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 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
126 * Validation of SM2 keys has been separated from the validation of regular EC
127 keys, allowing to improve the SM2 validation process to reject loaded private
128 keys that are not conforming to the SM2 ISO standard.
129 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
134 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
135 switches: a validation failure triggers an early exit, returning a failure
136 exit status to the parent process.
140 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
141 to ignore unknown ciphers.
145 * The `-cipher-commands` and `-digest-commands` options
146 of the command line utility `list` have been deprecated.
147 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
151 * All of the low level EC_KEY functions have been deprecated including:
153 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
154 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
155 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
156 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
157 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
158 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
159 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
160 EC_KEY_METHOD_get_verify,
161 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
162 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
163 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
164 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
165 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
166 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
167 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
168 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
169 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
170 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
171 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
172 Applications that need to implement an EC_KEY_METHOD need to consider
173 implementation of the functionality in a special provider.
174 For replacement of the functions manipulating the EC_KEY objects
175 see the EVP_PKEY-EC(7) manual page.
177 Additionally functions that read and write EC_KEY objects such as
178 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
179 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
180 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
181 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
182 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
183 have also been deprecated. Applications should instead use the
184 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
186 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
187 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
188 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
189 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
190 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
192 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
194 * Deprecated all the libcrypto and libssl error string loading
195 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
196 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
197 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
198 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
199 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
200 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
201 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
202 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
203 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
204 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
206 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
207 now loads error strings automatically.
211 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
212 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
213 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
214 are to be used by servers requiring ephemeral DH keys. Instead applications
215 should consider using the built-in DH parameters that are available by
216 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
217 necessary then applications can use the alternative functions
218 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
219 replacement for the "callback" functions. The callback was originally useful
220 in order to have different parameters for export and non-export ciphersuites.
221 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
222 functions should be replaced by one of the other methods described above.
226 * The `-crypt` option to the `passwd` command line tool has been removed.
230 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
235 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
237 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
238 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
239 The inverse ciphers use AES decryption for wrapping, and
240 AES encryption for unwrapping.
244 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
245 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
246 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
247 OpenSSL 3.0 these are replaced by the more generic functions
248 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
249 The old versions have been converted to deprecated macros that just call the
254 * The security callback, which can be customised by application code, supports
255 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
256 in the "other" parameter. In most places this is what is passed. All these
257 places occur server side. However there was one client side call of this
258 security operation and it passed a DH object instead. This is incorrect
259 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
260 of the other locations. Therefore this client side call has been changed to
261 pass an EVP_PKEY instead.
265 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
266 interface. Their functionality remains unchanged.
270 * Added new option for 'openssl list', '-providers', which will display the
271 list of loaded providers, their names, version and status. It optionally
272 displays their gettable parameters.
276 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
277 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
278 type is internally recognised so the workaround is no longer needed.
280 Functionality is still retained as it is, but will only work with
281 EVP_PKEYs with a legacy internal key.
285 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
286 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
290 * Changed all "STACK" functions to be macros instead of inline functions. Macro
291 parameters are still checked for type safety at compile time via helper
296 * Remove the RAND_DRBG API
298 The RAND_DRBG API did not fit well into the new provider concept as
299 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
300 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
301 and some of its API calls are rather low-level. This holds in particular
302 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
304 Adding a compatibility layer to continue supporting the RAND_DRBG API as
305 a legacy API for a regular deprecation period turned out to come at the
306 price of complicating the new provider API unnecessarily. Since the
307 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
310 *Paul Dale and Matthias St. Pierre*
312 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
313 as well as actual hostnames.
317 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
318 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
319 conversely, silently ignore DTLS protocol version bounds when configuring
320 TLS-based contexts. The commands can be repeated to set bounds of both
321 types. The same applies with the corresponding "min_protocol" and
322 "max_protocol" command-line switches, in case some application uses both TLS
325 SSL_CTX instances that are created for a fixed protocol version (e.g.
326 `TLSv1_server_method()`) also silently ignore version bounds. Previously
327 attempts to apply bounds to these protocol versions would result in an
328 error. Now only the "version-flexible" SSL_CTX instances are subject to
329 limits in configuration files in command-line options.
333 * Deprecated the `ENGINE` API. Engines should be replaced with providers
338 * Reworked the recorded ERR codes to make better space for system errors.
339 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
340 given code is a system error (true) or an OpenSSL error (false).
344 * Reworked the test perl framework to better allow parallel testing.
346 *Nicola Tuveri and David von Oheimb*
348 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
349 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
353 * 'Configure' has been changed to figure out the configuration target if
354 none is given on the command line. Consequently, the 'config' script is
355 now only a mere wrapper. All documentation is changed to only mention
358 *Rich Salz and Richard Levitte*
360 * Added a library context `OSSL_LIB_CTX` that applications as well as
361 other libraries can use to form a separate context within which
362 libcrypto operations are performed.
364 There are two ways this can be used:
366 - Directly, by passing a library context to functions that take
367 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
369 - Indirectly, by creating a new library context and then assigning
370 it as the new default, with `OSSL_LIB_CTX_set0_default`.
372 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
373 apart from the functions directly related to `OSSL_LIB_CTX`, accept
374 NULL to indicate that the default library context should be used.
376 Library code that changes the default library context using
377 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
378 second call before returning to the caller.
380 _(Note: the library context was initially called `OPENSSL_CTX` and
381 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
385 * Handshake now fails if Extended Master Secret extension is dropped
390 * Dropped interactive mode from the `openssl` program. From now on,
391 running it without arguments is equivalent to `openssl help`.
395 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
396 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
397 While the old function names have been retained for backward compatibility
398 they should not be used in new developments
399 because their return values are confusing: Unlike other `_cmp()` functions
400 they do not return 0 in case their arguments are equal.
404 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
405 `EC_GROUP_get_field_type()`.
409 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
410 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
411 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
412 Applications should rely on the library automatically assigning a suitable
413 EC_METHOD internally upon EC_GROUP construction.
417 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
418 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
419 assigned internally without application intervention.
420 Users of EC_GROUP_new() should switch to a different suitable constructor.
424 * Add CAdES-BES signature verification support, mostly derived
425 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
427 *Filipe Raimundo da Silva*
429 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
433 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
434 functions are not widely used and now OpenSSL automatically perform this
435 conversion when needed.
439 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
440 EC_KEY_precompute_mult(). These functions are not widely used and
441 applications should instead switch to named curves which OpenSSL has
442 hardcoded lookup tables for.
446 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
447 should instead use the L<EC_POINT_mul(3)> function.
451 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
452 that are not applicable to the new provider model. Applications should
453 instead use EVP_default_properties_is_fips_enabled() and
454 EVP_default_properties_enable_fips().
458 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
459 is set, an unexpected EOF is ignored, it pretends a close notify was received
460 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
464 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
465 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
466 used and applications should instead use the
467 L<EC_POINT_set_affine_coordinates(3)> and
468 L<EC_POINT_get_affine_coordinates(3)> functions.
472 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
473 arrays to be more easily constructed via a series of utility functions.
474 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
475 the various push functions and finally convert to a passable OSSL_PARAM
476 array using OSSL_PARAM_BLD_to_param().
480 * The security strength of SHA1 and MD5 based signatures in TLS has been
481 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
482 working at the default security level of 1 and instead requires security
483 level 0. The security level can be changed either using the cipher string
484 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
488 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
489 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
490 internal keys, if they correspond to one of those built in types.
494 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
495 contain a provider side internal key.
499 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
500 They are old functions that we don't use, and that you could disable with
501 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
505 * Project text documents not yet having a proper file name extension
506 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
507 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
511 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
512 have been converted to Markdown with the goal to produce documents
513 which not only look pretty when viewed online in the browser, but
514 remain well readable inside a plain text editor.
516 To achieve this goal, a 'minimalistic' Markdown style has been applied
517 which avoids formatting elements that interfere too much with the
518 reading flow in the text file. For example, it
520 * avoids [ATX headings][] and uses [setext headings][] instead
521 (which works for `<h1>` and `<h2>` headings only).
522 * avoids [inline links][] and uses [reference links][] instead.
523 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
525 [ATX headings]: https://github.github.com/gfm/#atx-headings
526 [setext headings]: https://github.github.com/gfm/#setext-headings
527 [inline links]: https://github.github.com/gfm/#inline-link
528 [reference links]: https://github.github.com/gfm/#reference-link
529 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
530 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
532 *Matthias St. Pierre*
534 * The test suite is changed to preserve results of each test recipe.
535 A new directory test-runs/ with subdirectories named like the
536 test recipes are created in the build tree for this purpose.
540 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
541 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
542 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
544 *David von Oheimb, Martin Peylo*
546 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
547 The legacy OCSP-focused and only partly documented API is retained for
548 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
552 * Added `util/check-format.pl`, a tool for checking adherence to the
553 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
554 The checks performed are incomplete and yield some false positives.
555 Still the tool should be useful for detecting most typical glitches.
559 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
560 If domain name resolution yields multiple IP addresses all of them are tried
561 after `connect()` failures.
565 * All of the low level RSA functions have been deprecated including:
567 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
568 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
569 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
570 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
571 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
572 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
573 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
574 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
575 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
576 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
577 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
578 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
579 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
580 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
581 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
582 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
583 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
584 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
585 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
586 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
587 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
588 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
589 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
590 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
591 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
592 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
593 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
594 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
596 Use of these low level functions has been informally discouraged for a long
597 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
598 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
599 L<EVP_PKEY_decrypt(3)>.
603 * X509 certificates signed using SHA1 are no longer allowed at security
605 In TLS/SSL the default security level is 1. It can be set either
606 using the cipher string with `@SECLEVEL`, or calling
607 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
608 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
610 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
611 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
612 options of the commands.
616 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
617 modified to use PKEY APIs. These commands are now in maintenance mode
618 and no new features will be added to them.
622 * The command line utility rsautl has been deprecated.
623 Instead use the pkeyutl program.
627 * The command line utilities genrsa and rsa have been modified to use PKEY
628 APIs. They now write PKCS#8 keys by default. These commands are now in
629 maintenance mode and no new features will be added to them.
633 * All of the low level DH functions have been deprecated including:
635 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
636 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
637 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
638 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
639 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
640 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
641 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
642 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
643 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
644 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
645 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
646 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
647 DH_meth_get_generate_params and DH_meth_set_generate_params.
649 Use of these low level functions has been informally discouraged for a long
650 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
651 and L<EVP_PKEY_derive(3)>.
653 Additionally functions that read and write DH objects such as d2i_DHparams,
654 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
655 functions have also been deprecated. Applications should instead use the
656 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
658 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
659 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
660 `EVP_PKEY_set1_DH()` are also deprecated.
661 Applications should instead either read or write an
662 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
663 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
665 *Paul Dale and Matt Caswell*
667 * All of the low level DSA functions have been deprecated including:
669 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
670 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
671 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
672 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
673 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
674 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
675 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
676 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
677 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
678 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
679 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
680 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
681 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
682 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
683 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
684 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
685 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
687 Use of these low level functions has been informally discouraged for a long
688 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
689 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
693 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
694 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
695 This means that applications don't have to look at the curve NID and
696 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
697 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
698 a no-op when the EVP_PKEY is already of the given type.
700 Parameter and key generation is also reworked to make it possible
701 to generate EVP_PKEY_SM2 parameters and keys without having to go
702 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
703 However, code that does the latter will still work as before.
707 * Deprecated low level ECDH and ECDSA functions. These include:
709 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
710 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
713 Use of these low level functions has been informally discouraged for a long
714 time. Instead applications should use the EVP_PKEY_derive(3),
715 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
719 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
720 and EVP_PKEY_decrypt() instead.
721 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
722 and EVP_PKEY_encrypt() instead.
726 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
727 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
728 a new formulation to include all the things it can be used for,
729 as well as words of caution.
733 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
734 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
738 * All of the low level HMAC functions have been deprecated including:
740 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
741 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
744 Use of these low level functions has been informally discouraged for a long
745 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
746 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
747 and L<EVP_MAC_final(3)>.
751 * Over two thousand fixes were made to the documentation, including:
752 - Common options (such as -rand/-writerand, TLS version control, etc)
753 were refactored and point to newly-enhanced descriptions in openssl.pod.
754 - Added style conformance for all options (with help from Richard Levitte),
755 documented all reported missing options, added a CI build to check
756 that all options are documented and that no unimplemented options
758 - Documented some internals, such as all use of environment variables.
759 - Addressed all internal broken L<> references.
763 * All of the low level CMAC functions have been deprecated including:
765 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
766 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
768 Use of these low level functions has been informally discouraged for a long
769 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
770 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
771 and L<EVP_MAC_final(3)>.
775 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
776 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
779 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
780 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
781 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
782 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
783 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
784 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
785 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
786 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
787 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
788 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
790 Use of these low level functions has been informally discouraged
791 for a long time. Applications should use the EVP_DigestInit_ex(3),
792 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
796 * Corrected the documentation of the return values from the `EVP_DigestSign*`
797 set of functions. The documentation mentioned negative values for some
798 errors, but this was never the case, so the mention of negative values
801 Code that followed the documentation and thereby check with something
802 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
806 * All of the low level cipher functions have been deprecated including:
808 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
809 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
810 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
811 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
812 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
813 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
814 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
815 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
816 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
817 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
818 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
819 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
820 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
821 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
822 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
823 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
824 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
825 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
826 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
827 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
828 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
829 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
830 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
831 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
832 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
833 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
834 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
835 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
836 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
838 Use of these low level functions has been informally discouraged for
839 a long time. Applications should use the high level EVP APIs, e.g.
840 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
841 equivalently named decrypt functions instead.
843 *Matt Caswell and Paul Dale*
845 * Removed include/openssl/opensslconf.h.in and replaced it with
846 include/openssl/configuration.h.in, which differs in not including
847 <openssl/macros.h>. A short header include/openssl/opensslconf.h
848 was added to include both.
850 This allows internal hacks where one might need to modify the set
851 of configured macros, for example this if deprecated symbols are
852 still supposed to be available internally:
854 #include <openssl/configuration.h>
856 #undef OPENSSL_NO_DEPRECATED
857 #define OPENSSL_SUPPRESS_DEPRECATED
859 #include <openssl/macros.h>
861 This should not be used by applications that use the exported
862 symbols, as that will lead to linking errors.
866 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
867 used in exponentiation with 512-bit moduli. No EC algorithms are
868 affected. Analysis suggests that attacks against 2-prime RSA1024,
869 3-prime RSA1536, and DSA1024 as a result of this defect would be very
870 difficult to perform and are not believed likely. Attacks against DH512
871 are considered just feasible. However, for an attack the target would
872 have to re-use the DH512 private key, which is not recommended anyway.
873 Also applications directly using the low level API BN_mod_exp may be
874 affected if they use BN_FLG_CONSTTIME.
879 * Most memory-debug features have been deprecated, and the functionality
880 replaced with no-ops.
884 * Added documentation for the STACK API.
888 * Introduced a new method type and API, OSSL_ENCODER, to
889 represent generic encoders. An implementation is expected to
890 be able to encode an object associated with a given name (such
891 as an algorithm name for an asymmetric key) into forms given by
892 implementation properties.
894 Encoders are primarily used from inside libcrypto, through
895 calls to functions like EVP_PKEY_print_private(),
896 PEM_write_bio_PrivateKey() and similar.
898 Encoders are specified in such a way that they can be made to
899 directly handle the provider side portion of an object, if this
900 provider side part comes from the same provider as the encoder
901 itself, but can also be made to handle objects in parametrized
902 form (as an OSSL_PARAM array of data). This allows a provider to
903 offer generic encoders as a service for any other provider.
907 * Added a .pragma directive to the syntax of configuration files, to
908 allow varying behavior in a supported and predictable manner.
909 Currently added pragma:
913 This allows dollar signs to be a keyword character unless it's
914 followed by a opening brace or parenthesis. This is useful for
915 platforms where dollar signs are commonly used in names, such as
916 volume names and system directory names on VMS.
920 * Added functionality to create an EVP_PKEY from user data. This
921 is effectively the same as creating a RSA, DH or DSA object and
922 then assigning them to an EVP_PKEY, but directly using algorithm
923 agnostic EVP functions. A benefit is that this should be future
924 proof for public key algorithms to come.
928 * Change the interpretation of the '--api' configuration option to
929 mean that this is a desired API compatibility level with no
930 further meaning. The previous interpretation, that this would
931 also mean to remove all deprecated symbols up to and including
932 the given version, no requires that 'no-deprecated' is also used
933 in the configuration.
935 When building applications, the desired API compatibility level
936 can be set with the OPENSSL_API_COMPAT macro like before. For
937 API compatibility version below 3.0, the old style numerical
938 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
939 For version 3.0 and on, the value is expected to be the decimal
940 value calculated from the major and minor version like this:
942 MAJOR * 10000 + MINOR * 100
946 -DOPENSSL_API_COMPAT=30000 For 3.0
947 -DOPENSSL_API_COMPAT=30200 For 3.2
949 To hide declarations that are deprecated up to and including the
950 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
951 given when building the application as well.
955 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
956 access to certificate and CRL stores via URIs and OSSL_STORE
959 This adds the following functions:
961 - X509_LOOKUP_store()
962 - X509_STORE_load_file()
963 - X509_STORE_load_path()
964 - X509_STORE_load_store()
965 - SSL_add_store_cert_subjects_to_stack()
966 - SSL_CTX_set_default_verify_store()
967 - SSL_CTX_load_verify_file()
968 - SSL_CTX_load_verify_dir()
969 - SSL_CTX_load_verify_store()
973 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
974 The presence of this system service is determined at run-time.
978 * Added functionality to create an EVP_PKEY context based on data
979 for methods from providers. This takes an algorithm name and a
980 property query string and simply stores them, with the intent
981 that any operation that uses this context will use those strings
982 to fetch the needed methods implicitly, thereby making the port
983 of application written for pre-3.0 OpenSSL easier.
987 * The undocumented function NCONF_WIN32() has been deprecated; for
988 conversion details see the HISTORY section of doc/man5/config.pod
992 * Introduced the new functions EVP_DigestSignInit_ex() and
993 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
994 EVP_DigestVerifyUpdate() have been converted to functions. See the man
995 pages for further details.
999 * Over two thousand fixes were made to the documentation, including:
1000 adding missing command flags, better style conformance, documentation
1003 *Rich Salz, Richard Levitte*
1005 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1006 X25519, X448, Ed25519 and Ed448.
1010 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1015 * Deprecated the public definition of `ERR_STATE` as well as the function
1016 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1021 * Added ERR functionality to give callers access to the stored function
1022 names that have replaced the older function code based functions.
1024 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1025 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1026 ERR_peek_error_all() and ERR_peek_last_error_all().
1028 These functions have become deprecated: ERR_get_error_line(),
1029 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1030 ERR_peek_last_error_line_data() and ERR_func_error_string().
1032 Users are recommended to use ERR_get_error_all(), or to pick information
1033 with ERR_peek functions and finish off with getting the error code by using
1038 * Extended testing to be verbose for failing tests only. The make variables
1039 VERBOSE_FAILURE or VF can be used to enable this:
1041 $ make VF=1 test # Unix
1042 $ mms /macro=(VF=1) test ! OpenVMS
1043 $ nmake VF=1 test # Windows
1047 * Added the `-copy_extensions` option to the `x509` command for use with
1048 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1049 all extensions in the request are copied to the certificate or vice versa.
1051 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1053 * Added the `-copy_extensions` option to the `req` command for use with
1054 `-x509`. When given with the `copy` or `copyall` argument,
1055 all extensions in the certification request are copied to the certificate.
1059 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1060 they generate are by default RFC 5280 compliant in the following sense:
1061 There is a subjectKeyIdentifier extension with a hash value of the public key
1062 and for not self-signed certs there is an authorityKeyIdentifier extension
1063 with a keyIdentifier field or issuer information identifying the signing key.
1064 This is done unless some configuration overrides the new default behavior,
1065 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1069 * Added several checks to `X509_verify_cert()` according to requirements in
1070 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1071 (which may be done by using the CLI option `-x509_strict`):
1072 * The basicConstraints of CA certificates must be marked critical.
1073 * CA certificates must explicitly include the keyUsage extension.
1074 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1075 * The issuer name of any certificate must not be empty.
1076 * The subject name of CA certs, certs with keyUsage crlSign,
1077 and certs without subjectAlternativeName must not be empty.
1078 * If a subjectAlternativeName extension is given it must not be empty.
1079 * The signatureAlgorithm field and the cert signature must be consistent.
1080 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1081 must not be marked critical.
1082 * The authorityKeyIdentifier must be given for X.509v3 certs
1083 unless they are self-signed.
1084 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1088 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1089 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1093 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1094 used even when parsing explicit parameters, when loading a encoded key
1095 or calling `EC_GROUP_new_from_ecpkparameters()`/
1096 `EC_GROUP_new_from_ecparameters()`.
1097 This prevents bypass of security hardening and performance gains,
1098 especially for curves with specialized EC_METHODs.
1099 By default, if a key encoded with explicit parameters is loaded and later
1100 encoded, the output is still encoded with explicit parameters, even if
1101 internally a "named" EC_GROUP is used for computation.
1105 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1106 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1107 NULL. After this change, only the cofactor parameter can be NULL. It also
1108 does some minimal sanity checks on the passed order.
1113 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1114 An attack is simple, if the first CMS_recipientInfo is valid but the
1115 second CMS_recipientInfo is chosen ciphertext. If the second
1116 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1117 encryption key will be replaced by garbage, and the message cannot be
1118 decoded, but if the RSA decryption fails, the correct encryption key is
1119 used and the recipient will not notice the attack.
1120 As a work around for this potential attack the length of the decrypted
1121 key must be equal to the cipher default key length, in case the
1122 certifiate is not given and all recipientInfo are tried out.
1123 The old behaviour can be re-enabled in the CMS code by setting the
1124 CMS_DEBUG_DECRYPT flag.
1128 * Early start up entropy quality from the DEVRANDOM seed source has been
1129 improved for older Linux systems. The RAND subsystem will wait for
1130 /dev/random to be producing output before seeding from /dev/urandom.
1131 The seeded state is stored for future library initialisations using
1132 a system global shared memory segment. The shared memory identifier
1133 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1134 the desired value. The default identifier is 114.
1138 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1139 when primes for RSA keys are computed.
1140 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1141 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1142 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1143 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1144 This avoids possible fingerprinting of newly generated RSA modules.
1148 * Correct the extended master secret constant on EBCDIC systems. Without this
1149 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1150 negotiate EMS will fail. Unfortunately this also means that TLS connections
1151 between EBCDIC systems with this fix, and EBCDIC systems without this
1152 fix will fail if they negotiate EMS.
1156 * Changed the library initialisation so that the config file is now loaded
1157 by default. This was already the case for libssl. It now occurs for both
1158 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1159 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1163 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1164 where the former acts as a replacement for `ERR_put_error()`, and the
1165 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1166 `ERR_raise_data()` adds more flexibility by taking a format string and
1167 an arbitrary number of arguments following it, to be processed with
1172 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1173 to check if a named provider is loaded and available. When called, it
1174 will also activate all fallback providers if such are still present.
1178 * Enforce a minimum DH modulus size of 512 bits.
1182 * Changed DH parameters to generate the order q subgroup instead of 2q.
1183 Previously generated DH parameters are still accepted by DH_check
1184 but DH_generate_key works around that by clearing bit 0 of the
1185 private key for those. This avoids leaking bit 0 of the private key.
1189 * Significantly reduce secure memory usage by the randomness pools.
1193 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1198 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1199 algorithms. An implementation of a key exchange algorithm can be obtained
1200 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1201 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1202 the older EVP_PKEY_derive_init() function. See the man pages for the new
1203 functions for further details.
1207 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1211 * Removed the function names from error messages and deprecated the
1214 * Removed NextStep support and the macro OPENSSL_UNISTD
1218 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1219 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1220 Also removed "export var as function" capability; we do not export
1221 variables, only functions.
1225 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1226 an error and 1 indicating success. In previous versions of OpenSSL this
1227 was a void type. If a key was set longer than the maximum possible this
1232 * Support SM2 signing and verification schemes with X509 certificate.
1236 * Use SHA256 as the default digest for TS query in the `ts` app.
1240 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1241 This checks that the salt length is at least 128 bits, the derived key
1242 length is at least 112 bits, and that the iteration count is at least 1000.
1243 For backwards compatibility these checks are disabled by default in the
1244 default provider, but are enabled by default in the fips provider.
1245 To enable or disable these checks use the control
1246 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1250 * Default cipher lists/suites are now available via a function, the
1251 #defines are deprecated.
1255 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1256 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1257 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1261 * Join the directories crypto/x509 and crypto/x509v3
1265 * Added command 'openssl kdf' that uses the EVP_KDF API.
1269 * Added command 'openssl mac' that uses the EVP_MAC API.
1273 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1274 as default directories. Also added the command 'openssl info'
1275 for scripting purposes.
1279 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1280 deprecated. These undocumented functions were never integrated into the EVP
1281 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1282 Bi-directional IGE mode. These modes were never formally standardised and
1283 usage of these functions is believed to be very small. In particular
1284 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1285 is ever used. The security implications are believed to be minimal, but
1286 this issue was never fixed for backwards compatibility reasons. New code
1287 should not use these modes.
1291 * Add prediction resistance to the DRBG reseeding process.
1295 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1296 mandated by IEEE Std 1619-2018.
1300 * Added newline escaping functionality to a filename when using openssl dgst.
1301 This output format is to replicate the output format found in the `*sum`
1302 checksum programs. This aims to preserve backward compatibility.
1304 *Matt Eaton, Richard Levitte, and Paul Dale*
1306 * Removed the heartbeat message in DTLS feature, as it has very
1307 little usage and doesn't seem to fulfill a valuable purpose.
1308 The configuration option is now deprecated.
1312 * Changed the output of 'openssl {digestname} < file' to display the
1313 digest name in its output.
1317 * Added a new generic trace API which provides support for enabling
1318 instrumentation through trace output. This feature is mainly intended
1319 as an aid for developers and is disabled by default. To utilize it,
1320 OpenSSL needs to be configured with the `enable-trace` option.
1322 If the tracing API is enabled, the application can activate trace output
1323 by registering BIOs as trace channels for a number of tracing and debugging
1326 The `openssl` program has been expanded to enable any of the types
1327 available via environment variables defined by the user, and serves as
1328 one possible example on how to use this functionality.
1330 *Richard Levitte & Matthias St. Pierre*
1332 * Added build tests for C++. These are generated files that only do one
1333 thing, to include one public OpenSSL head file each. This tests that
1334 the public header files can be usefully included in a C++ application.
1336 This test isn't enabled by default. It can be enabled with the option
1337 'enable-buildtest-c++'.
1341 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1345 * Add KMAC to EVP_MAC.
1349 * Added property based algorithm implementation selection framework to
1354 * Added SCA hardening for modular field inversion in EC_GROUP through
1355 a new dedicated field_inv() pointer in EC_METHOD.
1356 This also addresses a leakage affecting conversions from projective
1357 to affine coordinates.
1359 *Billy Bob Brumley, Nicola Tuveri*
1361 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1362 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1363 those algorithms that were already supported through the EVP_PKEY API
1364 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1365 and scrypt are now wrappers that call EVP_KDF.
1369 * Build devcrypto engine as a dynamic engine.
1371 *Eneas U de Queiroz*
1373 * Add keyed BLAKE2 to EVP_MAC.
1377 * Fix a bug in the computation of the endpoint-pair shared secret used
1378 by DTLS over SCTP. This breaks interoperability with older versions
1379 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1380 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1381 interoperability with such broken implementations. However, enabling
1382 this switch breaks interoperability with correct implementations.
1384 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1385 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1389 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1393 * Change the license to the Apache License v2.0.
1397 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1399 - Major releases (indicated by incrementing the MAJOR release number)
1400 may introduce incompatible API/ABI changes.
1401 - Minor releases (indicated by incrementing the MINOR release number)
1402 may introduce new features but retain API/ABI compatibility.
1403 - Patch releases (indicated by incrementing the PATCH number)
1404 are intended for bug fixes and other improvements of existing
1405 features only (like improving performance or adding documentation)
1406 and retain API/ABI compatibility.
1410 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1414 * Remove the 'dist' target and add a tarball building script. The
1415 'dist' target has fallen out of use, and it shouldn't be
1416 necessary to configure just to create a source distribution.
1420 * Recreate the OS390-Unix config target. It no longer relies on a
1421 special script like it did for OpenSSL pre-1.1.0.
1425 * Instead of having the source directories listed in Configure, add
1426 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1431 * Add GMAC to EVP_MAC.
1435 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1439 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1440 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1441 to facilitate the continued use of MACs through raw private keys in
1442 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1446 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1447 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1451 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1452 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1453 are retained for backwards compatibility.
1457 * AES-XTS mode now enforces that its two keys are different to mitigate
1458 the attacked described in "Efficient Instantiations of Tweakable
1459 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1460 Details of this attack can be obtained from:
1461 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1465 * Rename the object files, i.e. give them other names than in previous
1466 versions. Their names now include the name of the final product, as
1467 well as its type mnemonic (bin, lib, shlib).
1471 * Added new option for 'openssl list', '-objects', which will display the
1472 list of built in objects, i.e. OIDs with names.
1476 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1477 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1482 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1483 improves application performance by removing data copies and providing
1484 applications with zero-copy system calls such as sendfile and splice.
1488 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1489 option is set, openssl cleanses (zeroize) plaintext bytes from
1490 internal buffers after delivering them to the application. Note,
1491 the application is still responsible for cleansing other copies
1492 (e.g.: data received by SSL_read(3)).
1496 * `PKCS12_parse` now maintains the order of the parsed certificates
1497 when outputting them via `*ca` (rather than reversing it).
1501 * Deprecated pthread fork support methods. These were unused so no
1502 replacement is required.
1504 - OPENSSL_fork_prepare()
1505 - OPENSSL_fork_parent()
1506 - OPENSSL_fork_child()
1513 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1515 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1517 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1518 create a unique hash value based on the issuer and serial number data
1519 contained within an X509 certificate. However it was failing to correctly
1520 handle any errors that may occur while parsing the issuer field (which might
1521 occur if the issuer field is maliciously constructed). This may subsequently
1522 result in a NULL pointer deref and a crash leading to a potential denial of
1528 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1529 padding mode to correctly check for rollback attacks. This is considered a
1530 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1535 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1536 functions. Previously they could overflow the output length argument in some
1537 cases where the input length is close to the maximum permissable length for
1538 an integer on the platform. In such cases the return value from the function
1539 call would be 1 (indicating success), but the output length value would be
1540 negative. This could cause applications to behave incorrectly or crash.
1545 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1546 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1547 could be exploited in a side channel attack to recover the password. Since
1548 the attack is local host only this is outside of the current OpenSSL
1549 threat model and therefore no CVE is assigned.
1551 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1556 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1558 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1559 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1560 If an attacker can control both items being compared then this could lead
1561 to a possible denial of service attack. OpenSSL itself uses the
1562 GENERAL_NAME_cmp function for two purposes:
1563 1) Comparing CRL distribution point names between an available CRL and a
1564 CRL distribution point embedded in an X509 certificate
1565 2) When verifying that a timestamp response token signer matches the
1566 timestamp authority name (exposed via the API functions
1567 TS_RESP_verify_response and TS_RESP_verify_token)
1572 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1574 * Certificates with explicit curve parameters are now disallowed in
1575 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1579 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1580 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1581 conversely, silently ignore DTLS protocol version bounds when configuring
1582 TLS-based contexts. The commands can be repeated to set bounds of both
1583 types. The same applies with the corresponding "min_protocol" and
1584 "max_protocol" command-line switches, in case some application uses both TLS
1587 SSL_CTX instances that are created for a fixed protocol version (e.g.
1588 TLSv1_server_method()) also silently ignore version bounds. Previously
1589 attempts to apply bounds to these protocol versions would result in an
1590 error. Now only the "version-flexible" SSL_CTX instances are subject to
1591 limits in configuration files in command-line options.
1595 * Handshake now fails if Extended Master Secret extension is dropped
1600 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1602 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1604 * Fixed segmentation fault in SSL_check_chain()
1605 Server or client applications that call the SSL_check_chain() function
1606 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1607 dereference as a result of incorrect handling of the
1608 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1609 or unrecognised signature algorithm is received from the peer. This could
1610 be exploited by a malicious peer in a Denial of Service attack.
1615 * Added AES consttime code for no-asm configurations
1616 an optional constant time support for AES was added
1617 when building openssl for no-asm.
1618 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1619 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1620 At this time this feature is by default disabled.
1621 It will be enabled by default in 3.0.
1625 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1627 * Revert the change of EOF detection while reading in libssl to avoid
1628 regressions in applications depending on the current way of reporting
1629 the EOF. As the existing method is not fully accurate the change to
1630 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1631 branch and will be present in the 3.0 release.
1635 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1636 when primes for RSA keys are computed.
1637 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1638 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1639 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1640 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1641 This avoids possible fingerprinting of newly generated RSA modules.
1645 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1647 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1648 while reading in libssl then we would report an error back to the
1649 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1650 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1651 therefore give a hint as to what went wrong.
1655 * Check that ed25519 and ed448 are allowed by the security level. Previously
1656 signature algorithms not using an MD were not being checked that they were
1657 allowed by the security level.
1661 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1662 was not quite right. The behaviour was not consistent between resumption
1663 and normal handshakes, and also not quite consistent with historical
1664 behaviour. The behaviour in various scenarios has been clarified and
1665 it has been updated to make it match historical behaviour as closely as
1670 * *[VMS only]* The header files that the VMS compilers include automatically,
1671 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1672 that the C++ compiler doesn't understand. This is a shortcoming in the
1673 compiler, but can be worked around with `__cplusplus` guards.
