*
18219 * Generate errors when private/public key check is done.
18223 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18224 for some CRL extensions and new objects added.
18228 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18229 key usage extension and fuller support for authority key id.
18233 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18234 padding method for RSA, which is recommended for new applications in PKCS
18235 #1 v2.0 (RFC 2437, October 1998).
18236 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18237 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18238 against Bleichbacher's attack on RSA.
18239 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18242 * Updates to the new SSL compression code
18244 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18246 * Fix so that the version number in the master secret, when passed
18247 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18248 (because the server will not accept higher), that the version number
18249 is 0x03,0x01, not 0x03,0x00
18251 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18253 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18254 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18255 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18259 * Support for RAW extensions where an arbitrary extension can be
18260 created by including its DER encoding. See `apps/openssl.cnf` for
18265 * Make sure latest Perl versions don't interpret some generated C array
18266 code as Perl array code in the crypto/err/err_genc.pl script.
18268 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18270 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18271 not many people have the assembler. Various Win32 compilation fixes and
18272 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18273 build instructions.
18277 * Modify configure script 'Configure' to automatically create crypto/date.h
18278 file under Win32 and also build pem.h from pem.org. New script
18279 util/mkfiles.pl to create the MINFO file on environments that can't do a
18280 'make files': perl util/mkfiles.pl >MINFO should work.
18284 * Major rework of DES function declarations, in the pursuit of correctness
18285 and purity. As a result, many evil casts evaporated, and some weirdness,
18286 too. You may find this causes warnings in your code. Zapping your evil
18287 casts will probably fix them. Mostly.
18291 * Fix for a typo in asn1.h. Bug fix to object creation script
18292 obj_dat.pl. It considered a zero in an object definition to mean
18293 "end of object": none of the objects in objects.h have any zeros
18294 so it wasn't spotted.
18296 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18298 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18299 Masking (CBCM). In the absence of test vectors, the best I have been able
18300 to do is check that the decrypt undoes the encrypt, so far. Send me test
18301 vectors if you have them.
18305 * Correct calculation of key length for export ciphers (too much space was
18306 allocated for null ciphers). This has not been tested!
18310 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18311 message is now correct (it understands "crypto" and "ssl" on its
18312 command line). There is also now an "update" option. This will update
18313 the util/ssleay.num and util/libeay.num files with any new functions.
18315 perl util/mkdef.pl crypto ssl update
18316 it will update them.
18320 * Overhauled the Perl interface:
18321 - ported BN stuff to OpenSSL's different BN library
18322 - made the perl/ source tree CVS-aware
18323 - renamed the package from SSLeay to OpenSSL (the files still contain
18324 their history because I've copied them in the repository)
18325 - removed obsolete files (the test scripts will be replaced
18326 by better Test::Harness variants in the future)
18328 *Ralf S. Engelschall*
18330 * First cut for a very conservative source tree cleanup:
18331 1. merge various obsolete readme texts into doc/ssleay.txt
18332 where we collect the old documents and readme texts.
18333 2. remove the first part of files where I'm already sure that we no
18334 longer need them because of three reasons: either they are just temporary
18335 files which were left by Eric or they are preserved original files where
18336 I've verified that the diff is also available in the CVS via "cvs diff
18337 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18338 the crypto/md/ stuff).
18340 *Ralf S. Engelschall*
18342 * More extension code. Incomplete support for subject and issuer alt
18343 name, issuer and authority key id. Change the i2v function parameters
18344 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18345 what that's for :-) Fix to ASN1 macro which messed up
18346 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18350 * Preliminary support for ENUMERATED type. This is largely copied from the
18355 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18357 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18359 * Make sure `make rehash` target really finds the `openssl` program.
18361 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18363 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18364 like to hear about it if this slows down other processors.
18368 * Add CygWin32 platform information to Configure script.