1675 C++ applications that use OpenSSL libraries must be compiled using the
1676 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1677 functions. Otherwise, only functions with symbols of less than 31
1678 characters can be used, as the linker will not be able to successfully
1679 resolve symbols with longer names.
1683 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1684 The presence of this system service is determined at run-time.
1688 * Added newline escaping functionality to a filename when using openssl dgst.
1689 This output format is to replicate the output format found in the `*sum`
1690 checksum programs. This aims to preserve backward compatibility.
1692 *Matt Eaton, Richard Levitte, and Paul Dale*
1694 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1699 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1701 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1702 number generator (RNG). This was intended to include protection in the
1703 event of a fork() system call in order to ensure that the parent and child
1704 processes did not share the same RNG state. However this protection was not
1705 being used in the default case.
1707 A partial mitigation for this issue is that the output from a high
1708 precision timer is mixed into the RNG state so the likelihood of a parent
1709 and child process sharing state is significantly reduced.
1711 If an application already calls OPENSSL_init_crypto() explicitly using
1712 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1715 *Matthias St. Pierre*
1717 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1718 used even when parsing explicit parameters, when loading a encoded key
1719 or calling `EC_GROUP_new_from_ecpkparameters()`/
1720 `EC_GROUP_new_from_ecparameters()`.
1721 This prevents bypass of security hardening and performance gains,
1722 especially for curves with specialized EC_METHODs.
1723 By default, if a key encoded with explicit parameters is loaded and later
1724 encoded, the output is still encoded with explicit parameters, even if
1725 internally a "named" EC_GROUP is used for computation.
1729 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1730 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1731 NULL. After this change, only the cofactor parameter can be NULL. It also
1732 does some minimal sanity checks on the passed order.
1737 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1738 An attack is simple, if the first CMS_recipientInfo is valid but the
1739 second CMS_recipientInfo is chosen ciphertext. If the second
1740 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1741 encryption key will be replaced by garbage, and the message cannot be
1742 decoded, but if the RSA decryption fails, the correct encryption key is
1743 used and the recipient will not notice the attack.
1744 As a work around for this potential attack the length of the decrypted
1745 key must be equal to the cipher default key length, in case the
1746 certifiate is not given and all recipientInfo are tried out.
1747 The old behaviour can be re-enabled in the CMS code by setting the
1748 CMS_DEBUG_DECRYPT flag.
1753 * Early start up entropy quality from the DEVRANDOM seed source has been
1754 improved for older Linux systems. The RAND subsystem will wait for
1755 /dev/random to be producing output before seeding from /dev/urandom.
1756 The seeded state is stored for future library initialisations using
1757 a system global shared memory segment. The shared memory identifier
1758 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1759 the desired value. The default identifier is 114.
1763 * Correct the extended master secret constant on EBCDIC systems. Without this
1764 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1765 negotiate EMS will fail. Unfortunately this also means that TLS connections
1766 between EBCDIC systems with this fix, and EBCDIC systems without this
1767 fix will fail if they negotiate EMS.
1771 * Use Windows installation paths in the mingw builds
1773 Mingw isn't a POSIX environment per se, which means that Windows
1774 paths should be used for installation.
1779 * Changed DH_check to accept parameters with order q and 2q subgroups.
1780 With order 2q subgroups the bit 0 of the private key is not secret
1781 but DH_generate_key works around that by clearing bit 0 of the
1782 private key for those. This avoids leaking bit 0 of the private key.
1786 * Significantly reduce secure memory usage by the randomness pools.
1790 * Revert the DEVRANDOM_WAIT feature for Linux systems
1792 The DEVRANDOM_WAIT feature added a select() call to wait for the
1793 /dev/random device to become readable before reading from the
1794 /dev/urandom device.
1796 It turned out that this change had negative side effects on
1797 performance which were not acceptable. After some discussion it
1798 was decided to revert this feature and leave it up to the OS
1799 resp. the platform maintainer to ensure a proper initialization
1800 during early boot time.
1802 *Matthias St. Pierre*
1804 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1806 * Add build tests for C++. These are generated files that only do one
1807 thing, to include one public OpenSSL head file each. This tests that
1808 the public header files can be usefully included in a C++ application.
1810 This test isn't enabled by default. It can be enabled with the option
1811 'enable-buildtest-c++'.
1815 * Enable SHA3 pre-hashing for ECDSA and DSA.
1819 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1820 This changes the size when using the `genpkey` command when no size is given.
1821 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1822 generation commands to use 2048 bits by default.
1826 * Reorganize the manual pages to consistently have RETURN VALUES,
1827 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1828 util/fix-doc-nits accordingly.
1830 *Paul Yang, Joshua Lock*
1832 * Add the missing accessor EVP_PKEY_get0_engine()
1836 * Have commands like `s_client` and `s_server` output the signature scheme
1837 along with other cipher suite parameters when debugging.
1841 * Make OPENSSL_config() error agnostic again.
1845 * Do the error handling in RSA decryption constant time.
1849 * Prevent over long nonces in ChaCha20-Poly1305.
1851 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1852 for every encryption operation. RFC 7539 specifies that the nonce value
1853 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1854 and front pads the nonce with 0 bytes if it is less than 12
1855 bytes. However it also incorrectly allows a nonce to be set of up to 16
1856 bytes. In this case only the last 12 bytes are significant and any
1857 additional leading bytes are ignored.
1859 It is a requirement of using this cipher that nonce values are
1860 unique. Messages encrypted using a reused nonce value are susceptible to
1861 serious confidentiality and integrity attacks. If an application changes
1862 the default nonce length to be longer than 12 bytes and then makes a
1863 change to the leading bytes of the nonce expecting the new value to be a
1864 new unique nonce then such an application could inadvertently encrypt
1865 messages with a reused nonce.
1867 Additionally the ignored bytes in a long nonce are not covered by the
1868 integrity guarantee of this cipher. Any application that relies on the
1869 integrity of these ignored leading bytes of a long nonce may be further
1870 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1871 is safe because no such use sets such a long nonce value. However user
1872 applications that use this cipher directly and set a non-default nonce
1873 length to be longer than 12 bytes may be vulnerable.
1875 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1881 * Add DEVRANDOM_WAIT feature for Linux systems
1883 On older Linux systems where the getrandom() system call is not available,
1884 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1885 Contrary to getrandom(), the /dev/urandom device will not block during
1886 early boot when the kernel CSPRNG has not been seeded yet.
1888 To mitigate this known weakness, use select() to wait for /dev/random to
1889 become readable before reading from /dev/urandom.
1891 * Ensure that SM2 only uses SM3 as digest algorithm
1895 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1897 * Change the info callback signals for the start and end of a post-handshake
1898 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1899 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1900 confused by this and assume that a TLSv1.2 renegotiation has started. This
1901 can break KeyUpdate handling. Instead we no longer signal the start and end
1902 of a post handshake message exchange (although the messages themselves are
1903 still signalled). This could break some applications that were expecting
1904 the old signals. However without this KeyUpdate is not usable for many
1909 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1911 * Timing vulnerability in DSA signature generation
1913 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1914 timing side channel attack. An attacker could use variations in the signing
1915 algorithm to recover the private key.
1917 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1922 * Timing vulnerability in ECDSA signature generation
1924 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1925 timing side channel attack. An attacker could use variations in the signing
1926 algorithm to recover the private key.
1928 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1933 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1934 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1935 of two gigabytes and the error handling improved.
1937 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1938 categorized as a normal bug, not a security issue, because the DRBG reseeds
1939 automatically and is fully functional even without additional randomness
1940 provided by the application.
1942 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1944 * Add a new ClientHello callback. Provides a callback interface that gives
1945 the application the ability to adjust the nascent SSL object at the
1946 earliest stage of ClientHello processing, immediately after extensions have
1947 been collected but before they have been processed. In particular, this
1948 callback can adjust the supported TLS versions in response to the contents
1953 * Add SM2 base algorithm support.
1957 * s390x assembly pack: add (improved) hardware-support for the following
1958 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1959 aes-cfb/cfb8, aes-ecb.
1963 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1964 parameter is no longer accepted, as it leads to a corrupt table. NULL
1965 pem_str is reserved for alias entries only.
1969 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1970 step for prime curves. The new implementation is based on formulae from
1971 differential addition-and-doubling in homogeneous projective coordinates
1972 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1973 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1974 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1975 to work in projective coordinates.
1977 *Billy Bob Brumley, Nicola Tuveri*
1979 * Change generating and checking of primes so that the error rate of not
1980 being prime depends on the intended use based on the size of the input.
1981 For larger primes this will result in more rounds of Miller-Rabin.
1982 The maximal error rate for primes with more than 1080 bits is lowered
1985 *Kurt Roeckx, Annie Yousar*
1987 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1991 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1992 moving between systems, and to avoid confusion when a Windows build is
1993 done with mingw vs with MSVC. For POSIX installs, there's still a
1994 symlink or copy named 'tsget' to avoid that confusion as well.
1998 * Revert blinding in ECDSA sign and instead make problematic addition
1999 length-invariant. Switch even to fixed-length Montgomery multiplication.
2003 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2004 step for binary curves. The new implementation is based on formulae from
2005 differential addition-and-doubling in mixed Lopez-Dahab projective
2006 coordinates, modified to independently blind the operands.
2008 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2010 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2011 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2012 EC_METHODs to implement their own specialized "ladder step", to take
2013 advantage of more favorable coordinate systems or more efficient
2014 differential addition-and-doubling algorithms.
2016 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2018 * Modified the random device based seed sources to keep the relevant
2019 file descriptors open rather than reopening them on each access.
2020 This allows such sources to operate in a chroot() jail without
2021 the associated device nodes being available. This behaviour can be
2022 controlled using RAND_keep_random_devices_open().
2026 * Numerous side-channel attack mitigations have been applied. This may have
2027 performance impacts for some algorithms for the benefit of improved
2028 security. Specific changes are noted in this change log by their respective
2033 * AIX shared library support overhaul. Switch to AIX "natural" way of
2034 handling shared libraries, which means collecting shared objects of
2035 different versions and bitnesses in one common archive. This allows to
2036 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2037 doesn't affect the way 3rd party applications are linked, only how
2038 multi-version installation is managed.
2042 * Make ec_group_do_inverse_ord() more robust and available to other
2043 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2044 mitigations are applied to the fallback BN_mod_inverse().
2045 When using this function rather than BN_mod_inverse() directly, new
2046 EC cryptosystem implementations are then safer-by-default.
2050 * Add coordinate blinding for EC_POINT and implement projective
2051 coordinate blinding for generic prime curves as a countermeasure to
2052 chosen point SCA attacks.
2054 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2056 * Add blinding to ECDSA and DSA signatures to protect against side channel
2057 attacks discovered by Keegan Ryan (NCC Group).
2061 * Enforce checking in the `pkeyutl` command to ensure that the input
2062 length does not exceed the maximum supported digest length when performing
2063 a sign, verify or verifyrecover operation.
2067 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2068 I/O in combination with something like select() or poll() will hang. This
2069 can be turned off again using SSL_CTX_clear_mode().
2070 Many applications do not properly handle non-application data records, and
2071 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2072 around the problems in those applications, but can also break some.
2073 It's recommended to read the manpages about SSL_read(), SSL_write(),
2074 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2075 SSL_CTX_set_read_ahead() again.
2079 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2080 now allow empty (zero character) pass phrases.
2084 * Apply blinding to binary field modular inversion and remove patent
2085 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2089 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2090 binary and prime elliptic curves.
2094 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2095 constant time fixed point multiplication.
2099 * Revise elliptic curve scalar multiplication with timing attack
2100 defenses: ec_wNAF_mul redirects to a constant time implementation
2101 when computing fixed point and variable point multiplication (which
2102 in OpenSSL are mostly used with secret scalars in keygen, sign,
2103 ECDH derive operations).
2104 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2107 * Updated CONTRIBUTING
2111 * Updated DRBG / RAND to request nonce and additional low entropy
2112 randomness from the system.
2114 *Matthias St. Pierre*
2116 * Updated 'openssl rehash' to use OpenSSL consistent default.
2120 * Moved the load of the ssl_conf module to libcrypto, which helps
2121 loading engines that libssl uses before libssl is initialised.
2125 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2129 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2131 *Ingo Schwarze, Rich Salz*
2133 * Added output of accepting IP address and port for 'openssl s_server'
2137 * Added a new API for TLSv1.3 ciphersuites:
2138 SSL_CTX_set_ciphersuites()
2139 SSL_set_ciphersuites()
2143 * Memory allocation failures consistently add an error to the error
2148 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2149 in libcrypto when run as setuid/setgid.
2153 * Load any config file by default when libssl is used.
2157 * Added new public header file <openssl/rand_drbg.h> and documentation
2158 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2160 *Matthias St. Pierre*
2162 * QNX support removed (cannot find contributors to get their approval
2163 for the license change).
2167 * TLSv1.3 replay protection for early data has been implemented. See the
2168 SSL_read_early_data() man page for further details.
2172 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2173 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2174 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2175 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2176 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2177 configuration has been separated out. See the ciphers man page or the
2178 SSL_CTX_set_ciphersuites() man page for more information.
2182 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2183 in responder mode now supports the new "-multi" option, which
2184 spawns the specified number of child processes to handle OCSP
2185 requests. The "-timeout" option now also limits the OCSP
2186 responder's patience to wait to receive the full client request
2187 on a newly accepted connection. Child processes are respawned
2188 as needed, and the CA index file is automatically reloaded
2189 when changed. This makes it possible to run the "ocsp" responder
2190 as a long-running service, making the OpenSSL CA somewhat more
2191 feature-complete. In this mode, most diagnostic messages logged
2192 after entering the event loop are logged via syslog(3) rather than
2197 * Added support for X448 and Ed448. Heavily based on original work by
2202 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2203 objects loaded. This adds the functions OSSL_STORE_expect() and
2204 OSSL_STORE_find() as well as needed tools to construct searches and
2205 get the search data out of them.
2209 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2210 version of OpenSSL should review their configuration settings to ensure
2211 that they are still appropriate for TLSv1.3. For further information see:
2212 <https://wiki.openssl.org/index.php/TLS1.3>
2216 * Grand redesign of the OpenSSL random generator
2218 The default RAND method now utilizes an AES-CTR DRBG according to
2219 NIST standard SP 800-90Ar1. The new random generator is essentially
2220 a port of the default random generator from the OpenSSL FIPS 2.0
2221 object module. It is a hybrid deterministic random bit generator
2222 using an AES-CTR bit stream and which seeds and reseeds itself
2223 automatically using trusted system entropy sources.
2225 Some of its new features are:
2226 - Support for multiple DRBG instances with seed chaining.
2227 - The default RAND method makes use of a DRBG.
2228 - There is a public and private DRBG instance.
2229 - The DRBG instances are fork-safe.
2230 - Keep all global DRBG instances on the secure heap if it is enabled.
2231 - The public and private DRBG instance are per thread for lock free
2234 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2236 * Changed Configure so it only says what it does and doesn't dump
2237 so much data. Instead, ./configdata.pm should be used as a script
2238 to display all sorts of configuration data.
2242 * Added processing of "make variables" to Configure.
2246 * Added SHA512/224 and SHA512/256 algorithm support.
2250 * The last traces of Netware support, first removed in 1.1.0, have
2255 * Get rid of Makefile.shared, and in the process, make the processing
2256 of certain files (rc.obj, or the .def/.map/.opt files produced from
2257 the ordinal files) more visible and hopefully easier to trace and
2258 debug (or make silent).
2262 * Make it possible to have environment variable assignments as
2263 arguments to config / Configure.
2267 * Add multi-prime RSA (RFC 8017) support.
2271 * Add SM3 implemented according to GB/T 32905-2016
2272 *Jack Lloyd <jack.lloyd@ribose.com>,*
2273 *Ronald Tse <ronald.tse@ribose.com>,*
2274 *Erick Borsboom <erick.borsboom@ribose.com>*
2276 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2277 as documented in RFC6066.
2278 Based on a patch from Tomasz Moń
2280 *Filipe Raimundo da Silva*
2282 * Add SM4 implemented according to GB/T 32907-2016.
2283 *Jack Lloyd <jack.lloyd@ribose.com>,*
2284 *Ronald Tse <ronald.tse@ribose.com>,*
2285 *Erick Borsboom <erick.borsboom@ribose.com>*
2287 * Reimplement -newreq-nodes and ERR_error_string_n; the
2288 original author does not agree with the license change.
2292 * Add ARIA AEAD TLS support.
2296 * Some macro definitions to support VS6 have been removed. Visual
2297 Studio 6 has not worked since 1.1.0
2301 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2302 without clearing the errors.
2306 * Add "atfork" functions. If building on a system that without
2307 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2308 requirements. The RAND facility now uses/requires this.
2316 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2317 not possible to disable entirely. However, it's still possible to
2318 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2321 To disable, configure with 'no-ui-console'. 'no-ui' is still
2322 possible to use as an alias. Check at compile time with the
2323 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2324 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2328 * Add a STORE module, which implements a uniform and URI based reader of
2329 stores that can contain keys, certificates, CRLs and numerous other
2330 objects. The main API is loosely based on a few stdio functions,
2331 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2332 OSSL_STORE_error and OSSL_STORE_close.
2333 The implementation uses backends called "loaders" to implement arbitrary
2334 URI schemes. There is one built in "loader" for the 'file' scheme.
2338 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2339 then adjusted to work on FreeBSD 8.4 as well.
2340 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2341 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2345 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2346 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2347 error code calls like this:
2349 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2351 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2352 that can be encoded in C. For the foreseeable future, this will only
2355 *Richard Levitte and Tim Hudson*
2357 * Removed BSD cryptodev engine.
2361 * Add a build target 'build_all_generated', to build all generated files
2362 and only that. This can be used to prepare everything that requires
2363 things like perl for a system that lacks perl and then move everything
2364 to that system and do the rest of the build there.
2368 * In the UI interface, make it possible to duplicate the user data. This
2369 can be used by engines that need to retain the data for a longer time
2370 than just the call where this user data is passed.
2374 * Ignore the '-named_curve auto' value for compatibility of applications
2377 *Tomáš Mráz <tmraz@fedoraproject.org>*
2379 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2380 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2381 alerts across multiple records (some of which could be empty). In practice
2382 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2383 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2384 support this at all. Supporting it adds significant complexity to the
2385 record layer, and its removal is unlikely to cause interoperability
2390 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2391 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2392 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2397 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2398 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2400 *Richard Levitte, Andy Polyakov*
2402 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2407 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2408 platform rather than 'mingw'.
2412 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2413 success if they are asked to add an object which already exists
2414 in the store. This change cascades to other functions which load
2415 certificates and CRLs.
2419 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2420 facilitate stack unwinding even from assembly subroutines.
2424 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2425 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2429 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2430 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2431 which is the minimum version we support.
2435 * Certificate time validation (X509_cmp_time) enforces stricter
2436 compliance with RFC 5280. Fractional seconds and timezone offsets
2437 are no longer allowed.
2441 * Add support for ARIA
2445 * s_client will now send the Server Name Indication (SNI) extension by
2446 default unless the new "-noservername" option is used. The server name is
2447 based on the host provided to the "-connect" option unless overridden by
2448 using "-servername".
2452 * Add support for SipHash
2456 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2457 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2458 prevent issues where no progress is being made and the peer continually
2459 sends unrecognised record types, using up resources processing them.
2463 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2464 using the algorithm defined in
2465 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2469 * Heartbeat support has been removed; the ABI is changed for now.
2471 *Richard Levitte, Rich Salz*
2473 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2477 * The RSA "null" method, which was partially supported to avoid patent
2478 issues, has been replaced to always returns NULL.
2485 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2487 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2488 used even when parsing explicit parameters, when loading a encoded key
2489 or calling `EC_GROUP_new_from_ecpkparameters()`/
2490 `EC_GROUP_new_from_ecparameters()`.
2491 This prevents bypass of security hardening and performance gains,
2492 especially for curves with specialized EC_METHODs.
2493 By default, if a key encoded with explicit parameters is loaded and later
2494 encoded, the output is still encoded with explicit parameters, even if
2495 internally a "named" EC_GROUP is used for computation.
2499 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2500 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2501 NULL. After this change, only the cofactor parameter can be NULL. It also
2502 does some minimal sanity checks on the passed order.
2507 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2508 An attack is simple, if the first CMS_recipientInfo is valid but the
2509 second CMS_recipientInfo is chosen ciphertext. If the second
2510 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2511 encryption key will be replaced by garbage, and the message cannot be
2512 decoded, but if the RSA decryption fails, the correct encryption key is
2513 used and the recipient will not notice the attack.
2514 As a work around for this potential attack the length of the decrypted
2515 key must be equal to the cipher default key length, in case the
2516 certifiate is not given and all recipientInfo are tried out.
2517 The old behaviour can be re-enabled in the CMS code by setting the
2518 CMS_DEBUG_DECRYPT flag.
2523 * Use Windows installation paths in the mingw builds
2525 Mingw isn't a POSIX environment per se, which means that Windows
2526 paths should be used for installation.
2531 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2533 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2534 This changes the size when using the `genpkey` command when no size is given.
2535 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2536 generation commands to use 2048 bits by default.
2540 * Prevent over long nonces in ChaCha20-Poly1305.
2542 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2543 for every encryption operation. RFC 7539 specifies that the nonce value
2544 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2545 and front pads the nonce with 0 bytes if it is less than 12
2546 bytes. However it also incorrectly allows a nonce to be set of up to 16
2547 bytes. In this case only the last 12 bytes are significant and any
2548 additional leading bytes are ignored.
2550 It is a requirement of using this cipher that nonce values are
2551 unique. Messages encrypted using a reused nonce value are susceptible to
2552 serious confidentiality and integrity attacks. If an application changes
2553 the default nonce length to be longer than 12 bytes and then makes a
2554 change to the leading bytes of the nonce expecting the new value to be a
2555 new unique nonce then such an application could inadvertently encrypt
2556 messages with a reused nonce.
2558 Additionally the ignored bytes in a long nonce are not covered by the
2559 integrity guarantee of this cipher. Any application that relies on the
2560 integrity of these ignored leading bytes of a long nonce may be further
2561 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2562 is safe because no such use sets such a long nonce value. However user
2563 applications that use this cipher directly and set a non-default nonce
2564 length to be longer than 12 bytes may be vulnerable.
2566 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2572 * Added SCA hardening for modular field inversion in EC_GROUP through
2573 a new dedicated field_inv() pointer in EC_METHOD.
2574 This also addresses a leakage affecting conversions from projective
2575 to affine coordinates.
2577 *Billy Bob Brumley, Nicola Tuveri*
2579 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2580 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2584 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2588 * Remove the 'dist' target and add a tarball building script. The
2589 'dist' target has fallen out of use, and it shouldn't be
2590 necessary to configure just to create a source distribution.
2594 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2596 * Timing vulnerability in DSA signature generation
2598 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2599 timing side channel attack. An attacker could use variations in the signing
2600 algorithm to recover the private key.
2602 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2607 * Timing vulnerability in ECDSA signature generation
2609 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2610 timing side channel attack. An attacker could use variations in the signing
2611 algorithm to recover the private key.
2613 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2618 * Add coordinate blinding for EC_POINT and implement projective
2619 coordinate blinding for generic prime curves as a countermeasure to
2620 chosen point SCA attacks.
2622 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2624 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2626 * Client DoS due to large DH parameter
2628 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2629 malicious server can send a very large prime value to the client. This will
2630 cause the client to spend an unreasonably long period of time generating a
2631 key for this prime resulting in a hang until the client has finished. This
2632 could be exploited in a Denial Of Service attack.
2634 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2639 * Cache timing vulnerability in RSA Key Generation
2641 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2642 a cache timing side channel attack. An attacker with sufficient access to
2643 mount cache timing attacks during the RSA key generation process could
2644 recover the private key.
2646 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2647 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2652 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2653 parameter is no longer accepted, as it leads to a corrupt table. NULL
2654 pem_str is reserved for alias entries only.
2658 * Revert blinding in ECDSA sign and instead make problematic addition
2659 length-invariant. Switch even to fixed-length Montgomery multiplication.
2663 * Change generating and checking of primes so that the error rate of not
2664 being prime depends on the intended use based on the size of the input.
2665 For larger primes this will result in more rounds of Miller-Rabin.
2666 The maximal error rate for primes with more than 1080 bits is lowered
2669 *Kurt Roeckx, Annie Yousar*
2671 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2675 * Add blinding to ECDSA and DSA signatures to protect against side channel
2676 attacks discovered by Keegan Ryan (NCC Group).
2680 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2681 now allow empty (zero character) pass phrases.
2685 * Certificate time validation (X509_cmp_time) enforces stricter
2686 compliance with RFC 5280. Fractional seconds and timezone offsets
2687 are no longer allowed.
2691 * Fixed a text canonicalisation bug in CMS
2693 Where a CMS detached signature is used with text content the text goes
2694 through a canonicalisation process first prior to signing or verifying a
2695 signature. This process strips trailing space at the end of lines, converts
2696 line terminators to CRLF and removes additional trailing line terminators
2697 at the end of a file. A bug in the canonicalisation process meant that
2698 some characters, such as form-feed, were incorrectly treated as whitespace
2699 and removed. This is contrary to the specification (RFC5485). This fix
2700 could mean that detached text data signed with an earlier version of
2701 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2702 signed with a fixed OpenSSL may fail to verify with an earlier version of
2703 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2704 and use the "-binary" flag (for the "cms" command line application) or set
2705 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2709 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2711 * Constructed ASN.1 types with a recursive definition could exceed the stack
2713 Constructed ASN.1 types with a recursive definition (such as can be found
2714 in PKCS7) could eventually exceed the stack given malicious input with
2715 excessive recursion. This could result in a Denial Of Service attack. There
2716 are no such structures used within SSL/TLS that come from untrusted sources
2717 so this is considered safe.
2719 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2725 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2727 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2728 effectively reduced to only comparing the least significant bit of each
2729 byte. This allows an attacker to forge messages that would be considered as
2730 authenticated in an amount of tries lower than that guaranteed by the
2731 security claims of the scheme. The module can only be compiled by the
2732 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2734 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2740 * Add a build target 'build_all_generated', to build all generated files
2741 and only that. This can be used to prepare everything that requires
2742 things like perl for a system that lacks perl and then move everything
2743 to that system and do the rest of the build there.
2747 * Backport SSL_OP_NO_RENGOTIATION
2749 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2750 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2751 changes this is no longer possible in 1.1.0. Therefore the new
2752 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2753 1.1.0 to provide equivalent functionality.
2755 Note that if an application built against 1.1.0h headers (or above) is run
2756 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2757 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2761 * Removed the OS390-Unix config target. It relied on a script that doesn't
2766 * rsaz_1024_mul_avx2 overflow bug on x86_64
2768 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2769 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2770 Analysis suggests that attacks against RSA and DSA as a result of this
2771 defect would be very difficult to perform and are not believed likely.
2772 Attacks against DH1024 are considered just feasible, because most of the
2773 work necessary to deduce information about a private key may be performed
2774 offline. The amount of resources required for such an attack would be
2775 significant. However, for an attack on TLS to be meaningful, the server
2776 would have to share the DH1024 private key among multiple clients, which is
2777 no longer an option since CVE-2016-0701.
2779 This only affects processors that support the AVX2 but not ADX extensions
2780 like Intel Haswell (4th generation).
2782 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2783 was originally found via the OSS-Fuzz project.
2788 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2790 * bn_sqrx8x_internal carry bug on x86_64
2792 There is a carry propagating bug in the x86_64 Montgomery squaring
2793 procedure. No EC algorithms are affected. Analysis suggests that attacks
2794 against RSA and DSA as a result of this defect would be very difficult to
2795 perform and are not believed likely. Attacks against DH are considered just
2796 feasible (although very difficult) because most of the work necessary to
2797 deduce information about a private key may be performed offline. The amount
2798 of resources required for such an attack would be very significant and
2799 likely only accessible to a limited number of attackers. An attacker would
2800 additionally need online access to an unpatched system using the target
2801 private key in a scenario with persistent DH parameters and a private
2802 key that is shared between multiple clients.
2804 This only affects processors that support the BMI1, BMI2 and ADX extensions
2805 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2807 This issue was reported to OpenSSL by the OSS-Fuzz project.
2812 * Malformed X.509 IPAddressFamily could cause OOB read
2814 If an X.509 certificate has a malformed IPAddressFamily extension,
2815 OpenSSL could do a one-byte buffer overread. The most likely result
2816 would be an erroneous display of the certificate in text format.
2818 This issue was reported to OpenSSL by the OSS-Fuzz project.
2823 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2825 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2826 platform rather than 'mingw'.
2830 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2831 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2832 which is the minimum version we support.
2836 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2838 * Encrypt-Then-Mac renegotiation crash
2840 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2841 negotiated where it was not in the original handshake (or vice-versa) then
2842 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2843 and servers are affected.
2845 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2850 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2852 * Truncated packet could crash via OOB read
2854 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2855 cipher is being used, then a truncated packet can cause that host to
2856 perform an out-of-bounds read, usually resulting in a crash.
2858 This issue was reported to OpenSSL by Robert Święcki of Google.
2863 * Bad (EC)DHE parameters cause a client crash
2865 If a malicious server supplies bad parameters for a DHE or ECDHE key
2866 exchange then this can result in the client attempting to dereference a
2867 NULL pointer leading to a client crash. This could be exploited in a Denial
2870 This issue was reported to OpenSSL by Guido Vranken.
2875 * BN_mod_exp may produce incorrect results on x86_64
2877 There is a carry propagating bug in the x86_64 Montgomery squaring
2878 procedure. No EC algorithms are affected. Analysis suggests that attacks
2879 against RSA and DSA as a result of this defect would be very difficult to
2880 perform and are not believed likely. Attacks against DH are considered just
2881 feasible (although very difficult) because most of the work necessary to
2882 deduce information about a private key may be performed offline. The amount
2883 of resources required for such an attack would be very significant and
2884 likely only accessible to a limited number of attackers. An attacker would
2885 additionally need online access to an unpatched system using the target
2886 private key in a scenario with persistent DH parameters and a private
2887 key that is shared between multiple clients. For example this can occur by
2888 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2889 similar to CVE-2015-3193 but must be treated as a separate problem.
2891 This issue was reported to OpenSSL by the OSS-Fuzz project.
2896 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2898 * ChaCha20/Poly1305 heap-buffer-overflow
2900 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2901 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2902 crash. This issue is not considered to be exploitable beyond a DoS.
2904 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2909 * CMS Null dereference
2911 Applications parsing invalid CMS structures can crash with a NULL pointer
2912 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2913 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2914 structure callback if an attempt is made to free certain invalid encodings.
2915 Only CHOICE structures using a callback which do not handle NULL value are
2918 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2923 * Montgomery multiplication may produce incorrect results
2925 There is a carry propagating bug in the Broadwell-specific Montgomery
2926 multiplication procedure that handles input lengths divisible by, but
2927 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2928 and DH private keys are impossible. This is because the subroutine in
2929 question is not used in operations with the private key itself and an input
2930 of the attacker's direct choice. Otherwise the bug can manifest itself as
2931 transient authentication and key negotiation failures or reproducible
2932 erroneous outcome of public-key operations with specially crafted input.
2933 Among EC algorithms only Brainpool P-512 curves are affected and one
2934 presumably can attack ECDH key negotiation. Impact was not analyzed in
2935 detail, because pre-requisites for attack are considered unlikely. Namely
2936 multiple clients have to choose the curve in question and the server has to
2937 share the private key among them, neither of which is default behaviour.
2938 Even then only clients that chose the curve will be affected.
2940 This issue was publicly reported as transient failures and was not
2941 initially recognized as a security issue. Thanks to Richard Morgan for
2942 providing reproducible case.
2947 * Removed automatic addition of RPATH in shared libraries and executables,
2948 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2952 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2954 * Fix Use After Free for large message sizes
2956 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2957 message larger than approx 16k is received then the underlying buffer to
2958 store the incoming message is reallocated and moved. Unfortunately a
2959 dangling pointer to the old location is left which results in an attempt to
2960 write to the previously freed location. This is likely to result in a
2961 crash, however it could potentially lead to execution of arbitrary code.
2963 This issue only affects OpenSSL 1.1.0a.
2965 This issue was reported to OpenSSL by Robert Święcki.
2970 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2972 * OCSP Status Request extension unbounded memory growth
2974 A malicious client can send an excessively large OCSP Status Request
2975 extension. If that client continually requests renegotiation, sending a
2976 large OCSP Status Request extension each time, then there will be unbounded
2977 memory growth on the server. This will eventually lead to a Denial Of
2978 Service attack through memory exhaustion. Servers with a default
2979 configuration are vulnerable even if they do not support OCSP. Builds using
2980 the "no-ocsp" build time option are not affected.
2982 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2987 * SSL_peek() hang on empty record
2989 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2990 sends an empty record. This could be exploited by a malicious peer in a
2991 Denial Of Service attack.
2993 This issue was reported to OpenSSL by Alex Gaynor.
2998 * Excessive allocation of memory in tls_get_message_header() and
2999 dtls1_preprocess_fragment()
3001 A (D)TLS message includes 3 bytes for its length in the header for the
3002 message. This would allow for messages up to 16Mb in length. Messages of
3003 this length are excessive and OpenSSL includes a check to ensure that a
3004 peer is sending reasonably sized messages in order to avoid too much memory
3005 being consumed to service a connection. A flaw in the logic of version
3006 1.1.0 means that memory for the message is allocated too early, prior to
3007 the excessive message length check. Due to way memory is allocated in
3008 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3009 to service a connection. This could lead to a Denial of Service through
3010 memory exhaustion. However, the excessive message length check still takes
3011 place, and this would cause the connection to immediately fail. Assuming
3012 that the application calls SSL_free() on the failed connection in a timely
3013 manner then the 21Mb of allocated memory will then be immediately freed
3014 again. Therefore the excessive memory allocation will be transitory in
3015 nature. This then means that there is only a security impact if:
3017 1) The application does not call SSL_free() in a timely manner in the event
3018 that the connection fails
3020 2) The application is working in a constrained environment where there is
3021 very little free memory
3023 3) The attacker initiates multiple connection attempts such that there are
3024 multiple connections in a state where memory has been allocated for the
3025 connection; SSL_free() has not yet been called; and there is insufficient
3026 memory to service the multiple requests.
3028 Except in the instance of (1) above any Denial Of Service is likely to be
3029 transitory because as soon as the connection fails the memory is
3030 subsequently freed again in the SSL_free() call. However there is an
3031 increased risk during this period of application crashes due to the lack of
3032 memory - which would then mean a more serious Denial of Service.
3034 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3035 (CVE-2016-6307 and CVE-2016-6308)
3039 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3040 had to be removed. Primary reason is that vendor assembler can't
3041 assemble our modules with -KPIC flag. As result it, assembly
3042 support, was not even available as option. But its lack means
3043 lack of side-channel resistant code, which is incompatible with
3044 security by todays standards. Fortunately gcc is readily available
3045 prepackaged option, which we firmly point at...
3049 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3051 * Windows command-line tool supports UTF-8 opt-in option for arguments
3052 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3053 (to any value) allows Windows user to access PKCS#12 file generated
3054 with Windows CryptoAPI and protected with non-ASCII password, as well
3055 as files generated under UTF-8 locale on Linux also protected with
3060 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3061 have been disabled by default and removed from DEFAULT, just like RC4.
3062 See the RC4 item below to re-enable both.
3066 * The method for finding the storage location for the Windows RAND seed file
3067 has changed. First we check %RANDFILE%. If that is not set then we check
3068 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3069 all else fails we fall back to C:\.
3073 * The EVP_EncryptUpdate() function has had its return type changed from void
3074 to int. A return of 0 indicates and error while a return of 1 indicates
3079 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3080 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3081 off the constant time implementation for RSA, DSA and DH have been made
3082 no-ops and deprecated.
3086 * Windows RAND implementation was simplified to only get entropy by
3087 calling CryptGenRandom(). Various other RAND-related tickets
3090 *Joseph Wylie Yandle, Rich Salz*
3092 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3093 and `OPENSSL_LH_`, respectively. The old names are available
3094 with API compatibility. They new names are now completely documented.
3098 * Unify TYPE_up_ref(obj) methods signature.
3099 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3100 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3101 int (instead of void) like all others TYPE_up_ref() methods.
3102 So now these methods also check the return value of CRYPTO_atomic_add(),
3103 and the validity of object reference counter.
3105 *fdasilvayy@gmail.com*
3107 * With Windows Visual Studio builds, the .pdb files are installed
3108 alongside the installed libraries and executables. For a static
3109 library installation, ossl_static.pdb is the associate compiler
3110 generated .pdb file to be used when linking programs.
3114 * Remove openssl.spec. Packaging files belong with the packagers.
3118 * Automatic Darwin/OSX configuration has had a refresh, it will now
3119 recognise x86_64 architectures automatically. You can still decide
3120 to build for a different bitness with the environment variable
3121 KERNEL_BITS (can be 32 or 64), for example:
3123 KERNEL_BITS=32 ./config
3127 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3128 256 bit AES and HMAC with SHA256.
3132 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3136 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3140 * To enable users to have their own config files and build file templates,
3141 Configure looks in the directory indicated by the environment variable
3142 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3143 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3144 name and is used as is.
3148 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3149 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3150 X509_CERT_FILE_CTX was removed.
3154 * "shared" builds are now the default. To create only static libraries use
3155 the "no-shared" Configure option.
3159 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3160 All of these option have not worked for some while and are fundamental
3165 * Make various cleanup routines no-ops and mark them as deprecated. Most
3166 global cleanup functions are no longer required because they are handled
3167 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3168 Explicitly de-initing can cause problems (e.g. where a library that uses
3169 OpenSSL de-inits, but an application is still using it). The affected
3170 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3171 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3172 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3173 COMP_zlib_cleanup().
3177 * --strict-warnings no longer enables runtime debugging options
3178 such as REF_DEBUG. Instead, debug options are automatically
3179 enabled with '--debug' builds.
3181 *Andy Polyakov, Emilia Käsper*
3183 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3184 have been moved out of the public header files. New functions for managing
3185 these have been added.
3189 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3190 objects have been moved out of the public header files. New
3191 functions for managing these have been added.
3195 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3196 have been moved out of the public header files. New functions for managing
3197 these have been added.
3201 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3202 moved out of the public header files. New functions for managing these
3207 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3211 * Removed the mk1mf build scripts.
3215 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3216 it is always safe to #include a header now.
3220 * Removed the aged BC-32 config and all its supporting scripts
3224 * Removed support for Ultrix, Netware, and OS/2.
3228 * Add support for HKDF.
3230 *Alessandro Ghedini*
3232 * Add support for blake2b and blake2s
3236 * Added support for "pipelining". Ciphers that have the
3237 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3238 encryptions/decryptions simultaneously. There are currently no built-in
3239 ciphers with this property but the expectation is that engines will be able
3240 to offer it to significantly improve throughput. Support has been extended
3241 into libssl so that multiple records for a single connection can be
3242 processed in one go (for >=TLS 1.1).
3246 * Added the AFALG engine. This is an async capable engine which is able to
3247 offload work to the Linux kernel. In this initial version it only supports
3248 AES128-CBC. The kernel must be version 4.1.0 or greater.
3252 * OpenSSL now uses a new threading API. It is no longer necessary to
3253 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3254 are two supported threading models: pthreads and windows threads. It is
3255 also possible to configure OpenSSL at compile time for "no-threads". The
3256 old threading API should no longer be used. The functions have been
3257 replaced with "no-op" compatibility macros.
3259 *Alessandro Ghedini, Matt Caswell*
3261 * Modify behavior of ALPN to invoke callback after SNI/servername
3262 callback, such that updates to the SSL_CTX affect ALPN.
3266 * Add SSL_CIPHER queries for authentication and key-exchange.
3270 * Changes to the DEFAULT cipherlist:
3271 - Prefer (EC)DHE handshakes over plain RSA.
3272 - Prefer AEAD ciphers over legacy ciphers.
3273 - Prefer ECDSA over RSA when both certificates are available.
3274 - Prefer TLSv1.2 ciphers/PRF.
3275 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3280 * Change the ECC default curve list to be this, in order: x25519,
3281 secp256r1, secp521r1, secp384r1.
3285 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3286 disabled by default. They can be re-enabled using the
3287 enable-weak-ssl-ciphers option to Configure.
3291 * If the server has ALPN configured, but supports no protocols that the
3292 client advertises, send a fatal "no_application_protocol" alert.
3293 This behaviour is SHALL in RFC 7301, though it isn't universally
3294 implemented by other servers.
3298 * Add X25519 support.
3299 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3300 for public and private key encoding using the format documented in
3301 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3302 key generation and key derivation.
3304 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3309 * Deprecate SRP_VBASE_get_by_user.
3310 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3311 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3312 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3313 seed, even if the seed is configured.
3315 Users should use SRP_VBASE_get1_by_user instead. Note that in
3316 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3317 also that even though configuring the SRP seed attempts to hide
3318 invalid usernames by continuing the handshake with fake
3319 credentials, this behaviour is not constant time and no strong
3320 guarantees are made that the handshake is indistinguishable from
3321 that of a valid user.
3325 * Configuration change; it's now possible to build dynamic engines
3326 without having to build shared libraries and vice versa. This
3327 only applies to the engines in `engines/`, those in `crypto/engine/`
3328 will always be built into libcrypto (i.e. "static").
3330 Building dynamic engines is enabled by default; to disable, use
3331 the configuration option "disable-dynamic-engine".
3333 The only requirements for building dynamic engines are the
3334 presence of the DSO module and building with position independent
3335 code, so they will also automatically be disabled if configuring
3336 with "disable-dso" or "disable-pic".
3338 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3339 are also taken away from openssl/opensslconf.h, as they are
3344 * Configuration change; if there is a known flag to compile
3345 position independent code, it will always be applied on the
3346 libcrypto and libssl object files, and never on the application
3347 object files. This means other libraries that use routines from
3348 libcrypto / libssl can be made into shared libraries regardless
3349 of how OpenSSL was configured.
3351 If this isn't desirable, the configuration options "disable-pic"
3352 or "no-pic" can be used to disable the use of PIC. This will
3353 also disable building shared libraries and dynamic engines.
3357 * Removed JPAKE code. It was experimental and has no wide use.
3361 * The INSTALL_PREFIX Makefile variable has been renamed to
3362 DESTDIR. That makes for less confusion on what this variable
3363 is for. Also, the configuration option --install_prefix is
3368 * Heartbeat for TLS has been removed and is disabled by default
3369 for DTLS; configure with enable-heartbeats. Code that uses the
3370 old #define's might need to be updated.
3372 *Emilia Käsper, Rich Salz*
3374 * Rename REF_CHECK to REF_DEBUG.
3378 * New "unified" build system
3380 The "unified" build system is aimed to be a common system for all
3381 platforms we support. With it comes new support for VMS.
3383 This system builds supports building in a different directory tree
3384 than the source tree. It produces one Makefile (for unix family
3385 or lookalikes), or one descrip.mms (for VMS).
3387 The source of information to make the Makefile / descrip.mms is
3388 small files called 'build.info', holding the necessary
3389 information for each directory with source to compile, and a
3390 template in Configurations, like unix-Makefile.tmpl or
3393 With this change, the library names were also renamed on Windows
3394 and on VMS. They now have names that are closer to the standard
3395 on Unix, and include the major version number, and in certain
3396 cases, the architecture they are built for. See "Notes on shared
3397 libraries" in INSTALL.
3399 We rely heavily on the perl module Text::Template.
3403 * Added support for auto-initialisation and de-initialisation of the library.
3404 OpenSSL no longer requires explicit init or deinit routines to be called,
3405 except in certain circumstances. See the OPENSSL_init_crypto() and
3406 OPENSSL_init_ssl() man pages for further information.
3410 * The arguments to the DTLSv1_listen function have changed. Specifically the
3411 "peer" argument is now expected to be a BIO_ADDR object.
3413 * Rewrite of BIO networking library. The BIO library lacked consistent
3414 support of IPv6, and adding it required some more extensive
3415 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3416 which hold all types of addresses and chains of address information.
3417 It also introduces a new API, with functions like BIO_socket,
3418 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3419 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3420 have been adapted accordingly.
3424 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3429 * CRIME protection: disable compression by default, even if OpenSSL is
3430 compiled with zlib enabled. Applications can still enable compression
3431 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3432 using the SSL_CONF library to configure compression.
3436 * The signature of the session callback configured with
3437 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3438 was explicitly marked as `const unsigned char*` instead of
3443 * Always DPURIFY. Remove the use of uninitialized memory in the
3444 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3448 * Removed many obsolete configuration items, including
3449 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3450 MD2_CHAR, MD2_INT, MD2_LONG
3452 IDEA_SHORT, IDEA_LONG
3453 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3455 *Rich Salz, with advice from Andy Polyakov*
3457 * Many BN internals have been moved to an internal header file.
3459 *Rich Salz with help from Andy Polyakov*
3461 * Configuration and writing out the results from it has changed.
3462 Files such as Makefile include/openssl/opensslconf.h and are now
3463 produced through general templates, such as Makefile.in and
3464 crypto/opensslconf.h.in and some help from the perl module
3467 Also, the center of configuration information is no longer
3468 Makefile. Instead, Configure produces a perl module in
3469 configdata.pm which holds most of the config data (in the hash
3470 table %config), the target data that comes from the target
3471 configuration in one of the `Configurations/*.conf` files (in
3476 * To clarify their intended purposes, the Configure options
3477 --prefix and --openssldir change their semantics, and become more
3478 straightforward and less interdependent.
3480 --prefix shall be used exclusively to give the location INSTALLTOP
3481 where programs, scripts, libraries, include files and manuals are
3482 going to be installed. The default is now /usr/local.
3484 --openssldir shall be used exclusively to give the default
3485 location OPENSSLDIR where certificates, private keys, CRLs are
3486 managed. This is also where the default openssl.cnf gets
3488 If the directory given with this option is a relative path, the
3489 values of both the --prefix value and the --openssldir value will
3490 be combined to become OPENSSLDIR.
3491 The default for --openssldir is INSTALLTOP/ssl.
3493 Anyone who uses --openssldir to specify where OpenSSL is to be
3494 installed MUST change to use --prefix instead.
3498 * The GOST engine was out of date and therefore it has been removed. An up
3499 to date GOST engine is now being maintained in an external repository.
3500 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3501 support for GOST ciphersuites (these are only activated if a GOST engine
3506 * EGD is no longer supported by default; use enable-egd when
3509 *Ben Kaduk and Rich Salz*
3511 * The distribution now has Makefile.in files, which are used to
3512 create Makefile's when Configure is run. *Configure must be run
3513 before trying to build now.*
3517 * The return value for SSL_CIPHER_description() for error conditions
3522 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3524 Obtaining and performing DNSSEC validation of TLSA records is
3525 the application's responsibility. The application provides
3526 the TLSA records of its choice to OpenSSL, and these are then
3527 used to authenticate the peer.
3529 The TLSA records need not even come from DNS. They can, for
3530 example, be used to implement local end-entity certificate or
3531 trust-anchor "pinning", where the "pin" data takes the form
3532 of TLSA records, which can augment or replace verification
3533 based on the usual WebPKI public certification authorities.
3537 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3538 continues to support deprecated interfaces in default builds.
3539 However, applications are strongly advised to compile their
3540 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3541 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3542 or the 1.1.0 releases.
3544 In environments in which all applications have been ported to
3545 not use any deprecated interfaces OpenSSL's Configure script
3546 should be used with the --api=1.1.0 option to entirely remove
3547 support for the deprecated features from the library and
3548 unconditionally disable them in the installed headers.
3549 Essentially the same effect can be achieved with the "no-deprecated"
3550 argument to Configure, except that this will always restrict
3551 the build to just the latest API, rather than a fixed API
3554 As applications are ported to future revisions of the API,
3555 they should update their compile-time OPENSSL_API_COMPAT define
3556 accordingly, but in most cases should be able to continue to
3557 compile with later releases.
3559 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3560 0x10000000L and 0x00908000L, respectively. However those
3561 versions did not support the OPENSSL_API_COMPAT feature, and
3562 so applications are not typically tested for explicit support
3563 of just the undeprecated features of either release.
3567 * Add support for setting the minimum and maximum supported protocol.
3568 It can bet set via the SSL_set_min_proto_version() and
3569 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3570 MaxProtocol. It's recommended to use the new APIs to disable
3571 protocols instead of disabling individual protocols using
3572 SSL_set_options() or SSL_CONF's Protocol. This change also
3573 removes support for disabling TLS 1.2 in the OpenSSL TLS
3574 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3578 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3582 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3583 and integrates ECDSA and ECDH functionality into EC. Implementations can
3584 now redirect key generation and no longer need to convert to or from
3587 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3588 include the ec.h header file instead.
3592 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3593 ciphers who are no longer supported and drops support the ephemeral RSA key
3594 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3598 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3599 opaque. For HMAC_CTX, the following constructors and destructors
3602 HMAC_CTX *HMAC_CTX_new(void);
3603 void HMAC_CTX_free(HMAC_CTX *ctx);
3605 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3606 destroy such methods has been added. See EVP_MD_meth_new(3) and
3607 EVP_CIPHER_meth_new(3) for documentation.
3610 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3611 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3612 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3613 an already created structure.
3614 2) For consistency with the majority of our object creators and
3615 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3616 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3617 for deprecated builds.
3621 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3622 cryptographic operations to be performed asynchronously as long as an
3623 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3624 further details. Libssl has also had this capability integrated with the
3625 introduction of the new mode SSL_MODE_ASYNC and associated error
3626 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3627 pages. This work was developed in partnership with Intel Corp.
3631 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3632 always enabled now. If you want to disable the support you should
3633 exclude it using the list of supported ciphers. This also means that the
3634 "-no_ecdhe" option has been removed from s_server.
3638 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3639 SSL_{CTX_}set1_curves() which can set a list.
3643 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3644 curve you want to support using SSL_{CTX_}set1_curves().
3648 * State machine rewrite. The state machine code has been significantly
3649 refactored in order to remove much duplication of code and solve issues
3650 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3651 further details). This change does have some associated API changes.
3652 Notably the SSL_state() function has been removed and replaced by
3653 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3654 SSL_set_state() has been removed altogether. The previous handshake states
3655 defined in ssl.h and ssl3.h have also been removed.
3659 * All instances of the string "ssleay" in the public API were replaced
3660 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3661 Some error codes related to internal RSA_eay API's were renamed.
3665 * The demo files in crypto/threads were moved to demo/threads.
3669 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3672 *Matt Caswell, Rich Salz*
3674 * New ASN.1 embed macro.
3676 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3677 structure is not allocated: it is part of the parent. That is instead of
3685 This reduces memory fragmentation and make it impossible to accidentally
3686 set a mandatory field to NULL.
3688 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3689 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3690 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3695 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3699 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3700 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3701 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3702 DES and RC4 ciphersuites.
3706 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3707 This changes the decoding behaviour for some invalid messages,
3708 though the change is mostly in the more lenient direction, and
3709 legacy behaviour is preserved as much as possible.
3713 * Fix no-stdio build.
3714 *David Woodhouse <David.Woodhouse@intel.com> and also*
3715 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3717 * New testing framework
3718 The testing framework has been largely rewritten and is now using
3719 perl and the perl modules Test::Harness and an extended variant of
3720 Test::More called OpenSSL::Test to do its work. All test scripts in
3721 test/ have been rewritten into test recipes, and all direct calls to
3722 executables in test/Makefile have become individual recipes using the
3723 simplified testing OpenSSL::Test::Simple.
3725 For documentation on our testing modules, do:
3727 perldoc test/testlib/OpenSSL/Test/Simple.pm
3728 perldoc test/testlib/OpenSSL/Test.pm
3732 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3733 are used; the latter aborts on memory leaks (usually checked on exit).
3734 Some undocumented "set malloc, etc., hooks" functions were removed
3735 and others were changed. All are now documented.
3739 * In DSA_generate_parameters_ex, if the provided seed is too short,
3742 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3744 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3745 from RFC4279, RFC4785, RFC5487, RFC5489.
3747 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3748 original RSA_PSK patch.
3752 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3753 era flag was never set throughout the codebase (only read). Also removed
3754 SSL3_FLAGS_POP_BUFFER which was only used if
3755 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3759 * Changed the default name options in the "ca", "crl", "req" and "x509"
3760 to be "oneline" instead of "compat".
3764 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3765 not aware of clients that still exhibit this bug, and the workaround
3766 hasn't been working properly for a while.
3770 * The return type of BIO_number_read() and BIO_number_written() as well as
3771 the corresponding num_read and num_write members in the BIO structure has
3772 changed from unsigned long to uint64_t. On platforms where an unsigned
3773 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3778 * Given the pervasive nature of TLS extensions it is inadvisable to run
3779 OpenSSL without support for them. It also means that maintaining
3780 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3781 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3785 * Removed support for the two export grade static DH ciphersuites
3786 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3787 were newly added (along with a number of other static DH ciphersuites) to
3788 1.0.2. However the two export ones have *never* worked since they were
3789 introduced. It seems strange in any case to be adding new export
3790 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3794 * Version negotiation has been rewritten. In particular SSLv23_method(),
3795 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3796 and turned into macros which simply call the new preferred function names
3797 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3798 should use the new names instead. Also as part of this change the ssl23.h
3799 header file has been removed.
3803 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3804 code and the associated standard is no longer considered fit-for-purpose.
3808 * RT2547 was closed. When generating a private key, try to make the
3809 output file readable only by the owner. This behavior change might
3810 be noticeable when interacting with other software.
3812 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3817 * Added HTTP GET support to the ocsp command.
3821 * Changed default digest for the dgst and enc commands from MD5 to
3826 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3830 * Added support for TLS extended master secret from
3831 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3832 initial patch which was a great help during development.
3836 * All libssl internal structures have been removed from the public header
3837 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3838 now redundant). Users should not attempt to access internal structures
3839 directly. Instead they should use the provided API functions.
3843 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3844 Access to deprecated functions can be re-enabled by running config with
3845 "enable-deprecated". In addition applications wishing to use deprecated
3846 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3847 will, by default, disable some transitive includes that previously existed
3848 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3852 * Added support for OCB mode. OpenSSL has been granted a patent license
3853 compatible with the OpenSSL license for use of OCB. Details are available
3854 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3855 for OCB can be removed by calling config with no-ocb.
3859 * SSLv2 support has been removed. It still supports receiving a SSLv2
3860 compatible client hello.
3864 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3865 done while fixing the error code for the key-too-small case.
3867 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3869 * CA.sh has been removed; use CA.pl instead.
3873 * Removed old DES API.
3877 * Remove various unsupported platforms:
3883 Sinix/ReliantUNIX RM400
3888 16-bit platforms such as WIN16
3892 * Clean up OPENSSL_NO_xxx #define's
3893 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3894 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3895 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3896 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3897 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3898 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3899 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3900 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3901 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3902 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3906 * Cleaned up dead code
3907 Remove all but one '#ifdef undef' which is to be looked at.
3911 * Clean up calling of xxx_free routines.
3912 Just like free(), fix most of the xxx_free routines to accept
3913 NULL. Remove the non-null checks from callers. Save much code.
3917 * Add secure heap for storage of private keys (when possible).
3918 Add BIO_s_secmem(), CBIGNUM, etc.
3919 Contributed by Akamai Technologies under our Corporate CLA.
3923 * Experimental support for a new, fast, unbiased prime candidate generator,
3924 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3926 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3928 * New output format NSS in the sess_id command line tool. This allows
3929 exporting the session id and the master key in NSS keylog format.
3931 *Martin Kaiser <martin@kaiser.cx>*
3933 * Harmonize version and its documentation. -f flag is used to display
3936 *mancha <mancha1@zoho.com>*
3938 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3939 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3941 *mancha <mancha1@zoho.com>*
3943 * Fix some double frees. These are not thought to be exploitable.
3945 *mancha <mancha1@zoho.com>*
3947 * A missing bounds check in the handling of the TLS heartbeat extension
3948 can be used to reveal up to 64k of memory to a connected client or
3951 Thanks for Neel Mehta of Google Security for discovering this bug and to
3952 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3953 preparing the fix ([CVE-2014-0160])
3955 *Adam Langley, Bodo Moeller*
3957 * Fix for the attack described in the paper "Recovering OpenSSL
3958 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3959 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3960 <http://eprint.iacr.org/2014/140>
3962 Thanks to Yuval Yarom and Naomi Benger for discovering this
3963 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3965 *Yuval Yarom and Naomi Benger*
3967 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3968 this fixes a limitation in previous versions of OpenSSL.
3972 * Experimental encrypt-then-mac support.
3974 Experimental support for encrypt then mac from
3975 draft-gutmann-tls-encrypt-then-mac-02.txt
3977 To enable it set the appropriate extension number (0x42 for the test
3978 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3980 For non-compliant peers (i.e. just about everything) this should have no
3983 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3987 * Add EVP support for key wrapping algorithms, to avoid problems with
3988 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3989 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3990 algorithms and include tests cases.
3994 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3999 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4000 MGF1 digest and OAEP label.
4004 * Make openssl verify return errors.
4006 *Chris Palmer <palmer@google.com> and Ben Laurie*
4008 * New function ASN1_TIME_diff to calculate the difference between two
4009 ASN1_TIME structures or one structure and the current time.
4013 * Update fips_test_suite to support multiple command line options. New
4014 test to induce all self test errors in sequence and check expected
4019 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4020 sign or verify all in one operation.
4024 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4025 test programs and fips_test_suite. Includes functionality to parse
4026 the minimal script output of fipsalgest.pl directly.
4030 * Add authorisation parameter to FIPS_module_mode_set().
4034 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4038 * Use separate DRBG fields for internal and external flags. New function
4039 FIPS_drbg_health_check() to perform on demand health checking. Add
4040 generation tests to fips_test_suite with reduced health check interval to
4041 demonstrate periodic health checking. Add "nodh" option to
4042 fips_test_suite to skip very slow DH test.
4046 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4051 * More extensive health check for DRBG checking many more failure modes.
4052 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4053 combination: call this in fips_test_suite.
4057 * Add support for canonical generation of DSA parameter 'g'. See
4060 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4061 POST to handle HMAC cases.
4065 * Add functions FIPS_module_version() and FIPS_module_version_text()
4066 to return numerical and string versions of the FIPS module number.
4070 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4071 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4072 outside the validated module in the FIPS capable OpenSSL.
4076 * Minor change to DRBG entropy callback semantics. In some cases
4077 there is no multiple of the block length between min_len and
4078 max_len. Allow the callback to return more than max_len bytes
4079 of entropy but discard any extra: it is the callback's responsibility
4080 to ensure that the extra data discarded does not impact the
4081 requested amount of entropy.
4085 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4086 information in FIPS186-3, SP800-57 and SP800-131A.
4090 * CCM support via EVP. Interface is very similar to GCM case except we
4091 must supply all data in one chunk (i.e. no update, final) and the
4092 message length must be supplied if AAD is used. Add algorithm test
4097 * Initial version of POST overhaul. Add POST callback to allow the status
4098 of POST to be monitored and/or failures induced. Modify fips_test_suite
4099 to use callback. Always run all selftests even if one fails.
4103 * XTS support including algorithm test driver in the fips_gcmtest program.
4104 Note: this does increase the maximum key length from 32 to 64 bytes but
4105 there should be no binary compatibility issues as existing applications
4106 will never use XTS mode.
4110 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4111 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4112 performs algorithm blocking for unapproved PRNG types. Also do not
4113 set PRNG type in FIPS_mode_set(): leave this to the application.
4114 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4115 the standard OpenSSL PRNG: set additional data to a date time vector.
4119 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4120 This shouldn't present any incompatibility problems because applications
4121 shouldn't be using these directly and any that are will need to rethink
4122 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4126 * Extensive self tests and health checking required by SP800-90 DRBG.
4127 Remove strength parameter from FIPS_drbg_instantiate and always
4128 instantiate at maximum supported strength.
4132 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4136 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4140 * New function DH_compute_key_padded() to compute a DH key and pad with
4141 leading zeroes if needed: this complies with SP800-56A et al.
4145 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4146 anything, incomplete, subject to change and largely untested at present.
4150 * Modify fipscanisteronly build option to only build the necessary object
4151 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4155 * Add experimental option FIPSSYMS to give all symbols in
4156 fipscanister.o and FIPS or fips prefix. This will avoid
4157 conflicts with future versions of OpenSSL. Add perl script
4158 util/fipsas.pl to preprocess assembly language source files
4159 and rename any affected symbols.
4163 * Add selftest checks and algorithm block of non-fips algorithms in
4164 FIPS mode. Remove DES2 from selftests.
4168 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4169 return internal method without any ENGINE dependencies. Add new
4170 tiny fips sign and verify functions.
4174 * New build option no-ec2m to disable characteristic 2 code.
4178 * New build option "fipscanisteronly". This only builds fipscanister.o
4179 and (currently) associated fips utilities. Uses the file Makefile.fips
4180 instead of Makefile.org as the prototype.
4184 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4185 Update fips_gcmtest to use IV generator.
4189 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4190 setting output buffer to NULL. The `*Final` function must be
4191 called although it will not retrieve any additional data. The tag
4192 can be set or retrieved with a ctrl. The IV length is by default 12
4193 bytes (96 bits) but can be set to an alternative value. If the IV
4194 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4199 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4200 underlying do_cipher function handles all cipher semantics itself
4201 including padding and finalisation. This is useful if (for example)
4202 an ENGINE cipher handles block padding itself. The behaviour of
4203 do_cipher is subtly changed if this flag is set: the return value
4204 is the number of characters written to the output buffer (zero is
4205 no longer an error code) or a negative error code. Also if the
4206 input buffer is NULL and length 0 finalisation should be performed.
4210 * If a candidate issuer certificate is already part of the constructed
4211 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4215 * Improve forward-security support: add functions
4217 void SSL_CTX_set_not_resumable_session_callback(
4218 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4219 void SSL_set_not_resumable_session_callback(
4220 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4222 for use by SSL/TLS servers; the callback function will be called whenever a
4223 new session is created, and gets to decide whether the session may be
4224 cached to make it resumable (return 0) or not (return 1). (As by the
4225 SSL/TLS protocol specifications, the session_id sent by the server will be
4226 empty to indicate that the session is not resumable; also, the server will
4227 not generate RFC 4507 (RFC 5077) session tickets.)
4229 A simple reasonable callback implementation is to return is_forward_secure.
4230 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4231 by the SSL/TLS server library, indicating whether it can provide forward
4234 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4236 * New -verify_name option in command line utilities to set verification
4241 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4242 Add CMAC pkey methods.
4246 * Experimental renegotiation in s_server -www mode. If the client
4247 browses /reneg connection is renegotiated. If /renegcert it is
4248 renegotiated requesting a certificate.
4252 * Add an "external" session cache for debugging purposes to s_server. This
4253 should help trace issues which normally are only apparent in deployed
4254 multi-process servers.
4258 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4259 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4260 BIO_set_cipher() and some obscure PEM functions were changed so they
4261 can now return an error. The RAND changes required a change to the
4262 RAND_METHOD structure.
4266 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4267 a gcc attribute to warn if the result of a function is ignored. This
4268 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4269 whose return value is often ignored.
4273 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4274 These allow SCTs (signed certificate timestamps) to be requested and
4275 validated when establishing a connection.
4277 *Rob Percival <robpercival@google.com>*
4282 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4284 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4285 used even when parsing explicit parameters, when loading a encoded key
4286 or calling `EC_GROUP_new_from_ecpkparameters()`/
4287 `EC_GROUP_new_from_ecparameters()`.
4288 This prevents bypass of security hardening and performance gains,
4289 especially for curves with specialized EC_METHODs.
4290 By default, if a key encoded with explicit parameters is loaded and later
4291 encoded, the output is still encoded with explicit parameters, even if
4292 internally a "named" EC_GROUP is used for computation.
4296 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4297 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4298 NULL. After this change, only the cofactor parameter can be NULL. It also
4299 does some minimal sanity checks on the passed order.
4304 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4305 An attack is simple, if the first CMS_recipientInfo is valid but the
4306 second CMS_recipientInfo is chosen ciphertext. If the second
4307 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4308 encryption key will be replaced by garbage, and the message cannot be
4309 decoded, but if the RSA decryption fails, the correct encryption key is
4310 used and the recipient will not notice the attack.
4311 As a work around for this potential attack the length of the decrypted
4312 key must be equal to the cipher default key length, in case the
4313 certifiate is not given and all recipientInfo are tried out.
4314 The old behaviour can be re-enabled in the CMS code by setting the
4315 CMS_DEBUG_DECRYPT flag.
4320 * Document issue with installation paths in diverse Windows builds
4322 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4323 binaries and run-time config file.
4328 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4330 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4331 This changes the size when using the `genpkey` command when no size is given.
4332 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4333 generation commands to use 2048 bits by default.
4337 * Add FIPS support for Android Arm 64-bit
4339 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4340 Module in Version 2.0.10. For some reason, the corresponding target
4341 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4342 built with FIPS support on Android Arm 64-bit. This omission has been
4345 *Matthias St. Pierre*
4347 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4349 * 0-byte record padding oracle
4351 If an application encounters a fatal protocol error and then calls
4352 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4353 then OpenSSL can respond differently to the calling application if a 0 byte
4354 record is received with invalid padding compared to if a 0 byte record is
4355 received with an invalid MAC. If the application then behaves differently
4356 based on that in a way that is detectable to the remote peer, then this
4357 amounts to a padding oracle that could be used to decrypt data.
4359 In order for this to be exploitable "non-stitched" ciphersuites must be in
4360 use. Stitched ciphersuites are optimised implementations of certain
4361 commonly used ciphersuites. Also the application must call SSL_shutdown()
4362 twice even if a protocol error has occurred (applications should not do
4363 this but some do anyway).
4365 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4366 Aviram, with additional investigation by Steven Collison and Andrew
4367 Hourselt. It was reported to OpenSSL on 10th December 2018.
4372 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4376 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4378 * Microarchitecture timing vulnerability in ECC scalar multiplication
4380 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4381 shown to be vulnerable to a microarchitecture timing side channel attack.
4382 An attacker with sufficient access to mount local timing attacks during
4383 ECDSA signature generation could recover the private key.
4385 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4386 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4392 * Timing vulnerability in DSA signature generation
4394 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4395 timing side channel attack. An attacker could use variations in the signing
4396 algorithm to recover the private key.
4398 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4403 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4404 Module, accidentally introduced while backporting security fixes from the
4405 development branch and hindering the use of ECC in FIPS mode.
4409 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4411 * Client DoS due to large DH parameter
4413 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4414 malicious server can send a very large prime value to the client. This will
4415 cause the client to spend an unreasonably long period of time generating a
4416 key for this prime resulting in a hang until the client has finished. This
4417 could be exploited in a Denial Of Service attack.