18370 *Alan Batie <batie@aahz.jf.intel.com>*
18372 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18374 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18376 * New program nseq to manipulate netscape certificate sequences
18380 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18385 * Fixes to BN code. Previously the default was to define BN_RECURSION
18386 but the BN code had some problems that would cause failures when
18387 doing certificate verification and some other functions.
18389 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18391 * Add ASN1 and PEM code to support netscape certificate sequences.
18395 * Add ASN1 and PEM code to support netscape certificate sequences.
18399 * Add several PKIX and private extended key usage OIDs.
18403 * Modify the 'ca' program to handle the new extension code. Modify
18404 openssl.cnf for new extension format, add comments.
18408 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18409 and add a sample to openssl.cnf so req -x509 now adds appropriate
18414 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18415 error code, add initial support to X509_print() and x509 application.
18419 * Takes a deep breath and start adding X509 V3 extension support code. Add
18420 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18421 stuff is currently isolated and isn't even compiled yet.
18425 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18426 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18427 Removed the versions check from X509 routines when loading extensions:
18428 this allows certain broken certificates that don't set the version
18429 properly to be processed.
18433 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18434 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18435 can still be regenerated with "make depend".
18439 * Spelling mistake in C version of CAST-128.
18441 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18443 * Changes to the error generation code. The perl script err-code.pl
18444 now reads in the old error codes and retains the old numbers, only
18445 adding new ones if necessary. It also only changes the .err files if new
18446 codes are added. The makefiles have been modified to only insert errors
18447 when needed (to avoid needlessly modifying header files). This is done
18448 by only inserting errors if the .err file is newer than the auto generated
18449 C file. To rebuild all the error codes from scratch (the old behaviour)
18450 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18451 or delete all the .err files.
18455 * CAST-128 was incorrectly implemented for short keys. The C version has
18456 been fixed, but is untested. The assembler versions are also fixed, but
18457 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18458 to regenerate it if needed.
18459 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18460 Hagino <itojun@kame.net>*
18462 * File was opened incorrectly in randfile.c.
18464 *Ulf Möller <ulf@fitug.de>*
18466 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18467 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18468 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18469 al: it's just almost always a UTCTime. Note this patch adds new error
18470 codes so do a "make errors" if there are problems.
18474 * Correct Linux 1 recognition in config.
18476 *Ulf Möller <ulf@fitug.de>*
18478 * Remove pointless MD5 hash when using DSA keys in ca.
18480 *Anonymous <nobody@replay.com>*
18482 * Generate an error if given an empty string as a cert directory. Also
18483 generate an error if handed NULL (previously returned 0 to indicate an
18484 error, but didn't set one).
18486 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18488 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18492 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18493 parameters. This was causing a warning which killed off the Win32 compile.
18497 * Remove C++ style comments from crypto/bn/bn_local.h.
18499 *Neil Costigan <neil.costigan@celocom.com>*
18501 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18502 based on a text string, looking up short and long names and finally
18503 "dot" format. The "dot" format stuff didn't work. Added new function
18504 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18505 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18506 OID is not part of the table.
18510 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18511 X509_LOOKUP_by_alias().
18515 * Sort openssl functions by name.
18519 * Get the `gendsa` command working and add it to the `list` command. Remove
18520 encryption from sample DSA keys (in case anyone is interested the password
18525 * Make *all* `*_free` functions accept a NULL pointer.
18527 *Frans Heymans <fheymans@isaserver.be>*
18529 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18532 *Anonymous <nobody@replay.com>*
18534 * s_server should send the CAfile as acceptable CAs, not its own cert.
18536 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18538 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18540 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18542 * Temp key "for export" tests were wrong in s3_srvr.c.
18544 *Anonymous <nobody@replay.com>*
18546 * Add prototype for temp key callback functions
18547 SSL_CTX_set_tmp_{rsa,dh}_callback().
18551 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18552 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18556 * X509_name_add_entry() freed the wrong thing after an error.