4419 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4424 * Cache timing vulnerability in RSA Key Generation
4426 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4427 a cache timing side channel attack. An attacker with sufficient access to
4428 mount cache timing attacks during the RSA key generation process could
4429 recover the private key.
4431 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4432 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4437 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4438 parameter is no longer accepted, as it leads to a corrupt table. NULL
4439 pem_str is reserved for alias entries only.
4443 * Revert blinding in ECDSA sign and instead make problematic addition
4444 length-invariant. Switch even to fixed-length Montgomery multiplication.
4448 * Change generating and checking of primes so that the error rate of not
4449 being prime depends on the intended use based on the size of the input.
4450 For larger primes this will result in more rounds of Miller-Rabin.
4451 The maximal error rate for primes with more than 1080 bits is lowered
4454 *Kurt Roeckx, Annie Yousar*
4456 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4460 * Add blinding to ECDSA and DSA signatures to protect against side channel
4461 attacks discovered by Keegan Ryan (NCC Group).
4465 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4466 now allow empty (zero character) pass phrases.
4470 * Certificate time validation (X509_cmp_time) enforces stricter
4471 compliance with RFC 5280. Fractional seconds and timezone offsets
4472 are no longer allowed.
4476 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4478 * Constructed ASN.1 types with a recursive definition could exceed the stack
4480 Constructed ASN.1 types with a recursive definition (such as can be found
4481 in PKCS7) could eventually exceed the stack given malicious input with
4482 excessive recursion. This could result in a Denial Of Service attack. There
4483 are no such structures used within SSL/TLS that come from untrusted sources
4484 so this is considered safe.
4486 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4492 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4494 * Read/write after SSL object in error state
4496 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4497 mechanism. The intent was that if a fatal error occurred during a handshake
4498 then OpenSSL would move into the error state and would immediately fail if
4499 you attempted to continue the handshake. This works as designed for the
4500 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4501 SSL_connect()), however due to a bug it does not work correctly if
4502 SSL_read() or SSL_write() is called directly. In that scenario, if the
4503 handshake fails then a fatal error will be returned in the initial function
4504 call. If SSL_read()/SSL_write() is subsequently called by the application
4505 for the same SSL object then it will succeed and the data is passed without
4506 being decrypted/encrypted directly from the SSL/TLS record layer.
4508 In order to exploit this issue an application bug would have to be present
4509 that resulted in a call to SSL_read()/SSL_write() being issued after having
4510 already received a fatal error.
4512 This issue was reported to OpenSSL by David Benjamin (Google).
4517 * rsaz_1024_mul_avx2 overflow bug on x86_64
4519 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4520 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4521 Analysis suggests that attacks against RSA and DSA as a result of this
4522 defect would be very difficult to perform and are not believed likely.
4523 Attacks against DH1024 are considered just feasible, because most of the
4524 work necessary to deduce information about a private key may be performed
4525 offline. The amount of resources required for such an attack would be
4526 significant. However, for an attack on TLS to be meaningful, the server
4527 would have to share the DH1024 private key among multiple clients, which is
4528 no longer an option since CVE-2016-0701.
4530 This only affects processors that support the AVX2 but not ADX extensions
4531 like Intel Haswell (4th generation).
4533 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4534 was originally found via the OSS-Fuzz project.
4539 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4541 * bn_sqrx8x_internal carry bug on x86_64
4543 There is a carry propagating bug in the x86_64 Montgomery squaring
4544 procedure. No EC algorithms are affected. Analysis suggests that attacks
4545 against RSA and DSA as a result of this defect would be very difficult to
4546 perform and are not believed likely. Attacks against DH are considered just
4547 feasible (although very difficult) because most of the work necessary to
4548 deduce information about a private key may be performed offline. The amount
4549 of resources required for such an attack would be very significant and
4550 likely only accessible to a limited number of attackers. An attacker would
4551 additionally need online access to an unpatched system using the target
4552 private key in a scenario with persistent DH parameters and a private
4553 key that is shared between multiple clients.
4555 This only affects processors that support the BMI1, BMI2 and ADX extensions
4556 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4558 This issue was reported to OpenSSL by the OSS-Fuzz project.
4563 * Malformed X.509 IPAddressFamily could cause OOB read
4565 If an X.509 certificate has a malformed IPAddressFamily extension,
4566 OpenSSL could do a one-byte buffer overread. The most likely result
4567 would be an erroneous display of the certificate in text format.
4569 This issue was reported to OpenSSL by the OSS-Fuzz project.
4574 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4576 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4577 platform rather than 'mingw'.
4581 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4583 * Truncated packet could crash via OOB read
4585 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4586 cipher is being used, then a truncated packet can cause that host to
4587 perform an out-of-bounds read, usually resulting in a crash.
4589 This issue was reported to OpenSSL by Robert Święcki of Google.
4594 * BN_mod_exp may produce incorrect results on x86_64
4596 There is a carry propagating bug in the x86_64 Montgomery squaring
4597 procedure. No EC algorithms are affected. Analysis suggests that attacks
4598 against RSA and DSA as a result of this defect would be very difficult to
4599 perform and are not believed likely. Attacks against DH are considered just
4600 feasible (although very difficult) because most of the work necessary to
4601 deduce information about a private key may be performed offline. The amount
4602 of resources required for such an attack would be very significant and
4603 likely only accessible to a limited number of attackers. An attacker would
4604 additionally need online access to an unpatched system using the target
4605 private key in a scenario with persistent DH parameters and a private
4606 key that is shared between multiple clients. For example this can occur by
4607 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4608 similar to CVE-2015-3193 but must be treated as a separate problem.
4610 This issue was reported to OpenSSL by the OSS-Fuzz project.
4615 * Montgomery multiplication may produce incorrect results
4617 There is a carry propagating bug in the Broadwell-specific Montgomery
4618 multiplication procedure that handles input lengths divisible by, but
4619 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4620 and DH private keys are impossible. This is because the subroutine in
4621 question is not used in operations with the private key itself and an input
4622 of the attacker's direct choice. Otherwise the bug can manifest itself as
4623 transient authentication and key negotiation failures or reproducible
4624 erroneous outcome of public-key operations with specially crafted input.
4625 Among EC algorithms only Brainpool P-512 curves are affected and one
4626 presumably can attack ECDH key negotiation. Impact was not analyzed in
4627 detail, because pre-requisites for attack are considered unlikely. Namely
4628 multiple clients have to choose the curve in question and the server has to
4629 share the private key among them, neither of which is default behaviour.
4630 Even then only clients that chose the curve will be affected.
4632 This issue was publicly reported as transient failures and was not
4633 initially recognized as a security issue. Thanks to Richard Morgan for
4634 providing reproducible case.
4639 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4640 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4641 prevent issues where no progress is being made and the peer continually
4642 sends unrecognised record types, using up resources processing them.
4646 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4648 * Missing CRL sanity check
4650 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4651 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4652 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4654 This issue only affects the OpenSSL 1.0.2i
4659 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4661 * OCSP Status Request extension unbounded memory growth
4663 A malicious client can send an excessively large OCSP Status Request
4664 extension. If that client continually requests renegotiation, sending a
4665 large OCSP Status Request extension each time, then there will be unbounded
4666 memory growth on the server. This will eventually lead to a Denial Of
4667 Service attack through memory exhaustion. Servers with a default
4668 configuration are vulnerable even if they do not support OCSP. Builds using
4669 the "no-ocsp" build time option are not affected.
4671 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4676 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4679 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4685 * OOB write in MDC2_Update()
4687 An overflow can occur in MDC2_Update() either if called directly or
4688 through the EVP_DigestUpdate() function using MDC2. If an attacker
4689 is able to supply very large amounts of input data after a previous
4690 call to EVP_EncryptUpdate() with a partial block then a length check
4691 can overflow resulting in a heap corruption.
4693 The amount of data needed is comparable to SIZE_MAX which is impractical
4696 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4701 * Malformed SHA512 ticket DoS
4703 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4704 DoS attack where a malformed ticket will result in an OOB read which will
4707 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4708 a custom server callback and ticket lookup mechanism.
4710 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4715 * OOB write in BN_bn2dec()
4717 The function BN_bn2dec() does not check the return value of BN_div_word().
4718 This can cause an OOB write if an application uses this function with an
4719 overly large BIGNUM. This could be a problem if an overly large certificate
4720 or CRL is printed out from an untrusted source. TLS is not affected because
4721 record limits will reject an oversized certificate before it is parsed.
4723 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4728 * OOB read in TS_OBJ_print_bio()
4730 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4731 the total length the OID text representation would use and not the amount
4732 of data written. This will result in OOB reads when large OIDs are
4735 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4740 * Pointer arithmetic undefined behaviour
4742 Avoid some undefined pointer arithmetic
4744 A common idiom in the codebase is to check limits in the following manner:
4747 Where "p" points to some malloc'd data of SIZE bytes and
4750 "len" here could be from some externally supplied data (e.g. from a TLS
4753 The rules of C pointer arithmetic are such that "p + len" is only well
4754 defined where len <= SIZE. Therefore the above idiom is actually
4755 undefined behaviour.
4757 For example this could cause problems if some malloc implementation
4758 provides an address for "p" such that "p + len" actually overflows for
4759 values of len that are too big and therefore p + len < limit.
4761 This issue was reported to OpenSSL by Guido Vranken
4766 * Constant time flag not preserved in DSA signing
4768 Operations in the DSA signing algorithm should run in constant time in
4769 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4770 implementation means that a non-constant time codepath is followed for
4771 certain operations. This has been demonstrated through a cache-timing
4772 attack to be sufficient for an attacker to recover the private DSA key.
4774 This issue was reported by César Pereida (Aalto University), Billy Brumley
4775 (Tampere University of Technology), and Yuval Yarom (The University of
4776 Adelaide and NICTA).
4781 * DTLS buffered message DoS
4783 In a DTLS connection where handshake messages are delivered out-of-order
4784 those messages that OpenSSL is not yet ready to process will be buffered
4785 for later use. Under certain circumstances, a flaw in the logic means that
4786 those messages do not get removed from the buffer even though the handshake
4787 has been completed. An attacker could force up to approx. 15 messages to
4788 remain in the buffer when they are no longer required. These messages will
4789 be cleared when the DTLS connection is closed. The default maximum size for
4790 a message is 100k. Therefore the attacker could force an additional 1500k
4791 to be consumed per connection. By opening many simulataneous connections an
4792 attacker could cause a DoS attack through memory exhaustion.
4794 This issue was reported to OpenSSL by Quan Luo.
4799 * DTLS replay protection DoS
4801 A flaw in the DTLS replay attack protection mechanism means that records
4802 that arrive for future epochs update the replay protection "window" before
4803 the MAC for the record has been validated. This could be exploited by an
4804 attacker by sending a record for the next epoch (which does not have to
4805 decrypt or have a valid MAC), with a very large sequence number. This means
4806 that all subsequent legitimate packets are dropped causing a denial of
4807 service for a specific DTLS connection.
4809 This issue was reported to OpenSSL by the OCAP audit team.
4814 * Certificate message OOB reads
4816 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4817 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4818 theoretical DoS risk but this has not been observed in practice on common
4821 The messages affected are client certificate, client certificate request
4822 and server certificate. As a result the attack can only be performed
4823 against a client or a server which enables client authentication.
4825 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4830 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4832 * Prevent padding oracle in AES-NI CBC MAC check
4834 A MITM attacker can use a padding oracle attack to decrypt traffic
4835 when the connection uses an AES CBC cipher and the server support
4838 This issue was introduced as part of the fix for Lucky 13 padding
4839 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4840 constant time by making sure that always the same bytes are read and
4841 compared against either the MAC or padding bytes. But it no longer
4842 checked that there was enough data to have both the MAC and padding
4845 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4850 * Fix EVP_EncodeUpdate overflow
4852 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4853 Base64 encoding of binary data. If an attacker is able to supply very large
4854 amounts of input data then a length check can overflow resulting in a heap
4857 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4858 the `PEM_write_bio*` family of functions. These are mainly used within the
4859 OpenSSL command line applications, so any application which processes data
4860 from an untrusted source and outputs it as a PEM file should be considered
4861 vulnerable to this issue. User applications that call these APIs directly
4862 with large amounts of untrusted data may also be vulnerable.
4864 This issue was reported by Guido Vranken.
4869 * Fix EVP_EncryptUpdate overflow
4871 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4872 is able to supply very large amounts of input data after a previous call to
4873 EVP_EncryptUpdate() with a partial block then a length check can overflow
4874 resulting in a heap corruption. Following an analysis of all OpenSSL
4875 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4876 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4877 the first called function after an EVP_EncryptInit(), and therefore that
4878 specific call must be safe. The second form is where the length passed to
4879 EVP_EncryptUpdate() can be seen from the code to be some small value and
4880 therefore there is no possibility of an overflow. Since all instances are
4881 one of these two forms, it is believed that there can be no overflows in
4882 internal code due to this problem. It should be noted that
4883 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4884 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4885 of these calls have also been analysed too and it is believed there are no
4886 instances in internal usage where an overflow could occur.
4888 This issue was reported by Guido Vranken.
4893 * Prevent ASN.1 BIO excessive memory allocation
4895 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4896 a short invalid encoding can cause allocation of large amounts of memory
4897 potentially consuming excessive resources or exhausting memory.
4899 Any application parsing untrusted data through d2i BIO functions is
4900 affected. The memory based functions such as d2i_X509() are *not* affected.
4901 Since the memory based functions are used by the TLS library, TLS
4902 applications are not affected.
4904 This issue was reported by Brian Carpenter.
4911 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4912 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4913 in arbitrary stack data being returned in the buffer.
4915 This issue was reported by Guido Vranken.
4920 * Modify behavior of ALPN to invoke callback after SNI/servername
4921 callback, such that updates to the SSL_CTX affect ALPN.
4925 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4930 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4931 methods are enabled and ssl2 is disabled the methods return NULL.
4935 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4937 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4938 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4939 provide any "EXPORT" or "LOW" strength ciphers.
4943 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4944 is by default disabled at build-time. Builds that are not configured with
4945 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4946 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4947 will need to explicitly call either of:
4949 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4951 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4953 as appropriate. Even if either of those is used, or the application
4954 explicitly uses the version-specific SSLv2_method() or its client and
4955 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4956 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4957 ciphers, and SSLv2 56-bit DES are no longer available.
4962 * Fix a double-free in DSA code
4964 A double free bug was discovered when OpenSSL parses malformed DSA private
4965 keys and could lead to a DoS attack or memory corruption for applications
4966 that receive DSA private keys from untrusted sources. This scenario is
4969 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4975 * Disable SRP fake user seed to address a server memory leak.
4977 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4979 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4980 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4981 was changed to ignore the "fake user" SRP seed, even if the seed
4984 Users should use SRP_VBASE_get1_by_user instead. Note that in
4985 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4986 also that even though configuring the SRP seed attempts to hide
4987 invalid usernames by continuing the handshake with fake
4988 credentials, this behaviour is not constant time and no strong
4989 guarantees are made that the handshake is indistinguishable from
4990 that of a valid user.
4995 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4997 In the BN_hex2bn function the number of hex digits is calculated using an
4998 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4999 large values of `i` this can result in `bn_expand` not allocating any
5000 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5001 field as NULL leading to a subsequent NULL ptr deref. For very large values
5002 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5003 In this case memory is allocated to the internal BIGNUM data field, but it
5004 is insufficiently sized leading to heap corruption. A similar issue exists
5005 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5006 is ever called by user applications with very large untrusted hex/dec data.
5007 This is anticipated to be a rare occurrence.
5009 All OpenSSL internal usage of these functions use data that is not expected
5010 to be untrusted, e.g. config file data or application command line
5011 arguments. If user developed applications generate config file data based
5012 on untrusted data then it is possible that this could also lead to security
5013 consequences. This is also anticipated to be rare.
5015 This issue was reported to OpenSSL by Guido Vranken.
5020 * Fix memory issues in `BIO_*printf` functions
5022 The internal `fmtstr` function used in processing a "%s" format string in
5023 the `BIO_*printf` functions could overflow while calculating the length of a
5024 string and cause an OOB read when printing very long strings.
5026 Additionally the internal `doapr_outch` function can attempt to write to an
5027 OOB memory location (at an offset from the NULL pointer) in the event of a
5028 memory allocation failure. In 1.0.2 and below this could be caused where
5029 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5030 could be in processing a very long "%s" format string. Memory leaks can
5033 The first issue may mask the second issue dependent on compiler behaviour.
5034 These problems could enable attacks where large amounts of untrusted data
5035 is passed to the `BIO_*printf` functions. If applications use these functions
5036 in this way then they could be vulnerable. OpenSSL itself uses these
5037 functions when printing out human-readable dumps of ASN.1 data. Therefore
5038 applications that print this data could be vulnerable if the data is from
5039 untrusted sources. OpenSSL command line applications could also be
5040 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5041 as command line arguments.
5043 Libssl is not considered directly vulnerable. Additionally certificates etc
5044 received via remote connections via libssl are also unlikely to be able to
5045 trigger these issues because of message size limits enforced within libssl.
5047 This issue was reported to OpenSSL Guido Vranken.
5052 * Side channel attack on modular exponentiation
5054 A side-channel attack was found which makes use of cache-bank conflicts on
5055 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5056 of RSA keys. The ability to exploit this issue is limited as it relies on
5057 an attacker who has control of code in a thread running on the same
5058 hyper-threaded core as the victim thread which is performing decryptions.
5060 This issue was reported to OpenSSL by Yuval Yarom, The University of
5061 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5062 Nadia Heninger, University of Pennsylvania with more information at
5063 <http://cachebleed.info>.
5068 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5069 if no keysize is specified with default_bits. This fixes an
5070 omission in an earlier change that changed all RSA/DSA key generation
5071 commands to use 2048 bits by default.
5075 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5077 * DH small subgroups
5079 Historically OpenSSL only ever generated DH parameters based on "safe"
5080 primes. More recently (in version 1.0.2) support was provided for
5081 generating X9.42 style parameter files such as those required for RFC 5114
5082 support. The primes used in such files may not be "safe". Where an
5083 application is using DH configured with parameters based on primes that are
5084 not "safe" then an attacker could use this fact to find a peer's private
5085 DH exponent. This attack requires that the attacker complete multiple
5086 handshakes in which the peer uses the same private DH exponent. For example
5087 this could be used to discover a TLS server's private DH exponent if it's
5088 reusing the private DH exponent or it's using a static DH ciphersuite.
5090 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5091 TLS. It is not on by default. If the option is not set then the server
5092 reuses the same private DH exponent for the life of the server process and
5093 would be vulnerable to this attack. It is believed that many popular
5094 applications do set this option and would therefore not be at risk.
5096 The fix for this issue adds an additional check where a "q" parameter is
5097 available (as is the case in X9.42 based parameters). This detects the
5098 only known attack, and is the only possible defense for static DH
5099 ciphersuites. This could have some performance impact.
5101 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5102 default and cannot be disabled. This could have some performance impact.
5104 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5109 * SSLv2 doesn't block disabled ciphers
5111 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5112 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5113 been disabled, provided that the SSLv2 protocol was not also disabled via
5116 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5117 and Sebastian Schinzel.
5122 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5124 * BN_mod_exp may produce incorrect results on x86_64
5126 There is a carry propagating bug in the x86_64 Montgomery squaring
5127 procedure. No EC algorithms are affected. Analysis suggests that attacks
5128 against RSA and DSA as a result of this defect would be very difficult to
5129 perform and are not believed likely. Attacks against DH are considered just
5130 feasible (although very difficult) because most of the work necessary to
5131 deduce information about a private key may be performed offline. The amount
5132 of resources required for such an attack would be very significant and
5133 likely only accessible to a limited number of attackers. An attacker would
5134 additionally need online access to an unpatched system using the target
5135 private key in a scenario with persistent DH parameters and a private
5136 key that is shared between multiple clients. For example this can occur by
5137 default in OpenSSL DHE based SSL/TLS ciphersuites.
5139 This issue was reported to OpenSSL by Hanno Böck.
5144 * Certificate verify crash with missing PSS parameter
5146 The signature verification routines will crash with a NULL pointer
5147 dereference if presented with an ASN.1 signature using the RSA PSS
5148 algorithm and absent mask generation function parameter. Since these
5149 routines are used to verify certificate signature algorithms this can be
5150 used to crash any certificate verification operation and exploited in a
5151 DoS attack. Any application which performs certificate verification is
5152 vulnerable including OpenSSL clients and servers which enable client
5155 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5160 * X509_ATTRIBUTE memory leak
5162 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5163 memory. This structure is used by the PKCS#7 and CMS routines so any
5164 application which reads PKCS#7 or CMS data from untrusted sources is
5165 affected. SSL/TLS is not affected.
5167 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5173 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5174 This changes the decoding behaviour for some invalid messages,
5175 though the change is mostly in the more lenient direction, and
5176 legacy behaviour is preserved as much as possible.
5180 * In DSA_generate_parameters_ex, if the provided seed is too short,
5183 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5185 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5187 * Alternate chains certificate forgery
5189 During certificate verification, OpenSSL will attempt to find an
5190 alternative certificate chain if the first attempt to build such a chain
5191 fails. An error in the implementation of this logic can mean that an
5192 attacker could cause certain checks on untrusted certificates to be
5193 bypassed, such as the CA flag, enabling them to use a valid leaf
5194 certificate to act as a CA and "issue" an invalid certificate.
5196 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5201 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5203 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5204 incompatibility in the handling of HMAC. The previous ABI has now been
5209 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5211 * Malformed ECParameters causes infinite loop
5213 When processing an ECParameters structure OpenSSL enters an infinite loop
5214 if the curve specified is over a specially malformed binary polynomial
5217 This can be used to perform denial of service against any
5218 system which processes public keys, certificate requests or
5219 certificates. This includes TLS clients and TLS servers with
5220 client authentication enabled.
5222 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5227 * Exploitable out-of-bounds read in X509_cmp_time
5229 X509_cmp_time does not properly check the length of the ASN1_TIME
5230 string and can read a few bytes out of bounds. In addition,
5231 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5234 An attacker can use this to craft malformed certificates and CRLs of
5235 various sizes and potentially cause a segmentation fault, resulting in
5236 a DoS on applications that verify certificates or CRLs. TLS clients
5237 that verify CRLs are affected. TLS clients and servers with client
5238 authentication enabled may be affected if they use custom verification
5241 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5242 independently by Hanno Böck.
5247 * PKCS7 crash with missing EnvelopedContent
5249 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5250 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5251 with missing content and trigger a NULL pointer dereference on parsing.
5253 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5254 structures from untrusted sources are affected. OpenSSL clients and
5255 servers are not affected.
5257 This issue was reported to OpenSSL by Michal Zalewski (Google).
5262 * CMS verify infinite loop with unknown hash function
5264 When verifying a signedData message the CMS code can enter an infinite loop
5265 if presented with an unknown hash function OID. This can be used to perform
5266 denial of service against any system which verifies signedData messages using
5268 This issue was reported to OpenSSL by Johannes Bauer.
5273 * Race condition handling NewSessionTicket
5275 If a NewSessionTicket is received by a multi-threaded client when attempting to
5276 reuse a previous ticket then a race condition can occur potentially leading to
5277 a double free of the ticket data.
5282 * Only support 256-bit or stronger elliptic curves with the
5283 'ecdh_auto' setting (server) or by default (client). Of supported
5284 curves, prefer P-256 (both).
5288 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5290 * ClientHello sigalgs DoS fix
5292 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5293 invalid signature algorithms extension a NULL pointer dereference will
5294 occur. This can be exploited in a DoS attack against the server.
5296 This issue was was reported to OpenSSL by David Ramos of Stanford
5300 *Stephen Henson and Matt Caswell*
5302 * Multiblock corrupted pointer fix
5304 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5305 feature only applies on 64 bit x86 architecture platforms that support AES
5306 NI instructions. A defect in the implementation of "multiblock" can cause
5307 OpenSSL's internal write buffer to become incorrectly set to NULL when
5308 using non-blocking IO. Typically, when the user application is using a
5309 socket BIO for writing, this will only result in a failed connection.
5310 However if some other BIO is used then it is likely that a segmentation
5311 fault will be triggered, thus enabling a potential DoS attack.
5313 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5318 * Segmentation fault in DTLSv1_listen fix
5320 The DTLSv1_listen function is intended to be stateless and processes the
5321 initial ClientHello from many peers. It is common for user code to loop
5322 over the call to DTLSv1_listen until a valid ClientHello is received with
5323 an associated cookie. A defect in the implementation of DTLSv1_listen means
5324 that state is preserved in the SSL object from one invocation to the next
5325 that can lead to a segmentation fault. Errors processing the initial
5326 ClientHello can trigger this scenario. An example of such an error could be
5327 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5330 This issue was reported to OpenSSL by Per Allansson.
5335 * Segmentation fault in ASN1_TYPE_cmp fix
5337 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5338 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5339 certificate signature algorithm consistency this can be used to crash any
5340 certificate verification operation and exploited in a DoS attack. Any
5341 application which performs certificate verification is vulnerable including
5342 OpenSSL clients and servers which enable client authentication.
5347 * Segmentation fault for invalid PSS parameters fix
5349 The signature verification routines will crash with a NULL pointer
5350 dereference if presented with an ASN.1 signature using the RSA PSS
5351 algorithm and invalid parameters. Since these routines are used to verify
5352 certificate signature algorithms this can be used to crash any
5353 certificate verification operation and exploited in a DoS attack. Any
5354 application which performs certificate verification is vulnerable including
5355 OpenSSL clients and servers which enable client authentication.
5357 This issue was was reported to OpenSSL by Brian Carpenter.
5362 * ASN.1 structure reuse memory corruption fix
5364 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5365 memory corruption via an invalid write. Such reuse is and has been
5366 strongly discouraged and is believed to be rare.
5368 Applications that parse structures containing CHOICE or ANY DEFINED BY
5369 components may be affected. Certificate parsing (d2i_X509 and related
5370 functions) are however not affected. OpenSSL clients and servers are
5376 * PKCS7 NULL pointer dereferences fix
5378 The PKCS#7 parsing code does not handle missing outer ContentInfo
5379 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5380 missing content and trigger a NULL pointer dereference on parsing.
5382 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5383 otherwise parse PKCS#7 structures from untrusted sources are
5384 affected. OpenSSL clients and servers are not affected.
5386 This issue was reported to OpenSSL by Michal Zalewski (Google).
5391 * DoS via reachable assert in SSLv2 servers fix
5393 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5394 servers that both support SSLv2 and enable export cipher suites by sending
5395 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5397 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5398 (OpenSSL development team).
5403 * Empty CKE with client auth and DHE fix
5405 If client auth is used then a server can seg fault in the event of a DHE
5406 ciphersuite being selected and a zero length ClientKeyExchange message
5407 being sent by the client. This could be exploited in a DoS attack.
5412 * Handshake with unseeded PRNG fix
5414 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5415 with an unseeded PRNG. The conditions are:
5416 - The client is on a platform where the PRNG has not been seeded
5417 automatically, and the user has not seeded manually
5418 - A protocol specific client method version has been used (i.e. not
5419 SSL_client_methodv23)
5420 - A ciphersuite is used that does not require additional random data from
5421 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5423 If the handshake succeeds then the client random that has been used will
5424 have been generated from a PRNG with insufficient entropy and therefore the
5425 output may be predictable.
5427 For example using the following command with an unseeded openssl will
5428 succeed on an unpatched platform:
5430 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5435 * Use After Free following d2i_ECPrivatekey error fix
5437 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5438 could cause a use after free condition. This, in turn, could cause a double
5439 free in several private key parsing functions (such as d2i_PrivateKey
5440 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5441 for applications that receive EC private keys from untrusted
5442 sources. This scenario is considered rare.
5444 This issue was discovered by the BoringSSL project and fixed in their
5450 * X509_to_X509_REQ NULL pointer deref fix
5452 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5453 the certificate key is invalid. This function is rarely used in practice.
5455 This issue was discovered by Brian Carpenter.
5460 * Removed the export ciphers from the DEFAULT ciphers
5464 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5466 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5467 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5468 So far those who have to target multiple platforms would compromise
5469 and argue that binary targeting say ARMv5 would still execute on
5470 ARMv8. "Universal" build resolves this compromise by providing
5471 near-optimal performance even on newer platforms.
5475 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5476 (other platforms pending).
5478 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5480 * Add support for the SignedCertificateTimestampList certificate and
5481 OCSP response extensions from RFC6962.
5485 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5486 for corner cases. (Certain input points at infinity could lead to
5487 bogus results, with non-infinity inputs mapped to infinity too.)
5491 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5492 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5493 common cases are optimized and there still is room for further
5494 improvements. Vector Permutation AES for Altivec is also added.
5498 * Add support for little-endian ppc64 Linux target.
5500 *Marcelo Cerri (IBM)*
5502 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5503 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5504 are optimized and there still is room for further improvements.
5505 Both 32- and 64-bit modes are supported.
5507 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5509 * Improved ARMv7 NEON support.
5513 * Support for SPARC Architecture 2011 crypto extensions, first
5514 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5515 SHA256/512, MD5, GHASH and modular exponentiation.
5517 *Andy Polyakov, David Miller*
5519 * Accelerated modular exponentiation for Intel processors, a.k.a.
5522 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5524 * Support for new and upcoming Intel processors, including AVX2,
5525 BMI and SHA ISA extensions. This includes additional "stitched"
5526 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5529 This work was sponsored by Intel Corp.
5533 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5534 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5535 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5539 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5540 this fixes a limitation in previous versions of OpenSSL.
5544 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5545 MGF1 digest and OAEP label.
5549 * Add EVP support for key wrapping algorithms, to avoid problems with
5550 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5551 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5552 algorithms and include tests cases.
5556 * Add functions to allocate and set the fields of an ECDSA_METHOD
5559 *Douglas E. Engert, Steve Henson*
5561 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5562 difference in days and seconds between two tm or ASN1_TIME structures.
5566 * Add -rev test option to s_server to just reverse order of characters
5567 received by client and send back to server. Also prints an abbreviated
5568 summary of the connection parameters.
5572 * New option -brief for s_client and s_server to print out a brief summary
5573 of connection parameters.
5577 * Add callbacks for arbitrary TLS extensions.
5579 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5581 * New option -crl_download in several openssl utilities to download CRLs
5582 from CRLDP extension in certificates.
5586 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5590 * New function X509_CRL_diff to generate a delta CRL from the difference
5591 of two full CRLs. Add support to "crl" utility.
5595 * New functions to set lookup_crls function and to retrieve
5596 X509_STORE from X509_STORE_CTX.
5600 * Print out deprecated issuer and subject unique ID fields in
5605 * Extend OCSP I/O functions so they can be used for simple general purpose
5606 HTTP as well as OCSP. New wrapper function which can be used to download
5607 CRLs using the OCSP API.
5611 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5615 * `SSL_CONF*` functions. These provide a common framework for application
5616 configuration using configuration files or command lines.
5620 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5621 message callback and prints the results. Needs compile time option
5622 "enable-ssl-trace". New options to s_client and s_server to enable
5627 * New ctrl and macro to retrieve supported points extensions.
5628 Print out extension in s_server and s_client.
5632 * New functions to retrieve certificate signature and signature
5637 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5642 * New Suite B modes for TLS code. These use and enforce the requirements
5643 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5644 only use Suite B curves. The Suite B modes can be set by using the
5645 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5649 * New chain verification flags for Suite B levels of security. Check
5650 algorithms are acceptable when flags are set in X509_verify_cert.
5654 * Make tls1_check_chain return a set of flags indicating checks passed
5655 by a certificate chain. Add additional tests to handle client
5656 certificates: checks for matching certificate type and issuer name
5661 * If an attempt is made to use a signature algorithm not in the peer
5662 preference list abort the handshake. If client has no suitable
5663 signature algorithms in response to a certificate request do not
5664 use the certificate.
5668 * If server EC tmp key is not in client preference list abort handshake.
5672 * Add support for certificate stores in CERT structure. This makes it
5673 possible to have different stores per SSL structure or one store in
5674 the parent SSL_CTX. Include distinct stores for certificate chain
5675 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5676 to build and store a certificate chain in CERT structure: returning
5677 an error if the chain cannot be built: this will allow applications
5678 to test if a chain is correctly configured.
5680 Note: if the CERT based stores are not set then the parent SSL_CTX
5681 store is used to retain compatibility with existing behaviour.
5685 * New function ssl_set_client_disabled to set a ciphersuite disabled
5686 mask based on the current session, check mask when sending client
5687 hello and checking the requested ciphersuite.
5691 * New ctrls to retrieve and set certificate types in a certificate
5692 request message. Print out received values in s_client. If certificate
5693 types is not set with custom values set sensible values based on
5694 supported signature algorithms.
5698 * Support for distinct client and server supported signature algorithms.
5702 * Add certificate callback. If set this is called whenever a certificate
5703 is required by client or server. An application can decide which
5704 certificate chain to present based on arbitrary criteria: for example
5705 supported signature algorithms. Add very simple example to s_server.
5706 This fixes many of the problems and restrictions of the existing client
5707 certificate callback: for example you can now clear an existing
5708 certificate and specify the whole chain.
5712 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5713 the certificate can be used for (if anything). Set valid_flags field
5714 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5715 to have similar checks in it.
5717 Add new "cert_flags" field to CERT structure and include a "strict mode".
5718 This enforces some TLS certificate requirements (such as only permitting
5719 certificate signature algorithms contained in the supported algorithms
5720 extension) which some implementations ignore: this option should be used
5721 with caution as it could cause interoperability issues.
5725 * Update and tidy signature algorithm extension processing. Work out
5726 shared signature algorithms based on preferences and peer algorithms
5727 and print them out in s_client and s_server. Abort handshake if no
5728 shared signature algorithms.
5732 * Add new functions to allow customised supported signature algorithms
5733 for SSL and SSL_CTX structures. Add options to s_client and s_server
5738 * New function SSL_certs_clear() to delete all references to certificates
5739 from an SSL structure. Before this once a certificate had been added
5740 it couldn't be removed.
5744 * Integrate hostname, email address and IP address checking with certificate
5745 verification. New verify options supporting checking in openssl utility.
5749 * Fixes and wildcard matching support to hostname and email checking
5750 functions. Add manual page.
5752 *Florian Weimer (Red Hat Product Security Team)*
5754 * New functions to check a hostname email or IP address against a
5755 certificate. Add options x509 utility to print results of checks against
5760 * Fix OCSP checking.
5762 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5764 * Initial experimental support for explicitly trusted non-root CAs.
5765 OpenSSL still tries to build a complete chain to a root but if an
5766 intermediate CA has a trust setting included that is used. The first
5767 setting is used: whether to trust (e.g., -addtrust option to the x509
5772 * Add -trusted_first option which attempts to find certificates in the
5773 trusted store even if an untrusted chain is also supplied.
5777 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5778 platform support for Linux and Android.
5782 * Support for linux-x32, ILP32 environment in x86_64 framework.
5786 * Experimental multi-implementation support for FIPS capable OpenSSL.
5787 When in FIPS mode the approved implementations are used as normal,
5788 when not in FIPS mode the internal unapproved versions are used instead.
5789 This means that the FIPS capable OpenSSL isn't forced to use the
5790 (often lower performance) FIPS implementations outside FIPS mode.
5794 * Transparently support X9.42 DH parameters when calling
5795 PEM_read_bio_DHparameters. This means existing applications can handle
5796 the new parameter format automatically.
5800 * Initial experimental support for X9.42 DH parameter format: mainly
5801 to support use of 'q' parameter for RFC5114 parameters.
5805 * Add DH parameters from RFC5114 including test data to dhtest.
5809 * Support for automatic EC temporary key parameter selection. If enabled
5810 the most preferred EC parameters are automatically used instead of
5811 hardcoded fixed parameters. Now a server just has to call:
5812 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5813 support ECDH and use the most appropriate parameters.
5817 * Enhance and tidy EC curve and point format TLS extension code. Use
5818 static structures instead of allocation if default values are used.
5819 New ctrls to set curves we wish to support and to retrieve shared curves.
5820 Print out shared curves in s_server. New options to s_server and s_client
5821 to set list of supported curves.
5825 * New ctrls to retrieve supported signature algorithms and
5826 supported curve values as an array of NIDs. Extend openssl utility
5827 to print out received values.
5831 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5832 between NIDs and the more common NIST names such as "P-256". Enhance
5833 ecparam utility and ECC method to recognise the NIST names for curves.
5837 * Enhance SSL/TLS certificate chain handling to support different
5838 chains for each certificate instead of one chain in the parent SSL_CTX.
5842 * Support for fixed DH ciphersuite client authentication: where both
5843 server and client use DH certificates with common parameters.
5847 * Support for fixed DH ciphersuites: those requiring DH server
5852 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5854 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5855 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5856 X509_CINF_get_signature were reverted post internal team review.
5861 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5863 * OCSP Status Request extension unbounded memory growth
5865 A malicious client can send an excessively large OCSP Status Request
5866 extension. If that client continually requests renegotiation, sending a
5867 large OCSP Status Request extension each time, then there will be unbounded
5868 memory growth on the server. This will eventually lead to a Denial Of
5869 Service attack through memory exhaustion. Servers with a default
5870 configuration are vulnerable even if they do not support OCSP. Builds using
5871 the "no-ocsp" build time option are not affected.
5873 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5878 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5881 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5887 * OOB write in MDC2_Update()
5889 An overflow can occur in MDC2_Update() either if called directly or
5890 through the EVP_DigestUpdate() function using MDC2. If an attacker
5891 is able to supply very large amounts of input data after a previous
5892 call to EVP_EncryptUpdate() with a partial block then a length check
5893 can overflow resulting in a heap corruption.
5895 The amount of data needed is comparable to SIZE_MAX which is impractical
5898 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5903 * Malformed SHA512 ticket DoS
5905 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5906 DoS attack where a malformed ticket will result in an OOB read which will
5909 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5910 a custom server callback and ticket lookup mechanism.
5912 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5917 * OOB write in BN_bn2dec()
5919 The function BN_bn2dec() does not check the return value of BN_div_word().
5920 This can cause an OOB write if an application uses this function with an
5921 overly large BIGNUM. This could be a problem if an overly large certificate
5922 or CRL is printed out from an untrusted source. TLS is not affected because
5923 record limits will reject an oversized certificate before it is parsed.
5925 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5930 * OOB read in TS_OBJ_print_bio()
5932 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5933 the total length the OID text representation would use and not the amount
5934 of data written. This will result in OOB reads when large OIDs are
5937 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5942 * Pointer arithmetic undefined behaviour
5944 Avoid some undefined pointer arithmetic
5946 A common idiom in the codebase is to check limits in the following manner:
5949 Where "p" points to some malloc'd data of SIZE bytes and
5952 "len" here could be from some externally supplied data (e.g. from a TLS
5955 The rules of C pointer arithmetic are such that "p + len" is only well
5956 defined where len <= SIZE. Therefore the above idiom is actually
5957 undefined behaviour.
5959 For example this could cause problems if some malloc implementation
5960 provides an address for "p" such that "p + len" actually overflows for
5961 values of len that are too big and therefore p + len < limit.
5963 This issue was reported to OpenSSL by Guido Vranken
5968 * Constant time flag not preserved in DSA signing
5970 Operations in the DSA signing algorithm should run in constant time in
5971 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5972 implementation means that a non-constant time codepath is followed for
5973 certain operations. This has been demonstrated through a cache-timing
5974 attack to be sufficient for an attacker to recover the private DSA key.
5976 This issue was reported by César Pereida (Aalto University), Billy Brumley
5977 (Tampere University of Technology), and Yuval Yarom (The University of
5978 Adelaide and NICTA).
5983 * DTLS buffered message DoS
5985 In a DTLS connection where handshake messages are delivered out-of-order
5986 those messages that OpenSSL is not yet ready to process will be buffered
5987 for later use. Under certain circumstances, a flaw in the logic means that
5988 those messages do not get removed from the buffer even though the handshake
5989 has been completed. An attacker could force up to approx. 15 messages to
5990 remain in the buffer when they are no longer required. These messages will
5991 be cleared when the DTLS connection is closed. The default maximum size for
5992 a message is 100k. Therefore the attacker could force an additional 1500k
5993 to be consumed per connection. By opening many simulataneous connections an
5994 attacker could cause a DoS attack through memory exhaustion.
5996 This issue was reported to OpenSSL by Quan Luo.
6001 * DTLS replay protection DoS
6003 A flaw in the DTLS replay attack protection mechanism means that records
6004 that arrive for future epochs update the replay protection "window" before
6005 the MAC for the record has been validated. This could be exploited by an
6006 attacker by sending a record for the next epoch (which does not have to
6007 decrypt or have a valid MAC), with a very large sequence number. This means
6008 that all subsequent legitimate packets are dropped causing a denial of
6009 service for a specific DTLS connection.
6011 This issue was reported to OpenSSL by the OCAP audit team.
6016 * Certificate message OOB reads
6018 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6019 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6020 theoretical DoS risk but this has not been observed in practice on common
6023 The messages affected are client certificate, client certificate request
6024 and server certificate. As a result the attack can only be performed
6025 against a client or a server which enables client authentication.
6027 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6032 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6034 * Prevent padding oracle in AES-NI CBC MAC check
6036 A MITM attacker can use a padding oracle attack to decrypt traffic
6037 when the connection uses an AES CBC cipher and the server support
6040 This issue was introduced as part of the fix for Lucky 13 padding
6041 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6042 constant time by making sure that always the same bytes are read and
6043 compared against either the MAC or padding bytes. But it no longer
6044 checked that there was enough data to have both the MAC and padding
6047 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6052 * Fix EVP_EncodeUpdate overflow
6054 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6055 Base64 encoding of binary data. If an attacker is able to supply very large
6056 amounts of input data then a length check can overflow resulting in a heap
6059 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6060 the `PEM_write_bio*` family of functions. These are mainly used within the
6061 OpenSSL command line applications, so any application which processes data
6062 from an untrusted source and outputs it as a PEM file should be considered
6063 vulnerable to this issue. User applications that call these APIs directly
6064 with large amounts of untrusted data may also be vulnerable.
6066 This issue was reported by Guido Vranken.
6071 * Fix EVP_EncryptUpdate overflow
6073 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6074 is able to supply very large amounts of input data after a previous call to
6075 EVP_EncryptUpdate() with a partial block then a length check can overflow
6076 resulting in a heap corruption. Following an analysis of all OpenSSL
6077 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6078 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6079 the first called function after an EVP_EncryptInit(), and therefore that
6080 specific call must be safe. The second form is where the length passed to
6081 EVP_EncryptUpdate() can be seen from the code to be some small value and
6082 therefore there is no possibility of an overflow. Since all instances are
6083 one of these two forms, it is believed that there can be no overflows in
6084 internal code due to this problem. It should be noted that
6085 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6086 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6087 of these calls have also been analysed too and it is believed there are no
6088 instances in internal usage where an overflow could occur.
6090 This issue was reported by Guido Vranken.
6095 * Prevent ASN.1 BIO excessive memory allocation
6097 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6098 a short invalid encoding can casuse allocation of large amounts of memory
6099 potentially consuming excessive resources or exhausting memory.
6101 Any application parsing untrusted data through d2i BIO functions is
6102 affected. The memory based functions such as d2i_X509() are *not* affected.
6103 Since the memory based functions are used by the TLS library, TLS
6104 applications are not affected.
6106 This issue was reported by Brian Carpenter.
6113 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6114 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6115 in arbitrary stack data being returned in the buffer.
6117 This issue was reported by Guido Vranken.
6122 * Modify behavior of ALPN to invoke callback after SNI/servername
6123 callback, such that updates to the SSL_CTX affect ALPN.
6127 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6132 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6133 methods are enabled and ssl2 is disabled the methods return NULL.
6137 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6139 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6140 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6141 provide any "EXPORT" or "LOW" strength ciphers.
6145 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6146 is by default disabled at build-time. Builds that are not configured with
6147 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6148 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6149 will need to explicitly call either of:
6151 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6153 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6155 as appropriate. Even if either of those is used, or the application
6156 explicitly uses the version-specific SSLv2_method() or its client and
6157 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6158 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6159 ciphers, and SSLv2 56-bit DES are no longer available.
6164 * Fix a double-free in DSA code
6166 A double free bug was discovered when OpenSSL parses malformed DSA private
6167 keys and could lead to a DoS attack or memory corruption for applications
6168 that receive DSA private keys from untrusted sources. This scenario is
6171 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6177 * Disable SRP fake user seed to address a server memory leak.
6179 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6181 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6182 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6183 was changed to ignore the "fake user" SRP seed, even if the seed
6186 Users should use SRP_VBASE_get1_by_user instead. Note that in
6187 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6188 also that even though configuring the SRP seed attempts to hide
6189 invalid usernames by continuing the handshake with fake
6190 credentials, this behaviour is not constant time and no strong
6191 guarantees are made that the handshake is indistinguishable from
6192 that of a valid user.
6197 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6199 In the BN_hex2bn function the number of hex digits is calculated using an
6200 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6201 large values of `i` this can result in `bn_expand` not allocating any
6202 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6203 field as NULL leading to a subsequent NULL ptr deref. For very large values
6204 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6205 In this case memory is allocated to the internal BIGNUM data field, but it
6206 is insufficiently sized leading to heap corruption. A similar issue exists
6207 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6208 is ever called by user applications with very large untrusted hex/dec data.
6209 This is anticipated to be a rare occurrence.
6211 All OpenSSL internal usage of these functions use data that is not expected
6212 to be untrusted, e.g. config file data or application command line
6213 arguments. If user developed applications generate config file data based
6214 on untrusted data then it is possible that this could also lead to security
6215 consequences. This is also anticipated to be rare.
6217 This issue was reported to OpenSSL by Guido Vranken.
6222 * Fix memory issues in `BIO_*printf` functions
6224 The internal `fmtstr` function used in processing a "%s" format string in
6225 the `BIO_*printf` functions could overflow while calculating the length of a
6226 string and cause an OOB read when printing very long strings.
6228 Additionally the internal `doapr_outch` function can attempt to write to an
6229 OOB memory location (at an offset from the NULL pointer) in the event of a
6230 memory allocation failure. In 1.0.2 and below this could be caused where
6231 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6232 could be in processing a very long "%s" format string. Memory leaks can
6235 The first issue may mask the second issue dependent on compiler behaviour.
6236 These problems could enable attacks where large amounts of untrusted data
6237 is passed to the `BIO_*printf` functions. If applications use these functions
6238 in this way then they could be vulnerable. OpenSSL itself uses these
6239 functions when printing out human-readable dumps of ASN.1 data. Therefore
6240 applications that print this data could be vulnerable if the data is from
6241 untrusted sources. OpenSSL command line applications could also be
6242 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6243 as command line arguments.
6245 Libssl is not considered directly vulnerable. Additionally certificates etc
6246 received via remote connections via libssl are also unlikely to be able to
6247 trigger these issues because of message size limits enforced within libssl.
6249 This issue was reported to OpenSSL Guido Vranken.
6254 * Side channel attack on modular exponentiation
6256 A side-channel attack was found which makes use of cache-bank conflicts on
6257 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6258 of RSA keys. The ability to exploit this issue is limited as it relies on
6259 an attacker who has control of code in a thread running on the same
6260 hyper-threaded core as the victim thread which is performing decryptions.
6262 This issue was reported to OpenSSL by Yuval Yarom, The University of
6263 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6264 Nadia Heninger, University of Pennsylvania with more information at
6265 <http://cachebleed.info>.
6270 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6271 if no keysize is specified with default_bits. This fixes an
6272 omission in an earlier change that changed all RSA/DSA key generation
6273 commands to use 2048 bits by default.
6277 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6279 * Protection for DH small subgroup attacks
6281 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6282 switched on by default and cannot be disabled. This could have some
6287 * SSLv2 doesn't block disabled ciphers
6289 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6290 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6291 been disabled, provided that the SSLv2 protocol was not also disabled via
6294 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6295 and Sebastian Schinzel.
6300 * Reject DH handshakes with parameters shorter than 1024 bits.
6304 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6306 * Certificate verify crash with missing PSS parameter
6308 The signature verification routines will crash with a NULL pointer
6309 dereference if presented with an ASN.1 signature using the RSA PSS
6310 algorithm and absent mask generation function parameter. Since these
6311 routines are used to verify certificate signature algorithms this can be
6312 used to crash any certificate verification operation and exploited in a
6313 DoS attack. Any application which performs certificate verification is
6314 vulnerable including OpenSSL clients and servers which enable client
6317 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6322 * X509_ATTRIBUTE memory leak
6324 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6325 memory. This structure is used by the PKCS#7 and CMS routines so any
6326 application which reads PKCS#7 or CMS data from untrusted sources is
6327 affected. SSL/TLS is not affected.
6329 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6335 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6336 This changes the decoding behaviour for some invalid messages,
6337 though the change is mostly in the more lenient direction, and
6338 legacy behaviour is preserved as much as possible.
6342 * In DSA_generate_parameters_ex, if the provided seed is too short,
6343 use a random seed, as already documented.
6345 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6347 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6349 * Alternate chains certificate forgery
6351 During certificate verfification, OpenSSL will attempt to find an
6352 alternative certificate chain if the first attempt to build such a chain
6353 fails. An error in the implementation of this logic can mean that an
6354 attacker could cause certain checks on untrusted certificates to be
6355 bypassed, such as the CA flag, enabling them to use a valid leaf
6356 certificate to act as a CA and "issue" an invalid certificate.
6358 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6364 * Race condition handling PSK identify hint
6366 If PSK identity hints are received by a multi-threaded client then
6367 the values are wrongly updated in the parent SSL_CTX structure. This can
6368 result in a race condition potentially leading to a double free of the
6374 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6376 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6377 incompatibility in the handling of HMAC. The previous ABI has now been
6380 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6382 * Malformed ECParameters causes infinite loop
6384 When processing an ECParameters structure OpenSSL enters an infinite loop
6385 if the curve specified is over a specially malformed binary polynomial
6388 This can be used to perform denial of service against any
6389 system which processes public keys, certificate requests or
6390 certificates. This includes TLS clients and TLS servers with
6391 client authentication enabled.
6393 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6398 * Exploitable out-of-bounds read in X509_cmp_time
6400 X509_cmp_time does not properly check the length of the ASN1_TIME
6401 string and can read a few bytes out of bounds. In addition,
6402 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6405 An attacker can use this to craft malformed certificates and CRLs of
6406 various sizes and potentially cause a segmentation fault, resulting in
6407 a DoS on applications that verify certificates or CRLs. TLS clients
6408 that verify CRLs are affected. TLS clients and servers with client
6409 authentication enabled may be affected if they use custom verification
6412 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6413 independently by Hanno Böck.
6418 * PKCS7 crash with missing EnvelopedContent
6420 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6421 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6422 with missing content and trigger a NULL pointer dereference on parsing.
6424 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6425 structures from untrusted sources are affected. OpenSSL clients and
6426 servers are not affected.
6428 This issue was reported to OpenSSL by Michal Zalewski (Google).
6433 * CMS verify infinite loop with unknown hash function
6435 When verifying a signedData message the CMS code can enter an infinite loop
6436 if presented with an unknown hash function OID. This can be used to perform
6437 denial of service against any system which verifies signedData messages using
6439 This issue was reported to OpenSSL by Johannes Bauer.
6444 * Race condition handling NewSessionTicket
6446 If a NewSessionTicket is received by a multi-threaded client when attempting to
6447 reuse a previous ticket then a race condition can occur potentially leading to
6448 a double free of the ticket data.
6453 * Reject DH handshakes with parameters shorter than 768 bits.
6455 *Kurt Roeckx and Emilia Kasper*
6457 * dhparam: generate 2048-bit parameters by default.
6459 *Kurt Roeckx and Emilia Kasper*
6461 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6463 * Segmentation fault in ASN1_TYPE_cmp fix
6465 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6466 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6467 certificate signature algorithm consistency this can be used to crash any
6468 certificate verification operation and exploited in a DoS attack. Any
6469 application which performs certificate verification is vulnerable including
6470 OpenSSL clients and servers which enable client authentication.
6475 * ASN.1 structure reuse memory corruption fix
6477 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6478 memory corruption via an invalid write. Such reuse is and has been
6479 strongly discouraged and is believed to be rare.
6481 Applications that parse structures containing CHOICE or ANY DEFINED BY
6482 components may be affected. Certificate parsing (d2i_X509 and related
6483 functions) are however not affected. OpenSSL clients and servers are
6489 * PKCS7 NULL pointer dereferences fix
6491 The PKCS#7 parsing code does not handle missing outer ContentInfo
6492 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6493 missing content and trigger a NULL pointer dereference on parsing.
6495 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6496 otherwise parse PKCS#7 structures from untrusted sources are
6497 affected. OpenSSL clients and servers are not affected.
6499 This issue was reported to OpenSSL by Michal Zalewski (Google).
6504 * DoS via reachable assert in SSLv2 servers fix
6506 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6507 servers that both support SSLv2 and enable export cipher suites by sending
6508 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6510 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6511 (OpenSSL development team).
6516 * Use After Free following d2i_ECPrivatekey error fix
6518 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6519 could cause a use after free condition. This, in turn, could cause a double
6520 free in several private key parsing functions (such as d2i_PrivateKey
6521 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6522 for applications that receive EC private keys from untrusted
6523 sources. This scenario is considered rare.
6525 This issue was discovered by the BoringSSL project and fixed in their
6531 * X509_to_X509_REQ NULL pointer deref fix
6533 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6534 the certificate key is invalid. This function is rarely used in practice.
6536 This issue was discovered by Brian Carpenter.
6541 * Removed the export ciphers from the DEFAULT ciphers
6545 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6547 * Build fixes for the Windows and OpenVMS platforms
6549 *Matt Caswell and Richard Levitte*
6551 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6553 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6554 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6555 dereference. This could lead to a Denial Of Service attack. Thanks to
6556 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6561 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6562 dtls1_buffer_record function under certain conditions. In particular this
6563 could occur if an attacker sent repeated DTLS records with the same
6564 sequence number but for the next epoch. The memory leak could be exploited
6565 by an attacker in a Denial of Service attack through memory exhaustion.
6566 Thanks to Chris Mueller for reporting this issue.
6571 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6572 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6573 method would be set to NULL which could later result in a NULL pointer
6574 dereference. Thanks to Frank Schmirler for reporting this issue.
6579 * Abort handshake if server key exchange message is omitted for ephemeral
6582 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6583 reporting this issue.
6588 * Remove non-export ephemeral RSA code on client and server. This code
6589 violated the TLS standard by allowing the use of temporary RSA keys in
6590 non-export ciphersuites and could be used by a server to effectively
6591 downgrade the RSA key length used to a value smaller than the server
6592 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6593 INRIA or reporting this issue.
6598 * Fixed issue where DH client certificates are accepted without verification.
6599 An OpenSSL server will accept a DH certificate for client authentication
6600 without the certificate verify message. This effectively allows a client to
6601 authenticate without the use of a private key. This only affects servers
6602 which trust a client certificate authority which issues certificates
6603 containing DH keys: these are extremely rare and hardly ever encountered.
6604 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6610 * Ensure that the session ID context of an SSL is updated when its
6611 SSL_CTX is updated via SSL_set_SSL_CTX.
6613 The session ID context is typically set from the parent SSL_CTX,
6614 and can vary with the CTX.
6618 * Fix various certificate fingerprint issues.
6620 By using non-DER or invalid encodings outside the signed portion of a
6621 certificate the fingerprint can be changed without breaking the signature.
6622 Although no details of the signed portion of the certificate can be changed
6623 this can cause problems with some applications: e.g. those using the
6624 certificate fingerprint for blacklists.
6626 1. Reject signatures with non zero unused bits.
6628 If the BIT STRING containing the signature has non zero unused bits reject
6629 the signature. All current signature algorithms require zero unused bits.
6631 2. Check certificate algorithm consistency.
6633 Check the AlgorithmIdentifier inside TBS matches the one in the
6634 certificate signature. NB: this will result in signature failure
6635 errors for some broken certificates.
6637 Thanks to Konrad Kraszewski from Google for reporting this issue.
6639 3. Check DSA/ECDSA signatures use DER.
6641 Re-encode DSA/ECDSA signatures and compare with the original received
6642 signature. Return an error if there is a mismatch.
6644 This will reject various cases including garbage after signature
6645 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6646 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6647 (negative or with leading zeroes).
6649 Further analysis was conducted and fixes were developed by Stephen Henson
6650 of the OpenSSL core team.
6656 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6657 results on some platforms, including x86_64. This bug occurs at random
6658 with a very low probability, and is not known to be exploitable in any
6659 way, though its exact impact is difficult to determine. Thanks to Pieter
6660 Wuille (Blockstream) who reported this issue and also suggested an initial
6661 fix. Further analysis was conducted by the OpenSSL development team and
6662 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6663 the OpenSSL core team.
6668 * Do not resume sessions on the server if the negotiated protocol
6669 version does not match the session's version. Resuming with a different
6670 version, while not strictly forbidden by the RFC, is of questionable
6671 sanity and breaks all known clients.
6673 *David Benjamin, Emilia Käsper*
6675 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6676 early CCS messages during renegotiation. (Note that because
6677 renegotiation is encrypted, this early CCS was not exploitable.)
6681 * Tighten client-side session ticket handling during renegotiation:
6682 ensure that the client only accepts a session ticket if the server sends
6683 the extension anew in the ServerHello. Previously, a TLS client would
6684 reuse the old extension state and thus accept a session ticket if one was
6685 announced in the initial ServerHello.
6687 Similarly, ensure that the client requires a session ticket if one
6688 was advertised in the ServerHello. Previously, a TLS client would
6689 ignore a missing NewSessionTicket message.
6693 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6697 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6698 sends a carefully crafted handshake message, to cause OpenSSL to fail
6699 to free up to 64k of memory causing a memory leak. This could be
6700 exploited in a Denial Of Service attack. This issue affects OpenSSL
6701 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6702 whether SRTP is used or configured. Implementations of OpenSSL that
6703 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6705 The fix was developed by the OpenSSL team.
6710 * Session Ticket Memory Leak.
6712 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6713 integrity of that ticket is first verified. In the event of a session
6714 ticket integrity check failing, OpenSSL will fail to free memory
6715 causing a memory leak. By sending a large number of invalid session
6716 tickets an attacker could exploit this issue in a Denial Of Service
6722 * Build option no-ssl3 is incomplete.
6724 When OpenSSL is configured with "no-ssl3" as a build option, servers
6725 could accept and complete a SSL 3.0 handshake, and clients could be
6726 configured to send them.
6729 *Akamai and the OpenSSL team*
6731 * Add support for TLS_FALLBACK_SCSV.
6732 Client applications doing fallback retries should call
6733 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6736 *Adam Langley, Bodo Moeller*
6738 * Add additional DigestInfo checks.
6740 Re-encode DigestInto in DER and check against the original when
6741 verifying RSA signature: this will reject any improperly encoded
6742 DigestInfo structures.
6744 Note: this is a precautionary measure and no attacks are currently known.
6748 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6750 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6751 SRP code can be overrun an internal buffer. Add sanity check that
6752 g, A, B < N to SRP code.
6754 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6755 Group for discovering this issue.
6760 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6761 TLS 1.0 instead of higher protocol versions when the ClientHello message
6762 is badly fragmented. This allows a man-in-the-middle attacker to force a
6763 downgrade to TLS 1.0 even if both the server and the client support a
6764 higher protocol version, by modifying the client's TLS records.
6766 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6767 researching this issue.
6772 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6773 to a denial of service attack. A malicious server can crash the client
6774 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6775 ciphersuite and sending carefully crafted handshake messages.
6777 Thanks to Felix Gröbert (Google) for discovering and researching this
6783 * By sending carefully crafted DTLS packets an attacker could cause openssl
6784 to leak memory. This can be exploited through a Denial of Service attack.
6785 Thanks to Adam Langley for discovering and researching this issue.
6790 * An attacker can force openssl to consume large amounts of memory whilst
6791 processing DTLS handshake messages. This can be exploited through a
6792 Denial of Service attack.
6793 Thanks to Adam Langley for discovering and researching this issue.
6798 * An attacker can force an error condition which causes openssl to crash
6799 whilst processing DTLS packets due to memory being freed twice. This
6800 can be exploited through a Denial of Service attack.
6801 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6807 * If a multithreaded client connects to a malicious server using a resumed
6808 session and the server sends an ec point format extension it could write
6809 up to 255 bytes to freed memory.
6811 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6817 * A malicious server can crash an OpenSSL client with a null pointer
6818 dereference (read) by specifying an SRP ciphersuite even though it was not
6819 properly negotiated with the client. This can be exploited through a
6820 Denial of Service attack.
6822 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6823 discovering and researching this issue.
6828 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6829 X509_name_oneline, X509_name_print_ex et al. to leak some information
6830 from the stack. Applications may be affected if they echo pretty printing
6831 output to the attacker.
6833 Thanks to Ivan Fratric (Google) for discovering this issue.
6836 *Emilia Käsper, and Steve Henson*
6838 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6839 for corner cases. (Certain input points at infinity could lead to
6840 bogus results, with non-infinity inputs mapped to infinity too.)
6844 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6846 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6847 handshake can force the use of weak keying material in OpenSSL
6848 SSL/TLS clients and servers.
6850 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6851 researching this issue. ([CVE-2014-0224])
6853 *KIKUCHI Masashi, Steve Henson*
6855 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6856 OpenSSL DTLS client the code can be made to recurse eventually crashing
6859 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6862 *Imre Rad, Steve Henson*
6864 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6865 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6866 client or server. This is potentially exploitable to run arbitrary
6867 code on a vulnerable client or server.
6869 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6871 *Jüri Aedla, Steve Henson*
6873 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6874 are subject to a denial of service attack.
6876 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6877 this issue. ([CVE-2014-3470])
6879 *Felix Gröbert, Ivan Fratric, Steve Henson*
6881 * Harmonize version and its documentation. -f flag is used to display
6884 *mancha <mancha1@zoho.com>*
6886 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6887 in i2d_ECPrivateKey.
6889 *mancha <mancha1@zoho.com>*
6891 * Fix some double frees. These are not thought to be exploitable.
6893 *mancha <mancha1@zoho.com>*
6895 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6897 * A missing bounds check in the handling of the TLS heartbeat extension
6898 can be used to reveal up to 64k of memory to a connected client or
6901 Thanks for Neel Mehta of Google Security for discovering this bug and to
6902 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6903 preparing the fix ([CVE-2014-0160])
6905 *Adam Langley, Bodo Moeller*
6907 * Fix for the attack described in the paper "Recovering OpenSSL
6908 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6909 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6910 <http://eprint.iacr.org/2014/140>
6912 Thanks to Yuval Yarom and Naomi Benger for discovering this
6913 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6915 *Yuval Yarom and Naomi Benger*
6917 * TLS pad extension: draft-agl-tls-padding-03
6919 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6920 TLS client Hello record length value would otherwise be > 255 and
6921 less that 512 pad with a dummy extension containing zeroes so it
6922 is at least 512 bytes long.
6924 *Adam Langley, Steve Henson*
6926 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6928 * Fix for TLS record tampering bug. A carefully crafted invalid
6929 handshake could crash OpenSSL with a NULL pointer exception.
6930 Thanks to Anton Johansson for reporting this issues.
6933 * Keep original DTLS digest and encryption contexts in retransmission
6934 structures so we can use the previous session parameters if they need
6935 to be resent. ([CVE-2013-6450])
6939 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6940 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6941 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6942 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6943 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6944 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6946 *Rob Stradling, Adam Langley*
6948 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6950 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6951 supporting platforms or when small records were transferred.
6953 *Andy Polyakov, Steve Henson*
6955 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6957 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6959 This addresses the flaw in CBC record processing discovered by
6960 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6961 at: <http://www.isg.rhul.ac.uk/tls/>
6963 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6964 Security Group at Royal Holloway, University of London
6965 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6966 Emilia Käsper for the initial patch.
6969 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6971 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6972 ciphersuites which can be exploited in a denial of service attack.
6973 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6974 and detecting this bug and to Wolfgang Ettlinger
6975 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6980 * Return an error when checking OCSP signatures when key is NULL.
6981 This fixes a DoS attack. ([CVE-2013-0166])
6985 * Make openssl verify return errors.
6987 *Chris Palmer <palmer@google.com> and Ben Laurie*
6989 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6990 the right response is stapled. Also change SSL_get_certificate()
6991 so it returns the certificate actually sent.
6992 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6994 *Rob Stradling <rob.stradling@comodo.com>*
6996 * Fix possible deadlock when decoding public keys.
7000 * Don't use TLS 1.0 record version number in initial client hello
7005 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7007 * Sanity check record length before skipping explicit IV in TLS
7008 1.2, 1.1 and DTLS to fix DoS attack.
7010 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7011 fuzzing as a service testing platform.
7016 * Initialise tkeylen properly when encrypting CMS messages.
7017 Thanks to Solar Designer of Openwall for reporting this issue.
7021 * In FIPS mode don't try to use composite ciphers as they are not
7026 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7028 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7029 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7030 mean any application compiled against OpenSSL 1.0.0 headers setting
7031 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7032 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7033 0x10000000L Any application which was previously compiled against
7034 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7035 will need to be recompiled as a result. Letting be results in
7036 inability to disable specifically TLS 1.1 and in client context,
7037 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7041 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7042 disable just protocol X, but all protocols above X *if* there are
7043 protocols *below* X still enabled. In more practical terms it means
7044 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7045 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7046 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7051 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7053 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7054 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7055 in CRYPTO_realloc_clean.
7057 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7058 issue and to Adam Langley <agl@chromium.org> for fixing it.
7061 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7063 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7067 * Workarounds for some broken servers that "hang" if a client hello
7068 record length exceeds 255 bytes.
7070 1. Do not use record version number > TLS 1.0 in initial client
7071 hello: some (but not all) hanging servers will now work.
7072 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7073 the number of ciphers sent in the client hello. This should be
7074 set to an even number, such as 50, for example by passing:
7075 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7076 Most broken servers should now work.
7077 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7078 TLS 1.2 client support entirely.
7082 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7086 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7088 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7089 STRING form instead of a DigestInfo.
7093 * The format used for MDC2 RSA signatures is inconsistent between EVP
7094 and the RSA_sign/RSA_verify functions. This was made more apparent when
7095 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7096 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7097 the correct format in RSA_verify so both forms transparently work.
7101 * Some servers which support TLS 1.0 can choke if we initially indicate
7102 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7103 encrypted premaster secret. As a workaround use the maximum permitted
7104 client version in client hello, this should keep such servers happy
7105 and still work with previous versions of OpenSSL.
7109 * Add support for TLS/DTLS heartbeats.
7111 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7113 * Add support for SCTP.
7115 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7117 * Improved PRNG seeding for VOS.