18558 *Arne Ansper <arne@ats.cyber.ee>*
18560 * rsa_eay.c would attempt to free a NULL context.
18562 *Arne Ansper <arne@ats.cyber.ee>*
18564 * BIO_s_socket() had a broken should_retry() on Windoze.
18566 *Arne Ansper <arne@ats.cyber.ee>*
18568 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18570 *Arne Ansper <arne@ats.cyber.ee>*
18572 * Make sure the already existing X509_STORE->depth variable is initialized
18573 in X509_STORE_new(), but document the fact that this variable is still
18574 unused in the certificate verification process.
18576 *Ralf S. Engelschall*
18578 * Fix the various library and `apps/` files to free up pkeys obtained from
18579 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18583 * Fix reference counting in X509_PUBKEY_get(). This makes
18584 demos/maurice/example2.c work, amongst others, probably.
18586 *Steve Henson and Ben Laurie*
18588 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18589 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18590 are no longer created. This way we have a single and consistent command
18591 line interface `openssl <command>`, similar to `cvs <command>`.
18593 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18595 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18596 BIT STRING wrapper always have zero unused bits.
18600 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18604 * Make the top-level INSTALL documentation easier to understand.
18608 * Makefiles updated to exit if an error occurs in a sub-directory
18609 make (including if user presses ^C) [Paul Sutton]
18611 * Make Montgomery context stuff explicit in RSA data structure.
18615 * Fix build order of pem and err to allow for generated pem.h.
18619 * Fix renumbering bug in X509_NAME_delete_entry().
18623 * Enhanced the err-ins.pl script so it makes the error library number
18624 global and can add a library name. This is needed for external ASN1 and
18625 other error libraries.
18629 * Fixed sk_insert which never worked properly.
18633 * Fix ASN1 macros so they can handle indefinite length constructed
18634 EXPLICIT tags. Some non standard certificates use these: they can now
18639 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18640 into a single doc/ssleay.txt bundle. This way the information is still
18641 preserved but no longer messes up this directory. Now it's new room for
18642 the new set of documentation files.
18644 *Ralf S. Engelschall*
18646 * SETs were incorrectly DER encoded. This was a major pain, because they
18647 shared code with SEQUENCEs, which aren't coded the same. This means that
18648 almost everything to do with SETs or SEQUENCEs has either changed name or
18649 number of arguments.
18651 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18653 * Fix test data to work with the above.
18657 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18658 was already fixed by Eric for 0.9.1 it seems.
18660 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18662 * Autodetect FreeBSD3.
18666 * Fix various bugs in Configure. This affects the following platforms:
18670 unixware-2.0-pentium
18675 * Eliminate generated files from CVS. Reorder tests to regenerate files
18676 before they are needed.
18680 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18684 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18686 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18687 changed SSLeay to OpenSSL in version strings.
18689 *Ralf S. Engelschall*
18691 * Some fixups to the top-level documents.
18695 * Fixed the nasty bug where rsaref.h was not found under compile-time
18696 because the symlink to include/ was missing.
18698 *Ralf S. Engelschall*
18700 * Incorporated the popular no-RSA/DSA-only patches
18701 which allow to compile a RSA-free SSLeay.
18703 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18705 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18706 when "ssleay" is still not found.
18708 *Ralf S. Engelschall*
18710 * Added more platforms to Configure: Cray T3E, HPUX 11,
18712 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18714 * Updated the README file.
18716 *Ralf S. Engelschall*
18718 * Added various .cvsignore files in the CVS repository subdirs
18719 to make a "cvs update" really silent.
18721 *Ralf S. Engelschall*
18723 * Recompiled the error-definition header files and added
18724 missing symbols to the Win32 linker tables.