7119 *Paul Green <Paul.Green@stratus.com>*
7121 * Extensive assembler packs updates, most notably:
7123 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7124 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7125 - x86_64: bit-sliced AES implementation;
7126 - ARM: NEON support, contemporary platforms optimizations;
7127 - s390x: z196 support;
7128 - `*`: GHASH and GF(2^m) multiplication implementations;
7132 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7133 (removal of unnecessary code)
7135 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7137 * Add TLS key material exporter from RFC 5705.
7141 * Add DTLS-SRTP negotiation from RFC 5764.
7145 * Add Next Protocol Negotiation,
7146 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7147 disabled with a no-npn flag to config or Configure. Code donated
7150 *Adam Langley <agl@google.com> and Ben Laurie*
7152 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7153 NIST-P256, NIST-P521, with constant-time single point multiplication on
7154 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7155 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7156 Code made available under Apache License version 2.0.
7158 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7159 line to include this in your build of OpenSSL, and run "make depend" (or
7160 "make update"). This enables the following EC_METHODs:
7162 EC_GFp_nistp224_method()
7163 EC_GFp_nistp256_method()
7164 EC_GFp_nistp521_method()
7166 EC_GROUP_new_by_curve_name() will automatically use these (while
7167 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7170 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7172 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7173 all platforms. Move ssize_t definition from e_os.h to the public
7174 header file e_os2.h as it now appears in public header file cms.h
7178 * New -sigopt option to the ca, req and x509 utilities. Additional
7179 signature parameters can be passed using this option and in
7184 * Add RSA PSS signing function. This will generate and set the
7185 appropriate AlgorithmIdentifiers for PSS based on those in the
7186 corresponding EVP_MD_CTX structure. No application support yet.
7190 * Support for companion algorithm specific ASN1 signing routines.
7191 New function ASN1_item_sign_ctx() signs a pre-initialised
7192 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7193 the appropriate parameters.
7197 * Add new algorithm specific ASN1 verification initialisation function
7198 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7199 handling will be the same no matter what EVP_PKEY_METHOD is used.
7200 Add a PSS handler to support verification of PSS signatures: checked
7201 against a number of sample certificates.
7205 * Add signature printing for PSS. Add PSS OIDs.
7207 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7209 * Add algorithm specific signature printing. An individual ASN1 method
7210 can now print out signatures instead of the standard hex dump.
7212 More complex signatures (e.g. PSS) can print out more meaningful
7213 information. Include DSA version that prints out the signature
7218 * Password based recipient info support for CMS library: implementing
7223 * Split password based encryption into PBES2 and PBKDF2 functions. This
7224 neatly separates the code into cipher and PBE sections and is required
7225 for some algorithms that split PBES2 into separate pieces (such as
7226 password based CMS).
7230 * Session-handling fixes:
7231 - Fix handling of connections that are resuming with a session ID,
7232 but also support Session Tickets.
7233 - Fix a bug that suppressed issuing of a new ticket if the client
7234 presented a ticket with an expired session.
7235 - Try to set the ticket lifetime hint to something reasonable.
7236 - Make tickets shorter by excluding irrelevant information.
7237 - On the client side, don't ignore renewed tickets.
7239 *Adam Langley, Bodo Moeller (Google)*
7241 * Fix PSK session representation.
7245 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7247 This work was sponsored by Intel.
7251 * Add GCM support to TLS library. Some custom code is needed to split
7252 the IV between the fixed (from PRF) and explicit (from TLS record)
7253 portions. This adds all GCM ciphersuites supported by RFC5288 and
7254 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7255 add a special AESGCM string for GCM only.
7259 * Expand range of ctrls for AES GCM. Permit setting invocation
7260 field on decrypt and retrieval of invocation field only on encrypt.
7264 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7265 As required by RFC5289 these ciphersuites cannot be used if for
7266 versions of TLS earlier than 1.2.
7270 * For FIPS capable OpenSSL interpret a NULL default public key method
7271 as unset and return the appropriate default but do *not* set the default.
7272 This means we can return the appropriate method in applications that
7273 switch between FIPS and non-FIPS modes.
7277 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7278 ENGINE is used then we cannot handle that in the FIPS module so we
7279 keep original code iff non-FIPS operations are allowed.
7283 * Add -attime option to openssl utilities.
7285 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7287 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7291 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7292 FIPS EC methods unconditionally for now.
7296 * New build option no-ec2m to disable characteristic 2 code.
7300 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7301 all cases can be covered as some introduce binary incompatibilities.
7305 * Redirect RSA operations to FIPS module including keygen,
7306 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7310 * Add similar low level API blocking to ciphers.
7314 * Low level digest APIs are not approved in FIPS mode: any attempt
7315 to use these will cause a fatal error. Applications that *really* want
7316 to use them can use the `private_*` version instead.
7320 * Redirect cipher operations to FIPS module for FIPS builds.
7324 * Redirect digest operations to FIPS module for FIPS builds.
7328 * Update build system to add "fips" flag which will link in fipscanister.o
7329 for static and shared library builds embedding a signature if needed.
7333 * Output TLS supported curves in preference order instead of numerical
7334 order. This is currently hardcoded for the highest order curves first.
7335 This should be configurable so applications can judge speed vs strength.
7339 * Add TLS v1.2 server support for client authentication.
7343 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7348 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7349 FIPS modules versions.
7353 * Add TLS v1.2 client side support for client authentication. Keep cache
7354 of handshake records longer as we don't know the hash algorithm to use
7355 until after the certificate request message is received.
7359 * Initial TLS v1.2 client support. Add a default signature algorithms
7360 extension including all the algorithms we support. Parse new signature
7361 format in client key exchange. Relax some ECC signing restrictions for
7362 TLS v1.2 as indicated in RFC5246.
7366 * Add server support for TLS v1.2 signature algorithms extension. Switch
7367 to new signature format when needed using client digest preference.
7368 All server ciphersuites should now work correctly in TLS v1.2. No client
7369 support yet and no support for client certificates.
7373 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7374 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7375 ciphersuites. At present only RSA key exchange ciphersuites work with
7376 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7377 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7378 and version checking.
7382 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7383 with this defined it will not be affected by any changes to ssl internal
7384 structures. Add several utility functions to allow openssl application
7385 to work with OPENSSL_NO_SSL_INTERN defined.
7389 * A long standing patch to add support for SRP from EdelWeb (Peter
7390 Sylvester and Christophe Renou) was integrated.
7391 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7392 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7395 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7399 * Permit abbreviated handshakes when renegotiating using the function
7400 SSL_renegotiate_abbreviated().
7402 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7404 * Add call to ENGINE_register_all_complete() to
7405 ENGINE_load_builtin_engines(), so some implementations get used
7406 automatically instead of needing explicit application support.
7410 * Add support for TLS key exporter as described in RFC5705.
7412 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7414 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7415 a few changes are required:
7417 Add SSL_OP_NO_TLSv1_1 flag.
7418 Add TLSv1_1 methods.
7419 Update version checking logic to handle version 1.1.
7420 Add explicit IV handling (ported from DTLS code).
7421 Add command line options to s_client/s_server.
7428 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7430 * X509_ATTRIBUTE memory leak
7432 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7433 memory. This structure is used by the PKCS#7 and CMS routines so any
7434 application which reads PKCS#7 or CMS data from untrusted sources is
7435 affected. SSL/TLS is not affected.
7437 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7443 * Race condition handling PSK identify hint
7445 If PSK identity hints are received by a multi-threaded client then
7446 the values are wrongly updated in the parent SSL_CTX structure. This can
7447 result in a race condition potentially leading to a double free of the
7453 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7455 * Malformed ECParameters causes infinite loop
7457 When processing an ECParameters structure OpenSSL enters an infinite loop
7458 if the curve specified is over a specially malformed binary polynomial
7461 This can be used to perform denial of service against any
7462 system which processes public keys, certificate requests or
7463 certificates. This includes TLS clients and TLS servers with
7464 client authentication enabled.
7466 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7471 * Exploitable out-of-bounds read in X509_cmp_time
7473 X509_cmp_time does not properly check the length of the ASN1_TIME
7474 string and can read a few bytes out of bounds. In addition,
7475 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7478 An attacker can use this to craft malformed certificates and CRLs of
7479 various sizes and potentially cause a segmentation fault, resulting in
7480 a DoS on applications that verify certificates or CRLs. TLS clients
7481 that verify CRLs are affected. TLS clients and servers with client
7482 authentication enabled may be affected if they use custom verification
7485 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7486 independently by Hanno Böck.
7491 * PKCS7 crash with missing EnvelopedContent
7493 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7494 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7495 with missing content and trigger a NULL pointer dereference on parsing.
7497 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7498 structures from untrusted sources are affected. OpenSSL clients and
7499 servers are not affected.
7501 This issue was reported to OpenSSL by Michal Zalewski (Google).
7506 * CMS verify infinite loop with unknown hash function
7508 When verifying a signedData message the CMS code can enter an infinite loop
7509 if presented with an unknown hash function OID. This can be used to perform
7510 denial of service against any system which verifies signedData messages using
7512 This issue was reported to OpenSSL by Johannes Bauer.
7517 * Race condition handling NewSessionTicket
7519 If a NewSessionTicket is received by a multi-threaded client when attempting to
7520 reuse a previous ticket then a race condition can occur potentially leading to
7521 a double free of the ticket data.
7526 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7528 * Segmentation fault in ASN1_TYPE_cmp fix
7530 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7531 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7532 certificate signature algorithm consistency this can be used to crash any
7533 certificate verification operation and exploited in a DoS attack. Any
7534 application which performs certificate verification is vulnerable including
7535 OpenSSL clients and servers which enable client authentication.
7540 * ASN.1 structure reuse memory corruption fix
7542 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7543 memory corruption via an invalid write. Such reuse is and has been
7544 strongly discouraged and is believed to be rare.
7546 Applications that parse structures containing CHOICE or ANY DEFINED BY
7547 components may be affected. Certificate parsing (d2i_X509 and related
7548 functions) are however not affected. OpenSSL clients and servers are
7554 * PKCS7 NULL pointer dereferences fix
7556 The PKCS#7 parsing code does not handle missing outer ContentInfo
7557 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7558 missing content and trigger a NULL pointer dereference on parsing.
7560 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7561 otherwise parse PKCS#7 structures from untrusted sources are
7562 affected. OpenSSL clients and servers are not affected.
7564 This issue was reported to OpenSSL by Michal Zalewski (Google).
7569 * DoS via reachable assert in SSLv2 servers fix
7571 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7572 servers that both support SSLv2 and enable export cipher suites by sending
7573 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7575 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7576 (OpenSSL development team).
7581 * Use After Free following d2i_ECPrivatekey error fix
7583 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7584 could cause a use after free condition. This, in turn, could cause a double
7585 free in several private key parsing functions (such as d2i_PrivateKey
7586 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7587 for applications that receive EC private keys from untrusted
7588 sources. This scenario is considered rare.
7590 This issue was discovered by the BoringSSL project and fixed in their
7596 * X509_to_X509_REQ NULL pointer deref fix
7598 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7599 the certificate key is invalid. This function is rarely used in practice.
7601 This issue was discovered by Brian Carpenter.
7606 * Removed the export ciphers from the DEFAULT ciphers
7610 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7612 * Build fixes for the Windows and OpenVMS platforms
7614 *Matt Caswell and Richard Levitte*
7616 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7618 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7619 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7620 dereference. This could lead to a Denial Of Service attack. Thanks to
7621 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7626 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7627 dtls1_buffer_record function under certain conditions. In particular this
7628 could occur if an attacker sent repeated DTLS records with the same
7629 sequence number but for the next epoch. The memory leak could be exploited
7630 by an attacker in a Denial of Service attack through memory exhaustion.
7631 Thanks to Chris Mueller for reporting this issue.
7636 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7637 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7638 method would be set to NULL which could later result in a NULL pointer
7639 dereference. Thanks to Frank Schmirler for reporting this issue.
7644 * Abort handshake if server key exchange message is omitted for ephemeral
7647 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7648 reporting this issue.
7653 * Remove non-export ephemeral RSA code on client and server. This code
7654 violated the TLS standard by allowing the use of temporary RSA keys in
7655 non-export ciphersuites and could be used by a server to effectively
7656 downgrade the RSA key length used to a value smaller than the server
7657 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7658 INRIA or reporting this issue.
7663 * Fixed issue where DH client certificates are accepted without verification.
7664 An OpenSSL server will accept a DH certificate for client authentication
7665 without the certificate verify message. This effectively allows a client to
7666 authenticate without the use of a private key. This only affects servers
7667 which trust a client certificate authority which issues certificates
7668 containing DH keys: these are extremely rare and hardly ever encountered.
7669 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7675 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7676 results on some platforms, including x86_64. This bug occurs at random
7677 with a very low probability, and is not known to be exploitable in any
7678 way, though its exact impact is difficult to determine. Thanks to Pieter
7679 Wuille (Blockstream) who reported this issue and also suggested an initial
7680 fix. Further analysis was conducted by the OpenSSL development team and
7681 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7682 the OpenSSL core team.
7687 * Fix various certificate fingerprint issues.
7689 By using non-DER or invalid encodings outside the signed portion of a
7690 certificate the fingerprint can be changed without breaking the signature.
7691 Although no details of the signed portion of the certificate can be changed
7692 this can cause problems with some applications: e.g. those using the
7693 certificate fingerprint for blacklists.
7695 1. Reject signatures with non zero unused bits.
7697 If the BIT STRING containing the signature has non zero unused bits reject
7698 the signature. All current signature algorithms require zero unused bits.
7700 2. Check certificate algorithm consistency.
7702 Check the AlgorithmIdentifier inside TBS matches the one in the
7703 certificate signature. NB: this will result in signature failure
7704 errors for some broken certificates.
7706 Thanks to Konrad Kraszewski from Google for reporting this issue.
7708 3. Check DSA/ECDSA signatures use DER.
7710 Reencode DSA/ECDSA signatures and compare with the original received
7711 signature. Return an error if there is a mismatch.
7713 This will reject various cases including garbage after signature
7714 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7715 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7716 (negative or with leading zeroes).
7718 Further analysis was conducted and fixes were developed by Stephen Henson
7719 of the OpenSSL core team.
7725 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7727 * Session Ticket Memory Leak.
7729 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7730 integrity of that ticket is first verified. In the event of a session
7731 ticket integrity check failing, OpenSSL will fail to free memory
7732 causing a memory leak. By sending a large number of invalid session
7733 tickets an attacker could exploit this issue in a Denial Of Service
7739 * Build option no-ssl3 is incomplete.
7741 When OpenSSL is configured with "no-ssl3" as a build option, servers
7742 could accept and complete a SSL 3.0 handshake, and clients could be
7743 configured to send them.
7746 *Akamai and the OpenSSL team*
7748 * Add support for TLS_FALLBACK_SCSV.
7749 Client applications doing fallback retries should call
7750 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7753 *Adam Langley, Bodo Moeller*
7755 * Add additional DigestInfo checks.
7757 Reencode DigestInto in DER and check against the original when
7758 verifying RSA signature: this will reject any improperly encoded
7759 DigestInfo structures.
7761 Note: this is a precautionary measure and no attacks are currently known.
7765 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7767 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7768 to a denial of service attack. A malicious server can crash the client
7769 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7770 ciphersuite and sending carefully crafted handshake messages.
7772 Thanks to Felix Gröbert (Google) for discovering and researching this
7778 * By sending carefully crafted DTLS packets an attacker could cause openssl
7779 to leak memory. This can be exploited through a Denial of Service attack.
7780 Thanks to Adam Langley for discovering and researching this issue.
7785 * An attacker can force openssl to consume large amounts of memory whilst
7786 processing DTLS handshake messages. This can be exploited through a
7787 Denial of Service attack.
7788 Thanks to Adam Langley for discovering and researching this issue.
7793 * An attacker can force an error condition which causes openssl to crash
7794 whilst processing DTLS packets due to memory being freed twice. This
7795 can be exploited through a Denial of Service attack.
7796 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7802 * If a multithreaded client connects to a malicious server using a resumed
7803 session and the server sends an ec point format extension it could write
7804 up to 255 bytes to freed memory.
7806 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7812 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7813 X509_name_oneline, X509_name_print_ex et al. to leak some information
7814 from the stack. Applications may be affected if they echo pretty printing
7815 output to the attacker.
7817 Thanks to Ivan Fratric (Google) for discovering this issue.
7820 *Emilia Käsper, and Steve Henson*
7822 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7823 for corner cases. (Certain input points at infinity could lead to
7824 bogus results, with non-infinity inputs mapped to infinity too.)
7828 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7830 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7831 handshake can force the use of weak keying material in OpenSSL
7832 SSL/TLS clients and servers.
7834 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7835 researching this issue. ([CVE-2014-0224])
7837 *KIKUCHI Masashi, Steve Henson*
7839 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7840 OpenSSL DTLS client the code can be made to recurse eventually crashing
7843 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7846 *Imre Rad, Steve Henson*
7848 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7849 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7850 client or server. This is potentially exploitable to run arbitrary
7851 code on a vulnerable client or server.
7853 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7855 *Jüri Aedla, Steve Henson*
7857 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7858 are subject to a denial of service attack.
7860 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7861 this issue. ([CVE-2014-3470])
7863 *Felix Gröbert, Ivan Fratric, Steve Henson*
7865 * Harmonize version and its documentation. -f flag is used to display
7868 *mancha <mancha1@zoho.com>*
7870 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7871 in i2d_ECPrivateKey.
7873 *mancha <mancha1@zoho.com>*
7875 * Fix some double frees. These are not thought to be exploitable.
7877 *mancha <mancha1@zoho.com>*
7879 * Fix for the attack described in the paper "Recovering OpenSSL
7880 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7881 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7882 <http://eprint.iacr.org/2014/140>
7884 Thanks to Yuval Yarom and Naomi Benger for discovering this
7885 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7887 *Yuval Yarom and Naomi Benger*
7889 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7891 * Keep original DTLS digest and encryption contexts in retransmission
7892 structures so we can use the previous session parameters if they need
7893 to be resent. ([CVE-2013-6450])
7897 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7898 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7899 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7900 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7901 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7902 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7904 *Rob Stradling, Adam Langley*
7906 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7908 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7910 This addresses the flaw in CBC record processing discovered by
7911 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7912 at: <http://www.isg.rhul.ac.uk/tls/>
7914 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7915 Security Group at Royal Holloway, University of London
7916 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7917 Emilia Käsper for the initial patch.
7920 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7922 * Return an error when checking OCSP signatures when key is NULL.
7923 This fixes a DoS attack. ([CVE-2013-0166])
7927 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7928 the right response is stapled. Also change SSL_get_certificate()
7929 so it returns the certificate actually sent.
7930 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7931 (This is a backport)
7933 *Rob Stradling <rob.stradling@comodo.com>*
7935 * Fix possible deadlock when decoding public keys.
7939 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7941 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7944 * Sanity check record length before skipping explicit IV in DTLS
7947 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7948 fuzzing as a service testing platform.
7953 * Initialise tkeylen properly when encrypting CMS messages.
7954 Thanks to Solar Designer of Openwall for reporting this issue.
7958 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7960 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7961 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7962 in CRYPTO_realloc_clean.
7964 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7965 issue and to Adam Langley <agl@chromium.org> for fixing it.
7968 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7970 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7972 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7973 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7974 content decryption and always return the same error. Note: this attack
7975 needs on average 2^20 messages so it only affects automated senders. The
7976 old behaviour can be re-enabled in the CMS code by setting the
7977 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7978 an MMA defence is not necessary.
7979 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7980 this issue. ([CVE-2012-0884])
7984 * Fix CVE-2011-4619: make sure we really are receiving a
7985 client hello before rejecting multiple SGC restarts. Thanks to
7986 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7990 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7992 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7993 Thanks to Antonio Martin, Enterprise Secure Access Research and
7994 Development, Cisco Systems, Inc. for discovering this bug and
7995 preparing a fix. ([CVE-2012-0050])
7999 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8001 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8002 of the Vaudenay padding oracle attack on CBC mode encryption
8003 which enables an efficient plaintext recovery attack against
8004 the OpenSSL implementation of DTLS. Their attack exploits timing
8005 differences arising during decryption processing. A research
8006 paper describing this attack can be found at:
8007 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8008 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8009 Security Group at Royal Holloway, University of London
8010 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8011 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8012 for preparing the fix. ([CVE-2011-4108])
8014 *Robin Seggelmann, Michael Tuexen*
8016 * Clear bytes used for block padding of SSL 3.0 records.
8019 *Adam Langley (Google)*
8021 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8022 Kadianakis <desnacked@gmail.com> for discovering this issue and
8023 Adam Langley for preparing the fix. ([CVE-2011-4619])
8025 *Adam Langley (Google)*
8027 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8029 *Andrey Kulikov <amdeich@gmail.com>*
8031 * Prevent malformed RFC3779 data triggering an assertion failure.
8032 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8033 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8035 *Rob Austein <sra@hactrn.net>*
8037 * Improved PRNG seeding for VOS.
8039 *Paul Green <Paul.Green@stratus.com>*
8041 * Fix ssl_ciph.c set-up race.
8043 *Adam Langley (Google)*
8045 * Fix spurious failures in ecdsatest.c.
8047 *Emilia Käsper (Google)*
8049 * Fix the BIO_f_buffer() implementation (which was mixing different
8050 interpretations of the `..._len` fields).
8052 *Adam Langley (Google)*
8054 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8055 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8056 threads won't reuse the same blinding coefficients.
8058 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8059 lock to call BN_BLINDING_invert_ex, and avoids one use of
8060 BN_BLINDING_update for each BN_BLINDING structure (previously,
8061 the last update always remained unused).
8063 *Emilia Käsper (Google)*
8065 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8067 *Bob Buckholz (Google)*
8069 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8071 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8072 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8074 *Kaspar Brand <ossl@velox.ch>*
8076 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8077 for multi-threaded use of ECDH. ([CVE-2011-3210])
8079 *Adam Langley (Google)*
8081 * Fix x509_name_ex_d2i memory leak on bad inputs.
8085 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8086 signature public key algorithm by using OID xref utilities instead.
8087 Before this you could only use some ECC ciphersuites with SHA1 only.
8091 * Add protection against ECDSA timing attacks as mentioned in the paper
8092 by Billy Bob Brumley and Nicola Tuveri, see:
8093 <http://eprint.iacr.org/2011/232.pdf>
8095 *Billy Bob Brumley and Nicola Tuveri*
8097 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8099 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8101 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8103 * Fix bug in string printing code: if *any* escaping is enabled we must
8104 escape the escape character (backslash) or the resulting string is
8109 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8111 * Disable code workaround for ancient and obsolete Netscape browsers
8112 and servers: an attacker can use it in a ciphersuite downgrade attack.
8113 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8117 * Fixed J-PAKE implementation error, originally discovered by
8118 Sebastien Martini, further info and confirmation from Stefan
8119 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8123 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8125 * Fix extension code to avoid race conditions which can result in a buffer
8126 overrun vulnerability: resumed sessions must not be modified as they can
8127 be shared by multiple threads. CVE-2010-3864
8131 * Fix WIN32 build system to correctly link an ENGINE directory into
8136 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8138 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8141 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8143 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8145 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8146 context. The operation can be customised via the ctrl mechanism in
8147 case ENGINEs want to include additional functionality.
8151 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8155 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8156 output hashes compatible with older versions of OpenSSL.
8158 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8160 * Fix compression algorithm handling: if resuming a session use the
8161 compression algorithm of the resumed session instead of determining
8162 it from client hello again. Don't allow server to change algorithm.
8166 * Add load_crls() function to commands tidying load_certs() too. Add option
8167 to verify utility to allow additional CRLs to be included.
8171 * Update OCSP request code to permit adding custom headers to the request:
8172 some responders need this.
8176 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8179 *Julia Lawall <julia@diku.dk>*
8181 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8182 needlessly dereferenced structures, used obsolete functions and
8183 didn't handle all updated verify codes correctly.
8187 * Disable MD2 in the default configuration.
8191 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8192 indicate the initial BIO being pushed or popped. This makes it possible
8193 to determine whether the BIO is the one explicitly called or as a result
8194 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8195 it handles reference counts correctly and doesn't zero out the I/O bio
8196 when it is not being explicitly popped. WARNING: applications which
8197 included workarounds for the old buggy behaviour will need to be modified
8198 or they could free up already freed BIOs.
8202 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8203 renaming to all platforms (within the 0.9.8 branch, this was
8204 done conditionally on Netware platforms to avoid a name clash).
8206 *Guenter <lists@gknw.net>*
8208 * Add ECDHE and PSK support to DTLS.
8210 *Michael Tuexen <tuexen@fh-muenster.de>*
8212 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8217 * Add "missing" function EVP_MD_flags() (without this the only way to
8218 retrieve a digest flags is by accessing the structure directly. Update
8219 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8220 or cipher is registered as in the "from" argument. Print out all
8221 registered digests in the dgst usage message instead of manually
8222 attempting to work them out.
8226 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8227 this allows the use of compression and extensions. Change default cipher
8228 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8229 by default unless an application cipher string requests it.
8233 * Alter match criteria in PKCS12_parse(). It used to try to use local
8234 key ids to find matching certificates and keys but some PKCS#12 files
8235 don't follow the (somewhat unwritten) rules and this strategy fails.
8236 Now just gather all certificates together and the first private key
8237 then look for the first certificate that matches the key.
8241 * Support use of registered digest and cipher names for dgst and cipher
8242 commands instead of having to add each one as a special case. So now
8249 openssl dgst -sha256 foo
8251 and this works for ENGINE based algorithms too.
8255 * Update Gost ENGINE to support parameter files.
8257 *Victor B. Wagner <vitus@cryptocom.ru>*
8259 * Support GeneralizedTime in ca utility.
8261 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8263 * Enhance the hash format used for certificate directory links. The new
8264 form uses the canonical encoding (meaning equivalent names will work
8265 even if they aren't identical) and uses SHA1 instead of MD5. This form
8266 is incompatible with the older format and as a result c_rehash should
8267 be used to rebuild symbolic links.
8271 * Make PKCS#8 the default write format for private keys, replacing the
8272 traditional format. This form is standardised, more secure and doesn't
8273 include an implicit MD5 dependency.
8277 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8278 committed to OpenSSL should pass this lot as a minimum.
8282 * Add session ticket override functionality for use by EAP-FAST.
8284 *Jouni Malinen <j@w1.fi>*
8286 * Modify HMAC functions to return a value. Since these can be implemented
8287 in an ENGINE errors can occur.
8291 * Type-checked OBJ_bsearch_ex.
8295 * Type-checked OBJ_bsearch. Also some constification necessitated
8296 by type-checking. Still to come: TXT_DB, bsearch(?),
8297 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8302 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8303 seconds to a tm structure directly, instead of going through OS
8304 specific date routines. This avoids any issues with OS routines such
8305 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8306 and X509_time_adj_ex() to cover the extended range. The existing
8307 X509_time_adj() is still usable and will no longer have any date issues.
8311 * Delta CRL support. New use deltas option which will attempt to locate
8312 and search any appropriate delta CRLs available.
8314 This work was sponsored by Google.
8318 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8319 code and add additional score elements. Validate alternate CRL paths
8320 as part of the CRL checking and indicate a new error "CRL path validation
8321 error" in this case. Applications wanting additional details can use
8322 the verify callback and check the new "parent" field. If this is not
8323 NULL CRL path validation is taking place. Existing applications won't
8324 see this because it requires extended CRL support which is off by
8327 This work was sponsored by Google.
8331 * Support for freshest CRL extension.
8333 This work was sponsored by Google.
8337 * Initial indirect CRL support. Currently only supported in the CRLs
8338 passed directly and not via lookup. Process certificate issuer
8339 CRL entry extension and lookup CRL entries by bother issuer name
8340 and serial number. Check and process CRL issuer entry in IDP extension.
8342 This work was sponsored by Google.
8346 * Add support for distinct certificate and CRL paths. The CRL issuer
8347 certificate is validated separately in this case. Only enabled if
8348 an extended CRL support flag is set: this flag will enable additional
8349 CRL functionality in future.
8351 This work was sponsored by Google.
8355 * Add support for policy mappings extension.
8357 This work was sponsored by Google.
8361 * Fixes to pathlength constraint, self issued certificate handling,
8362 policy processing to align with RFC3280 and PKITS tests.
8364 This work was sponsored by Google.
8368 * Support for name constraints certificate extension. DN, email, DNS
8369 and URI types are currently supported.
8371 This work was sponsored by Google.
8375 * To cater for systems that provide a pointer-based thread ID rather
8376 than numeric, deprecate the current numeric thread ID mechanism and
8377 replace it with a structure and associated callback type. This
8378 mechanism allows a numeric "hash" to be extracted from a thread ID in
8379 either case, and on platforms where pointers are larger than 'long',
8380 mixing is done to help ensure the numeric 'hash' is usable even if it
8381 can't be guaranteed unique. The default mechanism is to use "&errno"
8382 as a pointer-based thread ID to distinguish between threads.
8384 Applications that want to provide their own thread IDs should now use
8385 CRYPTO_THREADID_set_callback() to register a callback that will call
8386 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8388 Note that ERR_remove_state() is now deprecated, because it is tied
8389 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8390 to free the current thread's error state should be replaced by
8391 ERR_remove_thread_state(NULL).
8393 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8394 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8395 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8396 application was previously providing a numeric thread callback that
8397 was inappropriate for distinguishing threads, then uniqueness might
8398 have been obtained with &errno that happened immediately in the
8399 intermediate development versions of OpenSSL; this is no longer the
8400 case, the numeric thread callback will now override the automatic use
8403 *Geoff Thorpe, with help from Bodo Moeller*
8405 * Initial support for different CRL issuing certificates. This covers a
8406 simple case where the self issued certificates in the chain exist and
8407 the real CRL issuer is higher in the existing chain.
8409 This work was sponsored by Google.
8413 * Removed effectively defunct crypto/store from the build.
8417 * Revamp of STACK to provide stronger type-checking. Still to come:
8418 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8419 ASN1_STRING, CONF_VALUE.
8423 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8424 RAM on SSL connections. This option can save about 34k per idle SSL.
8428 * Revamp of LHASH to provide stronger type-checking. Still to come:
8429 STACK, TXT_DB, bsearch, qsort.
8433 * Initial support for Cryptographic Message Syntax (aka CMS) based
8434 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8435 support for data, signedData, compressedData, digestedData and
8436 encryptedData, envelopedData types included. Scripts to check against
8437 RFC4134 examples draft and interop and consistency checks of many
8438 content types and variants.
8442 * Add options to enc utility to support use of zlib compression BIO.
8446 * Extend mk1mf to support importing of options and assembly language
8447 files from Configure script, currently only included in VC-WIN32.
8448 The assembly language rules can now optionally generate the source
8449 files from the associated perl scripts.
8453 * Implement remaining functionality needed to support GOST ciphersuites.
8454 Interop testing has been performed using CryptoPro implementations.
8456 *Victor B. Wagner <vitus@cryptocom.ru>*
8458 * s390x assembler pack.
8462 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8467 * Implement Opaque PRF Input TLS extension as specified in
8468 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8469 official specification yet and no extension type assignment by
8470 IANA exists, this extension (for now) will have to be explicitly
8471 enabled when building OpenSSL by providing the extension number
8472 to use. For example, specify an option
8474 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8476 to the "config" or "Configure" script to enable the extension,
8477 assuming extension number 0x9527 (which is a completely arbitrary
8478 and unofficial assignment based on the MD5 hash of the Internet
8479 Draft). Note that by doing so, you potentially lose
8480 interoperability with other TLS implementations since these might
8481 be using the same extension number for other purposes.
8483 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8484 opaque PRF input value to use in the handshake. This will create
8485 an internal copy of the length-'len' string at 'src', and will
8486 return non-zero for success.
8488 To get more control and flexibility, provide a callback function
8491 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8492 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8496 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8499 Callback function 'cb' will be called in handshakes, and is
8500 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8501 Argument 'arg' is for application purposes (the value as given to
8502 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8503 be provided to the callback function). The callback function
8504 has to return non-zero to report success: usually 1 to use opaque
8505 PRF input just if possible, or 2 to enforce use of the opaque PRF
8506 input. In the latter case, the library will abort the handshake
8507 if opaque PRF input is not successfully negotiated.
8509 Arguments 'peerinput' and 'len' given to the callback function
8510 will always be NULL and 0 in the case of a client. A server will
8511 see the client's opaque PRF input through these variables if
8512 available (NULL and 0 otherwise). Note that if the server
8513 provides an opaque PRF input, the length must be the same as the
8514 length of the client's opaque PRF input.
8516 Note that the callback function will only be called when creating
8517 a new session (session resumption can resume whatever was
8518 previously negotiated), and will not be called in SSL 2.0
8519 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8520 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8521 for applications that need to enforce opaque PRF input.
8525 * Update ssl code to support digests other than SHA1+MD5 for handshake
8528 *Victor B. Wagner <vitus@cryptocom.ru>*
8530 * Add RFC4507 support to OpenSSL. This includes the corrections in
8531 RFC4507bis. The encrypted ticket format is an encrypted encoded
8532 SSL_SESSION structure, that way new session features are automatically
8535 If a client application caches session in an SSL_SESSION structure
8536 support is transparent because tickets are now stored in the encoded
8539 The SSL_CTX structure automatically generates keys for ticket
8540 protection in servers so again support should be possible
8541 with no application modification.
8543 If a client or server wishes to disable RFC4507 support then the option
8544 SSL_OP_NO_TICKET can be set.
8546 Add a TLS extension debugging callback to allow the contents of any client
8547 or server extensions to be examined.
8549 This work was sponsored by Google.
8553 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8554 OpenSSL should now compile cleanly on gcc 4.2
8556 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8558 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8559 support including streaming MAC support: this is required for GOST
8560 ciphersuite support.
8562 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8564 * Add option -stream to use PKCS#7 streaming in smime utility. New
8565 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8566 to output in BER and PEM format.
8570 * Experimental support for use of HMAC via EVP_PKEY interface. This
8571 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8572 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8573 ENGINE support for HMAC keys which are unextractable. New -mac and
8574 -macopt options to dgst utility.
8578 * New option -sigopt to dgst utility. Update dgst to use
8579 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8580 alternative signing parameters such as X9.31 or PSS in the dgst
8585 * Change ssl_cipher_apply_rule(), the internal function that does
8586 the work each time a ciphersuite string requests enabling
8587 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8588 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8589 the order of disabled ciphersuites such that those ciphersuites
8590 that most recently went from enabled to disabled not only stay
8591 in order with respect to each other, but also have higher priority
8592 than other disabled ciphersuites the next time ciphersuites are
8595 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8596 the same ciphersuites as with "HIGH" alone, but in a specific
8597 order where the PSK ciphersuites come first (since they are the
8598 most recently disabled ciphersuites when "HIGH" is parsed).
8600 Also, change ssl_create_cipher_list() (using this new
8601 functionality) such that between otherwise identical
8602 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8607 * Change ssl_create_cipher_list() so that it automatically
8608 arranges the ciphersuites in reasonable order before starting
8609 to process the rule string. Thus, the definition for "DEFAULT"
8610 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8611 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8612 This makes it much easier to arrive at a reasonable default order
8613 in applications for which anonymous ciphers are OK (meaning
8614 that you can't actually use DEFAULT).
8616 *Bodo Moeller; suggested by Victor Duchovni*
8618 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8619 processing) into multiple integers instead of setting
8620 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8621 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8622 (These masks as well as the individual bit definitions are hidden
8623 away into the non-exported interface ssl/ssl_locl.h, so this
8624 change to the definition of the SSL_CIPHER structure shouldn't
8625 affect applications.) This give us more bits for each of these
8626 categories, so there is no longer a need to coagulate AES128 and
8627 AES256 into a single algorithm bit, and to coagulate Camellia128
8628 and Camellia256 into a single algorithm bit, which has led to all
8631 Thus, among other things, the kludge introduced in 0.9.7m and
8632 0.9.8e for masking out AES256 independently of AES128 or masking
8633 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8635 With the change, we also introduce new ciphersuite aliases that
8636 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8641 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8642 Use the leftmost N bytes of the signature input if the input is
8643 larger than the prime q (with N being the size in bytes of q).
8647 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8648 it yet and it is largely untested.
8652 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8656 * Initial incomplete changes to avoid need for function casts in OpenSSL
8657 some compilers (gcc 4.2 and later) reject their use. Safestack is
8658 reimplemented. Update ASN1 to avoid use of legacy functions.
8662 * Win32/64 targets are linked with Winsock2.
8666 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8667 to external functions. This can be used to increase CRL handling
8668 efficiency especially when CRLs are very large by (for example) storing
8669 the CRL revoked certificates in a database.
8673 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8674 new CRLs added to a directory can be used. New command line option
8675 -verify_return_error to s_client and s_server. This causes real errors
8676 to be returned by the verify callback instead of carrying on no matter
8677 what. This reflects the way a "real world" verify callback would behave.
8681 * GOST engine, supporting several GOST algorithms and public key formats.
8682 Kindly donated by Cryptocom.
8686 * Partial support for Issuing Distribution Point CRL extension. CRLs
8687 partitioned by DP are handled but no indirect CRL or reason partitioning
8688 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8689 selected via a scoring technique which handles IDP and AKID in CRLs.
8693 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8694 will ultimately be used for all verify operations: this will remove the
8695 X509_STORE dependency on certificate verification and allow alternative
8696 lookup methods. X509_STORE based implementations of these two callbacks.
8700 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8701 Modify get_crl() to find a valid (unexpired) CRL if possible.
8705 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8706 this would be called X509_CRL_cmp() but that name is already used by
8707 a function that just compares CRL issuer names. Cache several CRL
8708 extensions in X509_CRL structure and cache CRLDP in X509.
8712 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8713 this maps equivalent X509_NAME structures into a consistent structure.
8714 Name comparison can then be performed rapidly using memcmp().
8718 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8723 * Allow digests to supply their own micalg string for S/MIME type using
8724 the ctrl EVP_MD_CTRL_MICALG.
8728 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8729 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8730 ctrl. It can then customise the structure before and/or after signing
8735 * New function OBJ_add_sigid() to allow application defined signature OIDs
8736 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8737 to free up any added signature OIDs.
8741 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8742 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8743 digest and cipher tables. New options added to openssl utility:
8744 list-message-digest-algorithms and list-cipher-algorithms.
8748 * Change the array representation of binary polynomials: the list
8749 of degrees of non-zero coefficients is now terminated with -1.
8750 Previously it was terminated with 0, which was also part of the
8751 value; thus, the array representation was not applicable to
8752 polynomials where t^0 has coefficient zero. This change makes
8753 the array representation useful in a more general context.
8757 * Various modifications and fixes to SSL/TLS cipher string
8758 handling. For ECC, the code now distinguishes between fixed ECDH
8759 with RSA certificates on the one hand and with ECDSA certificates
8760 on the other hand, since these are separate ciphersuites. The
8761 unused code for Fortezza ciphersuites has been removed.
8763 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8764 (not "ECDHE"). For consistency with the code for DH
8765 certificates, use of ECDH certificates is now considered ECDH
8766 authentication, not RSA or ECDSA authentication (the latter is
8767 merely the CA's signing algorithm and not actively used in the
8770 The temporary ciphersuite alias "ECCdraft" is no longer
8771 available, and ECC ciphersuites are no longer excluded from "ALL"
8772 and "DEFAULT". The following aliases now exist for RFC 4492
8773 ciphersuites, most of these by analogy with the DH case:
8775 kECDHr - ECDH cert, signed with RSA
8776 kECDHe - ECDH cert, signed with ECDSA
8777 kECDH - ECDH cert (signed with either RSA or ECDSA)
8778 kEECDH - ephemeral ECDH
8779 ECDH - ECDH cert or ephemeral ECDH
8785 AECDH - anonymous ECDH
8786 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8790 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8791 Use correct micalg parameters depending on digest(s) in signed message.
8795 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8796 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8800 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8801 an engine to register a method. Add ENGINE lookups for methods and
8802 functional reference processing.
8806 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8807 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8812 * New -resign option to smime utility. This adds one or more signers
8813 to an existing PKCS#7 signedData structure. Also -md option to use an
8814 alternative message digest algorithm for signing.
8818 * Tidy up PKCS#7 routines and add new functions to make it easier to
8819 create PKCS7 structures containing multiple signers. Update smime
8820 application to support multiple signers.
8824 * New -macalg option to pkcs12 utility to allow setting of an alternative
8829 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8830 Reorganize PBE internals to lookup from a static table using NIDs,
8831 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8832 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8833 PRF which will be automatically used with PBES2.
8837 * Replace the algorithm specific calls to generate keys in "req" with the
8842 * Update PKCS#7 enveloped data routines to use new API. This is now
8843 supported by any public key method supporting the encrypt operation. A
8844 ctrl is added to allow the public key algorithm to examine or modify
8845 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8850 * Add a ctrl to asn1 method to allow a public key algorithm to express
8851 a default digest type to use. In most cases this will be SHA1 but some
8852 algorithms (such as GOST) need to specify an alternative digest. The
8853 return value indicates how strong the preference is 1 means optional and
8854 2 is mandatory (that is it is the only supported type). Modify
8855 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8856 use the default md. Update openssl utilities to use the default digest
8857 type for signing if it is not explicitly indicated.
8861 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8862 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8863 signing method from the key type. This effectively removes the link
8864 between digests and public key types.
8868 * Add an OID cross reference table and utility functions. Its purpose is to
8869 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8870 rsaEncryption. This will allow some of the algorithm specific hackery
8871 needed to use the correct OID to be removed.
8875 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8876 structures for PKCS7_sign(). They are now set up by the relevant public
8881 * Add provisional EC pkey method with support for ECDSA and ECDH.
8885 * Add support for key derivation (agreement) in the API, DH method and
8890 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8891 public and private key formats. As a side effect these add additional
8892 command line functionality not previously available: DSA signatures can be
8893 generated and verified using pkeyutl and DH key support and generation in
8900 *Oliver Tappe <zooey@hirschkaefer.de>*
8902 * New make target "install_html_docs" installs HTML renditions of the
8905 *Oliver Tappe <zooey@hirschkaefer.de>*
8907 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8908 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8909 support key and parameter generation and add initial key generation
8910 functionality for RSA.
8914 * Add functions for main EVP_PKEY_method operations. The undocumented
8915 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8916 `EVP_PKEY_{encrypt,decrypt}_old`.
8920 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8921 key API, doesn't do much yet.
8925 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8926 public key algorithms. New option to openssl utility:
8927 "list-public-key-algorithms" to print out info.
8931 * Implement the Supported Elliptic Curves Extension for
8932 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8936 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8937 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8941 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8942 utilities such as rsa, dsa, dsaparam etc except they process any key
8947 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8948 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8949 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8954 * Initial support for pluggable public key ASN1.
8955 De-spaghettify the public key ASN1 handling. Move public and private
8956 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8957 algorithm specific handling to a single module within the relevant
8958 algorithm directory. Add functions to allow (near) opaque processing
8959 of public and private key structures.
8963 * Implement the Supported Point Formats Extension for
8964 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8968 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8969 for the psk identity [hint] and the psk callback functions to the
8970 SSL_SESSION, SSL and SSL_CTX structure.
8973 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8977 SSL_CTX_use_psk_identity_hint
8978 SSL_get_psk_identity_hint
8979 SSL_get_psk_identity
8980 SSL_use_psk_identity_hint
8982 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8984 * Add RFC 3161 compliant time stamp request creation, response generation
8985 and response verification functionality.
8987 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8989 * Add initial support for TLS extensions, specifically for the server_name
8990 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8991 have new members for a host name. The SSL data structure has an
8992 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8993 stored in that context to allow for session resumption, even after the
8994 SSL has been switched to a new SSL_CTX in reaction to a client's
8995 server_name extension.
8997 New functions (subject to change):
8999 SSL_get_servername()
9000 SSL_get_servername_type()
9003 New CTRL codes and macros (subject to change):
9005 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9006 - SSL_CTX_set_tlsext_servername_callback()
9007 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9008 - SSL_CTX_set_tlsext_servername_arg()
9009 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9011 openssl s_client has a new '-servername ...' option.
9013 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9014 '-key2 ...', '-servername_fatal' (subject to change). This allows
9015 testing the HostName extension for a specific single host name ('-cert'
9016 and '-key' remain fallbacks for handshakes without HostName
9017 negotiation). If the unrecognized_name alert has to be sent, this by
9018 default is a warning; it becomes fatal with the '-servername_fatal'
9021 *Peter Sylvester, Remy Allais, Christophe Renou*
9023 * Whirlpool hash implementation is added.
9027 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9028 bn(64,32). Because of instruction set limitations it doesn't have
9029 any negative impact on performance. This was done mostly in order
9030 to make it possible to share assembler modules, such as bn_mul_mont
9031 implementations, between 32- and 64-bit builds without hassle.
9035 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9036 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9041 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9042 dedicated Montgomery multiplication procedure, is introduced.
9043 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9044 "64-bit" performance on certain 32-bit targets.
9048 * New option SSL_OP_NO_COMP to disable use of compression selectively
9049 in SSL structures. New SSL ctrl to set maximum send fragment size.
9050 Save memory by setting the I/O buffer sizes dynamically instead of
9051 using the maximum available value.
9055 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9056 in addition to the text details.
9060 * Very, very preliminary EXPERIMENTAL support for printing of general
9061 ASN1 structures. This currently produces rather ugly output and doesn't
9062 handle several customised structures at all.
9066 * Integrated support for PVK file format and some related formats such
9067 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9068 these in the 'rsa' and 'dsa' utilities.
9072 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9076 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9077 place for the (very old) "NETSCAPE" format certificates which are now
9078 handled using new ASN1 code equivalents.
9082 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9083 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9084 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9088 * Modify CRL distribution points extension code to print out previously
9089 unsupported fields. Enhance extension setting code to allow setting of
9094 * Add print and set support for Issuing Distribution Point CRL extension.
9098 * Change 'Configure' script to enable Camellia by default.
9105 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9107 * When rejecting SSL/TLS records due to an incorrect version number, never
9108 update s->server with a new major version number. As of
9109 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9110 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9111 the previous behavior could result in a read attempt at NULL when
9112 receiving specific incorrect SSL/TLS records once record payload
9113 protection is active. ([CVE-2010-0740])
9115 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9117 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9118 could be crashed if the relevant tables were not present (e.g. chrooted).
9120 *Tomas Hoger <thoger@redhat.com>*
9122 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9124 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9126 *Martin Olsson, Neel Mehta*
9128 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9129 accommodate for stack sorting, always a write lock!).
9133 * On some versions of WIN32 Heap32Next is very slow. This can cause
9134 excessive delays in the RAND_poll(): over a minute. As a workaround
9135 include a time check in the inner Heap32Next loop too.
9139 * The code that handled flushing of data in SSL/TLS originally used the
9140 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9141 the problem outlined in PR#1949. The fix suggested there however can
9142 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9143 of Apache). So instead simplify the code to flush unconditionally.
9144 This should be fine since flushing with no data to flush is a no op.
9148 * Handle TLS versions 2.0 and later properly and correctly use the
9149 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9150 off ancient servers have a habit of sticking around for a while...
9154 * Modify compression code so it frees up structures without using the
9155 ex_data callbacks. This works around a problem where some applications
9156 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9157 restarting) then use compression (e.g. SSL with compression) later.
9158 This results in significant per-connection memory leaks and
9159 has caused some security issues including CVE-2008-1678 and
9164 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9165 change when encrypting or decrypting.
9169 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9170 connect and renegotiate with servers which do not support RI.
9171 Until RI is more widely deployed this option is enabled by default.
9175 * Add "missing" ssl ctrls to clear options and mode.
9179 * If client attempts to renegotiate and doesn't support RI respond with
9180 a no_renegotiation alert as required by RFC5746. Some renegotiating
9181 TLS clients will continue a connection gracefully when they receive
9182 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9183 waiting for a server hello which it will never receive. Now we treat a
9184 received no_renegotiation alert as a fatal error. This is because
9185 applications requesting a renegotiation might well expect it to succeed
9186 and would have no code in place to handle the server denying it so the
9187 only safe thing to do is to terminate the connection.
9191 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9192 peer supports secure renegotiation and 0 otherwise. Print out peer
9193 renegotiation support in s_client/s_server.
9197 * Replace the highly broken and deprecated SPKAC certification method with
9198 the updated NID creation version. This should correctly handle UTF8.
9202 * Implement RFC5746. Re-enable renegotiation but require the extension
9203 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9204 turns out to be a bad idea. It has been replaced by
9205 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9206 SSL_CTX_set_options(). This is really not recommended unless you
9207 know what you are doing.
9209 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9211 * Fixes to stateless session resumption handling. Use initial_ctx when
9212 issuing and attempting to decrypt tickets in case it has changed during
9213 servername handling. Use a non-zero length session ID when attempting
9214 stateless session resumption: this makes it possible to determine if
9215 a resumption has occurred immediately after receiving server hello
9216 (several places in OpenSSL subtly assume this) instead of later in
9221 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9222 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9223 fixes for a few places where the return code is not checked
9226 *Julia Lawall <julia@diku.dk>*
9228 * Add --strict-warnings option to Configure script to include devteam
9229 warnings in other configurations.
9233 * Add support for --libdir option and LIBDIR variable in makefiles. This
9234 makes it possible to install openssl libraries in locations which
9235 have names other than "lib", for example "/usr/lib64" which some
9238 *Steve Henson, based on patch from Jeremy Utley*
9240 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9241 X690 8.9.12 and can produce some misleading textual output of OIDs.
9243 *Steve Henson, reported by Dan Kaminsky*
9245 * Delete MD2 from algorithm tables. This follows the recommendation in
9246 several standards that it is not used in new applications due to
9247 several cryptographic weaknesses. For binary compatibility reasons
9248 the MD2 API is still compiled in by default.
9252 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9257 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9258 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9261 *Guenter <lists@gknw.net>*
9263 * Fix the server certificate chain building code to use X509_verify_cert(),
9264 it used to have an ad-hoc builder which was unable to cope with anything
9265 other than a simple chain.
9267 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9269 * Don't check self signed certificate signatures in X509_verify_cert()
9270 by default (a flag can override this): it just wastes time without
9271 adding any security. As a useful side effect self signed root CAs
9272 with non-FIPS digests are now usable in FIPS mode.
9276 * In dtls1_process_out_of_seq_message() the check if the current message
9277 is already buffered was missing. For every new message was memory
9278 allocated, allowing an attacker to perform an denial of service attack
9279 with sending out of seq handshake messages until there is no memory
9280 left. Additionally every future message was buffered, even if the
9281 sequence number made no sense and would be part of another handshake.
9282 So only messages with sequence numbers less than 10 in advance will be
9283 buffered. ([CVE-2009-1378])
9285 *Robin Seggelmann, discovered by Daniel Mentz*
9287 * Records are buffered if they arrive with a future epoch to be
9288 processed after finishing the corresponding handshake. There is
9289 currently no limitation to this buffer allowing an attacker to perform
9290 a DOS attack with sending records with future epochs until there is no
9291 memory left. This patch adds the pqueue_size() function to determine
9292 the size of a buffer and limits the record buffer to 100 entries.
9295 *Robin Seggelmann, discovered by Daniel Mentz*
9297 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9298 parent structure is freed. ([CVE-2009-1379])
9302 * Handle non-blocking I/O properly in SSL_shutdown() call.
9304 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9306 * Add `2.5.4.*` OIDs
9308 *Ilya O. <vrghost@gmail.com>*
9310 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9312 * Disable renegotiation completely - this fixes a severe security
9313 problem ([CVE-2009-3555]) at the cost of breaking all
9314 renegotiation. Renegotiation can be re-enabled by setting
9315 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9316 run-time. This is really not recommended unless you know what
9321 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9323 * Don't set val to NULL when freeing up structures, it is freed up by
9324 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9325 zeroing past the valid field. ([CVE-2009-0789])
9327 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9329 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9330 checked correctly. This would allow some invalid signed attributes to
9331 appear to verify correctly. ([CVE-2009-0591])
9333 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9335 * Reject UniversalString and BMPString types with invalid lengths. This
9336 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9337 a legal length. ([CVE-2009-0590])
9341 * Set S/MIME signing as the default purpose rather than setting it
9342 unconditionally. This allows applications to override it at the store
9347 * Permit restricted recursion of ASN1 strings. This is needed in practice
9348 to handle some structures.
9352 * Improve efficiency of mem_gets: don't search whole buffer each time
9355 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9357 * New -hex option for openssl rand.
9361 * Print out UTF8String and NumericString when parsing ASN1.
9365 * Support NumericString type for name components.
9369 * Allow CC in the environment to override the automatically chosen
9370 compiler. Note that nothing is done to ensure flags work with the
9375 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9377 * Properly check EVP_VerifyFinal() and similar return values
9380 *Ben Laurie, Bodo Moeller, Google Security Team*
9382 * Enable TLS extensions by default.
9386 * Allow the CHIL engine to be loaded, whether the application is
9387 multithreaded or not. (This does not release the developer from the
9388 obligation to set up the dynamic locking callbacks.)
9390 *Sander Temme <sander@temme.net>*
9392 * Use correct exit code if there is an error in dgst command.
9394 *Steve Henson; problem pointed out by Roland Dirlewanger*
9396 * Tweak Configure so that you need to say "experimental-jpake" to enable
9397 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9401 * Add experimental JPAKE support, including demo authentication in
9402 s_client and s_server.
9406 * Set the comparison function in v3_addr_canonize().
9408 *Rob Austein <sra@hactrn.net>*
9410 * Add support for XMPP STARTTLS in s_client.
9412 *Philip Paeps <philip@freebsd.org>*
9414 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9415 to ensure that even with this option, only ciphersuites in the
9416 server's preference list will be accepted. (Note that the option
9417 applies only when resuming a session, so the earlier behavior was
9418 just about the algorithm choice for symmetric cryptography.)
9422 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9424 * Fix NULL pointer dereference if a DTLS server received
9425 ChangeCipherSpec as first record ([CVE-2009-1386]).
9429 * Fix a state transition in s3_srvr.c and d1_srvr.c
9430 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9434 * The fix in 0.9.8c that supposedly got rid of unsafe
9435 double-checked locking was incomplete for RSA blinding,
9436 addressing just one layer of what turns out to have been
9437 doubly unsafe triple-checked locking.
9439 So now fix this for real by retiring the MONT_HELPER macro
9440 in crypto/rsa/rsa_eay.c.
9442 *Bodo Moeller; problem pointed out by Marius Schilder*
9444 * Various precautionary measures:
9446 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9448 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9449 (NB: This would require knowledge of the secret session ticket key
9450 to exploit, in which case you'd be SOL either way.)
9452 - Change bn_nist.c so that it will properly handle input BIGNUMs
9453 outside the expected range.
9455 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9458 *Neel Mehta, Bodo Moeller*
9460 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9461 the load fails. Useful for distros.
9463 *Ben Laurie and the FreeBSD team*
9465 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9469 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9473 * Expand ENGINE to support engine supplied SSL client certificate functions.
9475 This work was sponsored by Logica.
9479 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9480 keystores. Support for SSL/TLS client authentication too.
9481 Not compiled unless enable-capieng specified to Configure.
9483 This work was sponsored by Logica.
9487 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9488 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9489 attribute creation routines such as certificate requests and PKCS#12
9494 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9496 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9497 handshake which could lead to a client crash as found using the
9498 Codenomicon TLS test suite ([CVE-2008-1672])
9500 *Steve Henson, Mark Cox*
9502 * Fix double free in TLS server name extensions which could lead to
9503 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9507 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9509 Clear the error queue to ensure that error entries left from
9510 older function calls do not interfere with the correct operation.
9512 *Lutz Jaenicke, Erik de Castro Lopo*
9514 * Remove root CA certificates of commercial CAs:
9516 The OpenSSL project does not recommend any specific CA and does not
9517 have any policy with respect to including or excluding any CA.
9518 Therefore it does not make any sense to ship an arbitrary selection
9519 of root CA certificates with the OpenSSL software.
9523 * RSA OAEP patches to fix two separate invalid memory reads.
9524 The first one involves inputs when 'lzero' is greater than
9525 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9526 before the beginning of from). The second one involves inputs where
9527 the 'db' section contains nothing but zeroes (there is a one-byte
9528 invalid read after the end of 'db').
9530 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9532 * Partial backport from 0.9.9-dev:
9534 Introduce bn_mul_mont (dedicated Montgomery multiplication
9535 procedure) as a candidate for BIGNUM assembler implementation.
9536 While 0.9.9-dev uses assembler for various architectures, only
9537 x86_64 is available by default here in the 0.9.8 branch, and
9538 32-bit x86 is available through a compile-time setting.
9540 To try the 32-bit x86 assembler implementation, use Configure
9541 option "enable-montasm" (which exists only for this backport).
9543 As "enable-montasm" for 32-bit x86 disclaims code stability
9544 anyway, in this constellation we activate additional code
9545 backported from 0.9.9-dev for further performance improvements,
9546 namely BN_from_montgomery_word. (To enable this otherwise,
9547 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9549 *Andy Polyakov (backport partially by Bodo Moeller)*
9551 * Add TLS session ticket callback. This allows an application to set
9552 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9553 values. This is useful for key rollover for example where several key
9554 sets may exist with different names.
9558 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9559 This was broken until now in 0.9.8 releases, such that the only way
9560 a registered ENGINE could be used (assuming it initialises
9561 successfully on the host) was to explicitly set it as the default
9562 for the relevant algorithms. This is in contradiction with 0.9.7
9563 behaviour and the documentation. With this fix, when an ENGINE is
9564 registered into a given algorithm's table of implementations, the
9565 'uptodate' flag is reset so that auto-discovery will be used next
9566 time a new context for that algorithm attempts to select an
9569 *Ian Lister (tweaked by Geoff Thorpe)*
9571 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9572 implementation in the following ways:
9574 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9577 Lack of BER streaming support means one pass streaming processing is
9578 only supported if data is detached: setting the streaming flag is
9579 ignored for embedded content.
9581 CMS support is disabled by default and must be explicitly enabled
9582 with the enable-cms configuration option.
9586 * Update the GMP engine glue to do direct copies between BIGNUM and
9587 mpz_t when openssl and GMP use the same limb size. Otherwise the
9588 existing "conversion via a text string export" trick is still used.
9590 *Paul Sheer <paulsheer@gmail.com>*
9592 * Zlib compression BIO. This is a filter BIO which compressed and
9593 uncompresses any data passed through it.
9597 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9598 RFC3394 compatible AES key wrapping.
9602 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9603 sets string data without copying. X509_ALGOR_set0() and
9604 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9605 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9606 from an X509_ATTRIBUTE structure optionally checking it occurs only
9607 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9612 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9613 to get the expected BN_FLG_CONSTTIME behavior.
9615 *Bodo Moeller (Google)*
9619 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9620 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9621 - added some more tests to do_tests.pl
9622 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9623 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9624 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9625 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9626 - various changes to netware.pl to enable gcc-cross builds on Win32
9628 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9629 - various changes to fix missing prototype warnings
9630 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9631 - added AES, WHIRLPOOL and CPUID assembler code to build files
9632 - added missing AES assembler make rules to mk1mf.pl
9633 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9635 *Guenter Knauf <eflash@gmx.net>*
9637 * Implement certificate status request TLS extension defined in RFC3546.
9638 A client can set the appropriate parameters and receive the encoded
9639 OCSP response via a callback. A server can query the supplied parameters
9640 and set the encoded OCSP response in the callback. Add simplified examples
9641 to s_client and s_server.
9645 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9648 + Binary incompatibility of ssl_ctx_st structure
9649 + DTLS interoperation with non-compliant servers
9650 + Don't call get_session_cb() without proposed session
9651 + Fix ia64 assembler code
9653 *Andy Polyakov, Steve Henson*
9655 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9657 * DTLS Handshake overhaul. There were longstanding issues with
9658 OpenSSL DTLS implementation, which were making it impossible for
9659 RFC 4347 compliant client to communicate with OpenSSL server.
9660 Unfortunately just fixing these incompatibilities would "cut off"
9661 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9662 server keeps tolerating non RFC compliant syntax. The opposite is
9663 not true, 0.9.8f client can not communicate with earlier server.
9664 This update even addresses CVE-2007-4995.
9668 * Changes to avoid need for function casts in OpenSSL: some compilers
9669 (gcc 4.2 and later) reject their use.
9670 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9673 * Add RFC4507 support to OpenSSL. This includes the corrections in
9674 RFC4507bis. The encrypted ticket format is an encrypted encoded
9675 SSL_SESSION structure, that way new session features are automatically
9678 If a client application caches session in an SSL_SESSION structure
9679 support is transparent because tickets are now stored in the encoded
9682 The SSL_CTX structure automatically generates keys for ticket
9683 protection in servers so again support should be possible
9684 with no application modification.
9686 If a client or server wishes to disable RFC4507 support then the option
9687 SSL_OP_NO_TICKET can be set.
9689 Add a TLS extension debugging callback to allow the contents of any client
9690 or server extensions to be examined.
9692 This work was sponsored by Google.
9696 * Add initial support for TLS extensions, specifically for the server_name
9697 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9698 have new members for a host name. The SSL data structure has an
9699 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9700 stored in that context to allow for session resumption, even after the
9701 SSL has been switched to a new SSL_CTX in reaction to a client's
9702 server_name extension.
9704 New functions (subject to change):
9706 SSL_get_servername()
9707 SSL_get_servername_type()
9710 New CTRL codes and macros (subject to change):
9712 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9713 - SSL_CTX_set_tlsext_servername_callback()
9714 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9715 - SSL_CTX_set_tlsext_servername_arg()
9716 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9718 openssl s_client has a new '-servername ...' option.
9720 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9721 '-key2 ...', '-servername_fatal' (subject to change). This allows
9722 testing the HostName extension for a specific single host name ('-cert'
9723 and '-key' remain fallbacks for handshakes without HostName
9724 negotiation). If the unrecognized_name alert has to be sent, this by
9725 default is a warning; it becomes fatal with the '-servername_fatal'
9728 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9730 * Add AES and SSE2 assembly language support to VC++ build.
9734 * Mitigate attack on final subtraction in Montgomery reduction.
9738 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9739 (which previously caused an internal error).
9743 * Squeeze another 10% out of IGE mode when in != out.
9747 * AES IGE mode speedup.
9749 *Dean Gaudet (Google)*
9751 * Add the Korean symmetric 128-bit cipher SEED (see
9752 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9753 add SEED ciphersuites from RFC 4162:
9755 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9756 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9757 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9758 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9760 To minimize changes between patchlevels in the OpenSSL 0.9.8
9761 series, SEED remains excluded from compilation unless OpenSSL
9762 is configured with 'enable-seed'.
9764 *KISA, Bodo Moeller*
9766 * Mitigate branch prediction attacks, which can be practical if a
9767 single processor is shared, allowing a spy process to extract
9768 information. For detailed background information, see
9769 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9770 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9771 and Necessary Software Countermeasures"). The core of the change
9772 are new versions BN_div_no_branch() and
9773 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9774 respectively, which are slower, but avoid the security-relevant
9775 conditional branches. These are automatically called by BN_div()
9776 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9777 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9778 remove a conditional branch.
9780 BN_FLG_CONSTTIME is the new name for the previous
9781 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9782 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9783 in the exponent causes BN_mod_exp_mont() to use the alternative
9784 implementation in BN_mod_exp_mont_consttime().) The old name
9785 remains as a deprecated alias.
9787 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9788 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9789 constant-time implementations for more than just exponentiation.
9790 Here too the old name is kept as a deprecated alias.
9792 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9793 the BN_BLINDING structure gets an independent copy of the
9794 modulus. This means that the previous `BIGNUM *m` argument to
9795 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9796 essentially becomes `const BIGNUM *m`, although we can't actually
9797 change this in the header file before 0.9.9. It allows
9798 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9799 enable BN_FLG_CONSTTIME.
9801 *Matthew D Wood (Intel Corp)*
9803 * In the SSL/TLS server implementation, be strict about session ID
9804 context matching (which matters if an application uses a single
9805 external cache for different purposes). Previously,
9806 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9807 set. This did ensure strict client verification, but meant that,
9808 with applications using a single external cache for quite
9809 different requirements, clients could circumvent ciphersuite
9810 restrictions for a given session ID context by starting a session
9811 in a different context.
9815 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9816 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9817 authentication-only ciphersuites.
9821 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9822 not complete and could lead to a possible single byte overflow
9823 ([CVE-2007-5135]) [Ben Laurie]
9825 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9827 * Since AES128 and AES256 (and similarly Camellia128 and
9828 Camellia256) share a single mask bit in the logic of
9829 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9830 kludge to work properly if AES128 is available and AES256 isn't
9831 (or if Camellia128 is available and Camellia256 isn't).
9835 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9836 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9837 When a point or a seed is encoded in a BIT STRING, we need to
9838 prevent the removal of trailing zero bits to get the proper DER
9839 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9840 of a NamedBitList, for which trailing 0 bits need to be removed.)
9844 * Have SSL/TLS server implementation tolerate "mismatched" record
9845 protocol version while receiving ClientHello even if the
9846 ClientHello is fragmented. (The server can't insist on the
9847 particular protocol version it has chosen before the ServerHello
9848 message has informed the client about his choice.)
9852 * Add RFC 3779 support.
9854 *Rob Austein for ARIN, Ben Laurie*
9856 * Load error codes if they are not already present instead of using a
9857 static variable. This allows them to be cleanly unloaded and reloaded.
9858 Improve header file function name parsing.
9862 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9863 or CAPABILITY handshake as required by RFCs.
9867 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9869 * Introduce limits to prevent malicious keys being able to
9870 cause a denial of service. ([CVE-2006-2940])
9872 *Steve Henson, Bodo Moeller*
9874 * Fix ASN.1 parsing of certain invalid structures that can result
9875 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9877 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9878 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9880 * Fix SSL client code which could crash if connecting to a
9881 malicious SSLv2 server. ([CVE-2006-4343])
9883 *Tavis Ormandy and Will Drewry, Google Security Team*
9885 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9886 match only those. Before that, "AES256-SHA" would be interpreted
9887 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9888 the same strength classification in 0.9.7h) as we currently only
9889 have a single AES bit in the ciphersuite description bitmap.
9890 That change, however, also applied to ciphersuite strings such as
9891 "RC4-MD5" that intentionally matched multiple ciphersuites --
9892 namely, SSL 2.0 ciphersuites in addition to the more common ones
9893 from SSL 3.0/TLS 1.0.
9895 So we change the selection algorithm again: Naming an explicit
9896 ciphersuite selects this one ciphersuite, and any other similar
9897 ciphersuite (same bitmap) from *other* protocol versions.
9898 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9899 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9901 Since SSL 2.0 does not have any ciphersuites for which the
9902 128/256 bit distinction would be relevant, this works for now.