18726 *Ralf S. Engelschall*
18728 * Cleaned up the top-level documents;
18729 o new files: CHANGES and LICENSE
18730 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18731 o merged COPYRIGHT into LICENSE
18732 o removed obsolete TODO file
18733 o renamed MICROSOFT to INSTALL.W32
18735 *Ralf S. Engelschall*
18737 * Removed dummy files from the 0.9.1b source tree:
18739 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18740 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18741 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18742 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18743 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18746 *Ralf S. Engelschall*
18748 * Added various platform portability fixes.
18752 * The Genesis of the OpenSSL rpject:
18753 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18754 Young and Tim J. Hudson created while they were working for C2Net until
18757 *The OpenSSL Project*
18759 ### Changes between 0.9.0b and 0.9.1b [not released]
18761 * Updated a few CA certificates under certs/
18765 * Changed some BIGNUM api stuff.
18769 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18770 DGUX x86, Linux Alpha, etc.
18774 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18775 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18780 * Add -strparse option to asn1pars program which parses nested
18783 *Dr Stephen Henson <shenson@bigfoot.com>*
18785 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18789 * DSA fix for "ca" program.
18793 * Added "-genkey" option to "dsaparam" program.
18797 * Added RIPE MD160 (rmd160) message digest.
18801 * Added -a (all) option to "ssleay version" command.
18805 * Added PLATFORM define which is the id given to Configure.
18809 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18813 * Extended the ASN.1 parser routines.
18817 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18821 * Added a BN_CTX to the BN library.
18825 * Fixed the weak key values in DES library
18829 * Changed API in EVP library for cipher aliases.
18833 * Added support for RC2/64bit cipher.
18837 * Converted the lhash library to the crypto/mem.c functions.
18841 * Added more recognized ASN.1 object ids.
18845 * Added more RSA padding checks for SSL/TLS.
18849 * Added BIO proxy/filter functionality.
18853 * Added extra_certs to SSL_CTX which can be used
18854 send extra CA certificates to the client in the CA cert chain sending
18855 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18859 * Now Fortezza is denied in the authentication phase because
18860 this is key exchange mechanism is not supported by SSLeay at all.
18864 * Additional PKCS1 checks.
18868 * Support the string "TLSv1" for all TLS v1 ciphers.
18872 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18873 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18877 * Fixed a few memory leaks.
18881 * Fixed various code and comment typos.
18885 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18886 bytes sent in the client random.
18888 *Edward Bishop <ebishop@spyglass.com>*
18892 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18893 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18894 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18895 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18896 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18897 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18898 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18899 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18900 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18901 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18902 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18903 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18904 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18905 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18906 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18907 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18908 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18909 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18910 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18911 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18912 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18913 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18914 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18915 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18916 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18917 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18918 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18919 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18920 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18921 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18922 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18923 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18924 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18925 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18926 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18927 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18928 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18929 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18930 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18931 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18932 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18933 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18934 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18935 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18936 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18937 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18938 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18939 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18940 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18941 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18942 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18943 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18944 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18945 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18946 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18947 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18948 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18949 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18950 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18951 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18952 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18953 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18954 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18955 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18956 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18957 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18958 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18959 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18960 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18961 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18962 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18963 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18964 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18965 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18966 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18967 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18968 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18969 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18970 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18971 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18972 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18973 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18974 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18975 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18976 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18977 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18978 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18979 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18980 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18981 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18982 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18983 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18984 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18985 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18986 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18987 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18988 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18989 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18990 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18991 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18992 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18993 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18994 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18995 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18996 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18997 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18998 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18999 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19000 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19001 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19002 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19003 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19004 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19005 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19006 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19007 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19008 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19009 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19010 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19011 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19012 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19013 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19014 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19015 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19016 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19017 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19018 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19019 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19020 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19021 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19022 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19023 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19024 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19025 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19026 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19027 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19028 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19029 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19030 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19031 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19032 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19033 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19034 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19035 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19036 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19037 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19038 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19039 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19040 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19041 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19042 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19043 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19044 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19045 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19046 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19047 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19048 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19049 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19050 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19051 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19052 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19053 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655