9903 The proper fix will be to use different bits for AES128 and
9904 AES256, which would have avoided the problems from the beginning;
9905 however, bits are scarce, so we can only do this in a new release
9906 (not just a patchlevel) when we can change the SSL_CIPHER
9907 definition to split the single 'unsigned long mask' bitmap into
9908 multiple values to extend the available space.
9912 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9914 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9915 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9917 * Add AES IGE and biIGE modes.
9921 * Change the Unix randomness entropy gathering to use poll() when
9922 possible instead of select(), since the latter has some
9923 undesirable limitations.
9925 *Darryl Miles via Richard Levitte and Bodo Moeller*
9927 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9928 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9929 cannot be implicitly activated as part of, e.g., the "AES" alias.
9930 However, please upgrade to OpenSSL 0.9.9[-dev] for
9931 non-experimental use of the ECC ciphersuites to get TLS extension
9932 support, which is required for curve and point format negotiation
9933 to avoid potential handshake problems.
9937 * Disable rogue ciphersuites:
9939 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9940 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9941 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9943 The latter two were purportedly from
9944 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9947 Also deactivate the remaining ciphersuites from
9948 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9949 unofficial, and the ID has long expired.
9953 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9954 dual-core machines) and other potential thread-safety issues.
9958 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9959 versions), which is now available for royalty-free use
9960 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9961 Also, add Camellia TLS ciphersuites from RFC 4132.
9963 To minimize changes between patchlevels in the OpenSSL 0.9.8
9964 series, Camellia remains excluded from compilation unless OpenSSL
9965 is configured with 'enable-camellia'.
9969 * Disable the padding bug check when compression is in use. The padding
9970 bug check assumes the first packet is of even length, this is not
9971 necessarily true if compression is enabled and can result in false
9972 positives causing handshake failure. The actual bug test is ancient
9973 code so it is hoped that implementations will either have fixed it by
9974 now or any which still have the bug do not support compression.
9978 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9980 * When applying a cipher rule check to see if string match is an explicit
9981 cipher suite and only match that one cipher suite if it is.
9985 * Link in manifests for VC++ if needed.
9987 *Austin Ziegler <halostatue@gmail.com>*
9989 * Update support for ECC-based TLS ciphersuites according to
9990 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9991 TLS extensions, which are supported starting with the 0.9.9
9992 branch, not in the OpenSSL 0.9.8 branch).
9996 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9997 opaque EVP_CIPHER_CTX handling.
10001 * Fixes and enhancements to zlib compression code. We now only use
10002 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10003 to conform with the standards mentioned here:
10004 <http://www.zlib.net/DLL_FAQ.txt>
10005 Static zlib linking now works on Windows and the new --with-zlib-include
10006 --with-zlib-lib options to Configure can be used to supply the location
10007 of the headers and library. Gracefully handle case where zlib library
10012 * Several fixes and enhancements to the OID generation code. The old code
10013 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10014 handle numbers larger than ULONG_MAX, truncated printing and had a
10015 non standard OBJ_obj2txt() behaviour.
10019 * Add support for building of engines under engine/ as shared libraries
10020 under VC++ build system.
10024 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10025 Hopefully, we will not see any false combination of paths any more.
10029 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10031 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10032 (part of SSL_OP_ALL). This option used to disable the
10033 countermeasure against man-in-the-middle protocol-version
10034 rollback in the SSL 2.0 server implementation, which is a bad
10035 idea. ([CVE-2005-2969])
10037 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10038 for Information Security, National Institute of Advanced Industrial
10039 Science and Technology [AIST], Japan)*
10041 * Add two function to clear and return the verify parameter flags.
10045 * Keep cipherlists sorted in the source instead of sorting them at
10046 runtime, thus removing the need for a lock.
10050 * Avoid some small subgroup attacks in Diffie-Hellman.
10052 *Nick Mathewson and Ben Laurie*
10054 * Add functions for well-known primes.
10058 * Extended Windows CE support.
10060 *Satoshi Nakamura and Andy Polyakov*
10062 * Initialize SSL_METHOD structures at compile time instead of during
10063 runtime, thus removing the need for a lock.
10067 * Make PKCS7_decrypt() work even if no certificate is supplied by
10068 attempting to decrypt each encrypted key in turn. Add support to
10073 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10075 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10078 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10082 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10083 key into the same file any more.
10087 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10091 * Add -utf8 command line and config file option to 'ca'.
10093 *Stefan <stf@udoma.org*
10095 * Removed the macro des_crypt(), as it seems to conflict with some
10096 libraries. Use DES_crypt().
10100 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10101 involves renaming the source and generated shared-libs for
10102 both. The engines will accept the corrected or legacy ids
10103 ('ncipher' and '4758_cca' respectively) when binding. NB,
10104 this only applies when building 'shared'.
10106 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10108 * Add attribute functions to EVP_PKEY structure. Modify
10109 PKCS12_create() to recognize a CSP name attribute and
10110 use it. Make -CSP option work again in pkcs12 utility.
10114 * Add new functionality to the bn blinding code:
10115 - automatic re-creation of the BN_BLINDING parameters after
10116 a fixed number of uses (currently 32)
10117 - add new function for parameter creation
10118 - introduce flags to control the update behaviour of the
10119 BN_BLINDING parameters
10120 - hide BN_BLINDING structure
10121 Add a second BN_BLINDING slot to the RSA structure to improve
10122 performance when a single RSA object is shared among several
10127 * Add support for DTLS.
10129 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10131 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10132 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10136 * Remove buggy and incomplete DH cert support from
10137 ssl/ssl_rsa.c and ssl/s3_both.c
10141 * Use SHA-1 instead of MD5 as the default digest algorithm for
10142 the `apps/openssl` commands.
10146 * Compile clean with "-Wall -Wmissing-prototypes
10147 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10148 DEBUG_SAFESTACK must also be set.
10152 * Change ./Configure so that certain algorithms can be disabled by default.
10153 The new counterpiece to "no-xxx" is "enable-xxx".
10155 The patented RC5 and MDC2 algorithms will now be disabled unless
10156 "enable-rc5" and "enable-mdc2", respectively, are specified.
10158 (IDEA remains enabled despite being patented. This is because IDEA
10159 is frequently required for interoperability, and there is no license
10160 fee for non-commercial use. As before, "no-idea" can be used to
10161 avoid this algorithm.)
10165 * Add processing of proxy certificates (see RFC 3820). This work was
10166 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10167 EGEE (Enabling Grids for E-science in Europe).
10171 * RC4 performance overhaul on modern architectures/implementations, such
10172 as Intel P4, IA-64 and AMD64.
10176 * New utility extract-section.pl. This can be used specify an alternative
10177 section number in a pod file instead of having to treat each file as
10178 a separate case in Makefile. This can be done by adding two lines to the
10181 =for comment openssl_section:XXX
10183 The blank line is mandatory.
10187 * New arguments -certform, -keyform and -pass for s_client and s_server
10188 to allow alternative format key and certificate files and passphrase
10193 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10194 update associated structures and add various utility functions.
10196 Add new policy related verify parameters, include policy checking in
10197 standard verify code. Enhance 'smime' application with extra parameters
10198 to support policy checking and print out.
10202 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10203 Nehemiah processors. These extensions support AES encryption in hardware
10204 as well as RNG (though RNG support is currently disabled).
10206 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10208 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10212 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10214 *Andy Polyakov and a number of other people*
10216 * Improved PowerPC platform support. Most notably BIGNUM assembler
10217 implementation contributed by IBM.
10219 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10221 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10222 exponent rather than 'unsigned long'. There is a corresponding change to
10223 the new 'rsa_keygen' element of the RSA_METHOD structure.
10225 *Jelte Jansen, Geoff Thorpe*
10227 * Functionality for creating the initial serial number file is now
10228 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10230 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10231 number file to 1, which is bound to cause problems. To avoid
10232 the problems while respecting compatibility between different 0.9.7
10233 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10234 CA.pl for serial number initialization. With the new release 0.9.8,
10235 we can fix the problem directly in the 'ca' utility.)
10239 * Reduced header interdependencies by declaring more opaque objects in
10240 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10241 give fewer recursive includes, which could break lazy source code - so
10242 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10243 developers should define this symbol when building and using openssl to
10244 ensure they track the recommended behaviour, interfaces, [etc], but
10245 backwards-compatible behaviour prevails when this isn't defined.
10249 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10253 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10254 This will generate a random key of the appropriate length based on the
10255 cipher context. The EVP_CIPHER can provide its own random key generation
10256 routine to support keys of a specific form. This is used in the des and
10257 3des routines to generate a key of the correct parity. Update S/MIME
10258 code to use new functions and hence generate correct parity DES keys.
10259 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10260 valid (weak or incorrect parity).
10264 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10265 as looking them up. This is useful when the verified structure may contain
10266 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10267 present unless the new PKCS7_NO_CRL flag is asserted.
10271 * Extend ASN1 oid configuration module. It now additionally accepts the
10274 shortName = some long name, 1.2.3.4
10278 * Reimplemented the BN_CTX implementation. There is now no more static
10279 limitation on the number of variables it can handle nor the depth of the
10280 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10281 information can now expand as required, and rather than having a single
10282 static array of bignums, BN_CTX now uses a linked-list of such arrays
10283 allowing it to expand on demand whilst maintaining the usefulness of
10284 BN_CTX's "bundling".
10288 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10289 to allow all RSA operations to function using a single BN_CTX.
10293 * Preliminary support for certificate policy evaluation and checking. This
10294 is initially intended to pass the tests outlined in "Conformance Testing
10295 of Relying Party Client Certificate Path Processing Logic" v1.07.
10299 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10300 remained unused and not that useful. A variety of other little bignum
10301 tweaks and fixes have also been made continuing on from the audit (see
10306 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10307 associated ASN1, EVP and SSL functions and old ASN1 macros.
10311 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10312 and this should never fail. So the return value from the use of
10313 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10314 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10318 * BN_CTX_get() should return zero-valued bignums, providing the same
10319 initialised value as BN_new().
10321 *Geoff Thorpe, suggested by Ulf Möller*
10323 * Support for inhibitAnyPolicy certificate extension.
10327 * An audit of the BIGNUM code is underway, for which debugging code is
10328 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10329 is considered valid when processing BIGNUMs, and causes execution to
10330 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10331 further steps are taken to deliberately pollute unused data in BIGNUM
10332 structures to try and expose faulty code further on. For now, openssl will
10333 (in its default mode of operation) continue to tolerate the inconsistent
10334 forms that it has tolerated in the past, but authors and packagers should
10335 consider trying openssl and their own applications when compiled with
10336 these debugging symbols defined. It will help highlight potential bugs in
10337 their own code, and will improve the test coverage for OpenSSL itself. At
10338 some point, these tighter rules will become openssl's default to improve
10339 maintainability, though the assert()s and other overheads will remain only
10340 in debugging configurations. See bn.h for more details.
10342 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10344 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10345 that can only be obtained through BN_CTX_new() (which implicitly
10346 initialises it). The presence of this function only made it possible
10347 to overwrite an existing structure (and cause memory leaks).
10351 * Because of the callback-based approach for implementing LHASH as a
10352 template type, lh_insert() adds opaque objects to hash-tables and
10353 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10354 to clean up those corresponding objects before destroying the hash table
10355 (and losing the object pointers). So some over-zealous constifications in
10356 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10357 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10358 prototyped to have "const" restrictions on the object pointers they are
10359 given (and so aren't required to cast them away any more).
10363 * The tmdiff.h API was so ugly and minimal that our own timing utility
10364 (speed) prefers to use its own implementation. The two implementations
10365 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10366 its object type properly exposed (MS_TM) instead of casting to/from
10367 `char *`. This may still change yet if someone realises MS_TM and
10369 aren't necessarily the greatest nomenclatures - but this is what was used
10370 internally to the implementation so I've used that for now.
10374 * Ensure that deprecated functions do not get compiled when
10375 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10376 the self-tests were still using deprecated key-generation functions so
10377 these have been updated also.
10381 * Reorganise PKCS#7 code to separate the digest location functionality
10382 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10383 New function PKCS7_set_digest() to set the digest type for PKCS#7
10384 digestedData type. Add additional code to correctly generate the
10385 digestedData type and add support for this type in PKCS7 initialization
10390 * New function PKCS7_set0_type_other() this initializes a PKCS7
10391 structure of type "other".
10395 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10396 sure the loop does correctly stop and breaking ("division by zero")
10397 modulus operations are not performed. The (pre-generated) prime
10398 table crypto/bn/bn_prime.h was already correct, but it could not be
10399 re-generated on some platforms because of the "division by zero"
10400 situation in the script.
10402 *Ralf S. Engelschall*
10404 * Update support for ECC-based TLS ciphersuites according to
10405 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10406 SHA-1 now is only used for "small" curves (where the
10407 representation of a field element takes up to 24 bytes); for
10408 larger curves, the field element resulting from ECDH is directly
10409 used as premaster secret.
10411 *Douglas Stebila (Sun Microsystems Laboratories)*
10413 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10414 curve secp160r1 to the tests.
10416 *Douglas Stebila (Sun Microsystems Laboratories)*
10418 * Add the possibility to load symbols globally with DSO.
10420 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10422 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10423 control of the error stack.
10427 * Add support for STORE in ENGINE.
10431 * Add the STORE type. The intention is to provide a common interface
10432 to certificate and key stores, be they simple file-based stores, or
10433 HSM-type store, or LDAP stores, or...
10434 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10438 * Add a generic structure called OPENSSL_ITEM. This can be used to
10439 pass a list of arguments to any function as well as provide a way
10440 for a function to pass data back to the caller.
10444 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10445 works like BUF_strdup() but can be used to duplicate a portion of
10446 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10451 * Add the function sk_find_ex() which works like sk_find(), but will
10452 return an index to an element even if an exact match couldn't be
10453 found. The index is guaranteed to point at the element where the
10454 searched-for key would be inserted to preserve sorting order.
10458 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10459 takes an extra flags argument for optional functionality. Currently,
10460 the following flags are defined:
10462 OBJ_BSEARCH_VALUE_ON_NOMATCH
10463 This one gets OBJ_bsearch_ex() to return a pointer to the first
10464 element where the comparing function returns a negative or zero
10467 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10468 This one gets OBJ_bsearch_ex() to return a pointer to the first
10469 element where the comparing function returns zero. This is useful
10470 if there are more than one element where the comparing function
10475 * Make it possible to create self-signed certificates with 'openssl ca'
10476 in such a way that the self-signed certificate becomes part of the
10477 CA database and uses the same mechanisms for serial number generation
10478 as all other certificate signing. The new flag '-selfsign' enables
10479 this functionality. Adapt CA.sh and CA.pl.in.
10483 * Add functionality to check the public key of a certificate request
10484 against a given private. This is useful to check that a certificate
10485 request can be signed by that key (self-signing).
10489 * Make it possible to have multiple active certificates with the same
10490 subject in the CA index file. This is done only if the keyword
10491 'unique_subject' is set to 'no' in the main CA section (default
10492 if 'CA_default') of the configuration file. The value is saved
10493 with the database itself in a separate index attribute file,
10494 named like the index file with '.attr' appended to the name.
10498 * Generate multi-valued AVAs using '+' notation in config files for
10503 * Support for nameConstraints certificate extension.
10507 * Support for policyConstraints certificate extension.
10511 * Support for policyMappings certificate extension.
10515 * Make sure the default DSA_METHOD implementation only uses its
10516 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10517 and change its own handlers to be NULL so as to remove unnecessary
10518 indirection. This lets alternative implementations fallback to the
10519 default implementation more easily.
10523 * Support for directoryName in GeneralName related extensions
10528 * Make it possible to link applications using Makefile.shared.
10529 Make that possible even when linking against static libraries!
10533 * Support for single pass processing for S/MIME signing. This now
10534 means that S/MIME signing can be done from a pipe, in addition
10535 cleartext signing (multipart/signed type) is effectively streaming
10536 and the signed data does not need to be all held in memory.
10538 This is done with a new flag PKCS7_STREAM. When this flag is set
10539 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10540 is done after the data is output (and digests calculated) in
10541 SMIME_write_PKCS7().
10545 * Add full support for -rpath/-R, both in shared libraries and
10546 applications, at least on the platforms where it's known how
10551 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10552 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10553 will now compute a table of multiples of the generator that
10554 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10555 faster (notably in the case of a single point multiplication,
10556 scalar * generator).
10558 *Nils Larsch, Bodo Moeller*
10560 * IPv6 support for certificate extensions. The various extensions
10561 which use the IP:a.b.c.d can now take IPv6 addresses using the
10562 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10567 * Added an ENGINE that implements RSA by performing private key
10568 exponentiations with the GMP library. The conversions to and from
10569 GMP's mpz_t format aren't optimised nor are any montgomery forms
10570 cached, and on x86 it appears OpenSSL's own performance has caught up.
10571 However there are likely to be other architectures where GMP could
10572 provide a boost. This ENGINE is not built in by default, but it can be
10573 specified at Configure time and should be accompanied by the necessary
10574 linker additions, eg;
10575 ./config -DOPENSSL_USE_GMP -lgmp
10579 * "openssl engine" will not display ENGINE/DSO load failure errors when
10580 testing availability of engines with "-t" - the old behaviour is
10581 produced by increasing the feature's verbosity with "-tt".
10585 * ECDSA routines: under certain error conditions uninitialized BN objects
10586 could be freed. Solution: make sure initialization is performed early
10587 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10592 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10593 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10594 software implementations. For DSA and DH, parameter generation can
10595 also be overridden by providing the appropriate method callbacks.
10599 * Change the "progress" mechanism used in key-generation and
10600 primality testing to functions that take a new BN_GENCB pointer in
10601 place of callback/argument pairs. The new API functions have `_ex`
10602 postfixes and the older functions are reimplemented as wrappers for
10603 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10604 declarations of the old functions to help (graceful) attempts to
10605 migrate to the new functions. Also, the new key-generation API
10606 functions operate on a caller-supplied key-structure and return
10607 success/failure rather than returning a key or NULL - this is to
10608 help make "keygen" another member function of RSA_METHOD etc.
10610 Example for using the new callback interface:
10612 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10613 void *my_arg = ...;
10616 BN_GENCB_set(&my_cb, my_callback, my_arg);
10618 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10619 /* For the meaning of a, b in calls to my_callback(), see the
10620 * documentation of the function that calls the callback.
10621 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10622 * my_callback should return 1 if it wants BN_is_prime_ex()
10623 * to continue, or 0 to stop.
10628 * Change the ZLIB compression method to be stateful, and make it
10629 available to TLS with the number defined in
10630 draft-ietf-tls-compression-04.txt.
10634 * Add the ASN.1 structures and functions for CertificatePair, which
10635 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10637 CertificatePair ::= SEQUENCE {
10638 forward [0] Certificate OPTIONAL,
10639 reverse [1] Certificate OPTIONAL,
10640 -- at least one of the pair shall be present -- }
10642 Also implement the PEM functions to read and write certificate
10643 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10645 This needed to be defined, mostly for the sake of the LDAP
10646 attribute crossCertificatePair, but may prove useful elsewhere as
10651 * Make it possible to inhibit symlinking of shared libraries in
10652 Makefile.shared, for Cygwin's sake.
10656 * Extend the BIGNUM API by creating a function
10657 void BN_set_negative(BIGNUM *a, int neg);
10658 and a macro that behave like
10659 int BN_is_negative(const BIGNUM *a);
10661 to avoid the need to access 'a->neg' directly in applications.
10665 * Implement fast modular reduction for pseudo-Mersenne primes
10666 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10667 EC_GROUP_new_curve_GFp() will now automatically use this
10670 *Nils Larsch <nla@trustcenter.de>*
10672 * Add new lock type (CRYPTO_LOCK_BN).
10676 * Change the ENGINE framework to automatically load engines
10677 dynamically from specific directories unless they could be
10678 found to already be built in or loaded. Move all the
10679 current engines except for the cryptodev one to a new
10680 directory engines/.
10681 The engines in engines/ are built as shared libraries if
10682 the "shared" options was given to ./Configure or ./config.
10683 Otherwise, they are inserted in libcrypto.a.
10684 /usr/local/ssl/engines is the default directory for dynamic
10685 engines, but that can be overridden at configure time through
10686 the usual use of --prefix and/or --openssldir, and at run
10687 time with the environment variable OPENSSL_ENGINES.
10689 *Geoff Thorpe and Richard Levitte*
10691 * Add Makefile.shared, a helper makefile to build shared
10692 libraries. Adapt Makefile.org.
10696 * Add version info to Win32 DLLs.
10698 *Peter 'Luna' Runestig" <peter@runestig.com>*
10700 * Add new 'medium level' PKCS#12 API. Certificates and keys
10701 can be added using this API to created arbitrary PKCS#12
10702 files while avoiding the low level API.
10704 New options to PKCS12_create(), key or cert can be NULL and
10705 will then be omitted from the output file. The encryption
10706 algorithm NIDs can be set to -1 for no encryption, the mac
10707 iteration count can be set to 0 to omit the mac.
10709 Enhance pkcs12 utility by making the -nokeys and -nocerts
10710 options work when creating a PKCS#12 file. New option -nomac
10711 to omit the mac, NONE can be set for an encryption algorithm.
10712 New code is modified to use the enhanced PKCS12_create()
10713 instead of the low level API.
10717 * Extend ASN1 encoder to support indefinite length constructed
10718 encoding. This can output sequences tags and octet strings in
10719 this form. Modify pk7_asn1.c to support indefinite length
10720 encoding. This is experimental and needs additional code to
10721 be useful, such as an ASN1 bio and some enhanced streaming
10724 Extend template encode functionality so that tagging is passed
10725 down to the template encoder.
10729 * Let 'openssl req' fail if an argument to '-newkey' is not
10730 recognized instead of using RSA as a default.
10734 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10735 As these are not official, they are not included in "ALL";
10736 the "ECCdraft" ciphersuite group alias can be used to select them.
10738 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10740 * Add ECDH engine support.
10742 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10744 * Add ECDH in new directory crypto/ecdh/.
10746 *Douglas Stebila (Sun Microsystems Laboratories)*
10748 * Let BN_rand_range() abort with an error after 100 iterations
10749 without success (which indicates a broken PRNG).
10753 * Change BN_mod_sqrt() so that it verifies that the input value
10754 is really the square of the return value. (Previously,
10755 BN_mod_sqrt would show GIGO behaviour.)
10759 * Add named elliptic curves over binary fields from X9.62, SECG,
10760 and WAP/WTLS; add OIDs that were still missing.
10762 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10764 * Extend the EC library for elliptic curves over binary fields
10765 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10768 EC_GF2m_simple_method
10772 EC_GROUP_new_curve_GF2m
10773 EC_GROUP_set_curve_GF2m
10774 EC_GROUP_get_curve_GF2m
10775 EC_POINT_set_affine_coordinates_GF2m
10776 EC_POINT_get_affine_coordinates_GF2m
10777 EC_POINT_set_compressed_coordinates_GF2m
10779 Point compression for binary fields is disabled by default for
10780 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10783 As binary polynomials are represented as BIGNUMs, various members
10784 of the EC_GROUP and EC_POINT data structures can be shared
10785 between the implementations for prime fields and binary fields;
10786 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10787 are essentially identical to their `..._GFp` counterparts.
10788 (For simplicity, the `..._GFp` prefix has been dropped from
10789 various internal method names.)
10791 An internal 'field_div' method (similar to 'field_mul' and
10792 'field_sqr') has been added; this is used only for binary fields.
10794 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10796 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10797 through methods ('mul', 'precompute_mult').
10799 The generic implementations (now internally called 'ec_wNAF_mul'
10800 and 'ec_wNAF_precomputed_mult') remain the default if these
10801 methods are undefined.
10803 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10805 * New function EC_GROUP_get_degree, which is defined through
10806 EC_METHOD. For curves over prime fields, this returns the bit
10807 length of the modulus.
10809 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10811 * New functions EC_GROUP_dup, EC_POINT_dup.
10812 (These simply call ..._new and ..._copy).
10814 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10816 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10817 Polynomials are represented as BIGNUMs (where the sign bit is not
10818 used) in the following functions [macros]:
10821 BN_GF2m_sub [= BN_GF2m_add]
10822 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10823 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10824 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10826 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10827 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10828 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10829 BN_GF2m_cmp [= BN_ucmp]
10831 (Note that only the 'mod' functions are actually for fields GF(2^m).
10832 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10834 For some functions, an the irreducible polynomial defining a
10835 field can be given as an 'unsigned int[]' with strictly
10836 decreasing elements giving the indices of those bits that are set;
10837 i.e., p[] represents the polynomial
10838 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10840 p[0] > p[1] > ... > p[k] = 0.
10841 This applies to the following functions:
10844 BN_GF2m_mod_mul_arr
10845 BN_GF2m_mod_sqr_arr
10846 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10847 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10848 BN_GF2m_mod_exp_arr
10849 BN_GF2m_mod_sqrt_arr
10850 BN_GF2m_mod_solve_quad_arr
10854 Conversion can be performed by the following functions:
10859 bntest.c has additional tests for binary polynomial arithmetic.
10861 Two implementations for BN_GF2m_mod_div() are available.
10862 The default algorithm simply uses BN_GF2m_mod_inv() and
10863 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10864 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10865 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10867 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10869 * Add new error code 'ERR_R_DISABLED' that can be used when some
10870 functionality is disabled at compile-time.
10872 *Douglas Stebila <douglas.stebila@sun.com>*
10874 * Change default behaviour of 'openssl asn1parse' so that more
10875 information is visible when viewing, e.g., a certificate:
10877 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10878 mode the content of non-printable OCTET STRINGs is output in a
10879 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10880 avoid the appearance of a printable string.
10882 *Nils Larsch <nla@trustcenter.de>*
10884 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10886 EC_GROUP_set_asn1_flag()
10887 EC_GROUP_get_asn1_flag()
10888 EC_GROUP_set_point_conversion_form()
10889 EC_GROUP_get_point_conversion_form()
10890 These control ASN1 encoding details:
10891 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10892 has been set to OPENSSL_EC_NAMED_CURVE.
10893 - Points are encoded in uncompressed form by default; options for
10894 asn1_for are as for point2oct, namely
10895 POINT_CONVERSION_COMPRESSED
10896 POINT_CONVERSION_UNCOMPRESSED
10897 POINT_CONVERSION_HYBRID
10899 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10901 EC_GROUP_set_seed()
10902 EC_GROUP_get0_seed()
10903 EC_GROUP_get_seed_len()
10904 This is used only for ASN1 purposes (so far).
10906 *Nils Larsch <nla@trustcenter.de>*
10908 * Add 'field_type' member to EC_METHOD, which holds the NID
10909 of the appropriate field type OID. The new function
10910 EC_METHOD_get_field_type() returns this value.
10912 *Nils Larsch <nla@trustcenter.de>*
10915 EC_POINT_point2bn()
10916 EC_POINT_bn2point()
10917 EC_POINT_point2hex()
10918 EC_POINT_hex2point()
10919 providing useful interfaces to EC_POINT_point2oct() and
10920 EC_POINT_oct2point().
10922 *Nils Larsch <nla@trustcenter.de>*
10924 * Change internals of the EC library so that the functions
10925 EC_GROUP_set_generator()
10926 EC_GROUP_get_generator()
10927 EC_GROUP_get_order()
10928 EC_GROUP_get_cofactor()
10929 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10930 to methods, which would lead to unnecessary code duplication when
10931 adding different types of curves.
10933 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10935 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10936 arithmetic, and such that modified wNAFs are generated
10937 (which avoid length expansion in many cases).
10941 * Add a function EC_GROUP_check_discriminant() (defined via
10942 EC_METHOD) that verifies that the curve discriminant is non-zero.
10944 Add a function EC_GROUP_check() that makes some sanity tests
10945 on a EC_GROUP, its generator and order. This includes
10946 EC_GROUP_check_discriminant().
10948 *Nils Larsch <nla@trustcenter.de>*
10950 * Add ECDSA in new directory crypto/ecdsa/.
10952 Add applications 'openssl ecparam' and 'openssl ecdsa'
10953 (these are based on 'openssl dsaparam' and 'openssl dsa').
10955 ECDSA support is also included in various other files across the
10956 library. Most notably,
10957 - 'openssl req' now has a '-newkey ecdsa:file' option;
10958 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10959 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10960 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10961 them suitable for ECDSA where domain parameters must be
10962 extracted before the specific public key;
10963 - ECDSA engine support has been added.
10965 *Nils Larsch <nla@trustcenter.de>*
10967 * Include some named elliptic curves, and add OIDs from X9.62,
10968 SECG, and WAP/WTLS. Each curve can be obtained from the new
10970 EC_GROUP_new_by_curve_name(),
10971 and the list of available named curves can be obtained with
10972 EC_get_builtin_curves().
10973 Also add a 'curve_name' member to EC_GROUP objects, which can be
10975 EC_GROUP_set_curve_name()
10976 EC_GROUP_get_curve_name()
10978 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10980 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10981 was actually never needed) and in BN_mul(). The removal in BN_mul()
10982 required a small change in bn_mul_part_recursive() and the addition
10983 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10984 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10985 bn_sub_words() and bn_add_words() except they take arrays with
10990 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10992 * Cleanse PEM buffers before freeing them since they may contain
10995 *Benjamin Bennett <ben@psc.edu>*
10997 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10998 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10999 authentication-only ciphersuites.
11003 * Since AES128 and AES256 share a single mask bit in the logic of
11004 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11005 kludge to work properly if AES128 is available and AES256 isn't.
11009 * Expand security boundary to match 1.1.1 module.
11013 * Remove redundant features: hash file source, editing of test vectors
11014 modify fipsld to use external fips_premain.c signature.
11018 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11019 run algorithm test programs.
11023 * Make algorithm test programs more tolerant of whitespace.
11027 * Have SSL/TLS server implementation tolerate "mismatched" record
11028 protocol version while receiving ClientHello even if the
11029 ClientHello is fragmented. (The server can't insist on the
11030 particular protocol version it has chosen before the ServerHello
11031 message has informed the client about his choice.)
11035 * Load error codes if they are not already present instead of using a
11036 static variable. This allows them to be cleanly unloaded and reloaded.
11040 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11042 * Introduce limits to prevent malicious keys being able to
11043 cause a denial of service. ([CVE-2006-2940])
11045 *Steve Henson, Bodo Moeller*
11047 * Fix ASN.1 parsing of certain invalid structures that can result
11048 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11050 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11051 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11053 * Fix SSL client code which could crash if connecting to a
11054 malicious SSLv2 server. ([CVE-2006-4343])
11056 *Tavis Ormandy and Will Drewry, Google Security Team*
11058 * Change ciphersuite string processing so that an explicit
11059 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11060 will no longer include "AES128-SHA"), and any other similar
11061 ciphersuite (same bitmap) from *other* protocol versions (so that
11062 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11063 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11064 changes from 0.9.8b and 0.9.8d.
11068 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11070 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11071 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11073 * Change the Unix randomness entropy gathering to use poll() when
11074 possible instead of select(), since the latter has some
11075 undesirable limitations.
11077 *Darryl Miles via Richard Levitte and Bodo Moeller*
11079 * Disable rogue ciphersuites:
11081 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11082 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11083 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11085 The latter two were purportedly from
11086 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11089 Also deactivate the remaining ciphersuites from
11090 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11091 unofficial, and the ID has long expired.
11095 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11096 dual-core machines) and other potential thread-safety issues.
11100 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11102 * Adapt fipsld and the build system to link against the validated FIPS
11103 module in FIPS mode.
11107 * Fixes for VC++ 2005 build under Windows.
11111 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11112 from a Windows bash shell such as MSYS. It is autodetected from the
11113 "config" script when run from a VC++ environment. Modify standard VC++
11114 build to use fipscanister.o from the GNU make build.
11118 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11120 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11121 The value now differs depending on if you build for FIPS or not.
11122 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11123 safely run with a non-FIPSed libcrypto, as it may crash because of
11124 the difference induced by this change.
11128 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11130 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11131 (part of SSL_OP_ALL). This option used to disable the
11132 countermeasure against man-in-the-middle protocol-version
11133 rollback in the SSL 2.0 server implementation, which is a bad
11134 idea. ([CVE-2005-2969])
11136 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11137 for Information Security, National Institute of Advanced Industrial
11138 Science and Technology [AIST, Japan)]*
11140 * Minimal support for X9.31 signatures and PSS padding modes. This is
11141 mainly for FIPS compliance and not fully integrated at this stage.
11145 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11146 the exponentiation using a fixed-length exponent. (Otherwise,
11147 the information leaked through timing could expose the secret key
11148 after many signatures; cf. Bleichenbacher's attack on DSA with
11153 * Make a new fixed-window mod_exp implementation the default for
11154 RSA, DSA, and DH private-key operations so that the sequence of
11155 squares and multiplies and the memory access pattern are
11156 independent of the particular secret key. This will mitigate
11157 cache-timing and potential related attacks.
11159 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11160 and this is automatically used by BN_mod_exp_mont() if the new flag
11161 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11162 will use this BN flag for private exponents unless the flag
11163 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11164 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11166 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11168 * Change the client implementation for SSLv23_method() and
11169 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11170 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11171 (Previously, the SSL 2.0 backwards compatible Client Hello
11172 message format would be used even with SSL_OP_NO_SSLv2.)
11176 * Add support for smime-type MIME parameter in S/MIME messages which some
11181 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11182 a threadsafe manner. Modify rsa code to use new function and add calls
11183 to dsa and dh code (which had race conditions before).
11187 * Include the fixed error library code in the C error file definitions
11188 instead of fixing them up at runtime. This keeps the error code
11189 structures constant.
11193 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11195 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11198 * Fixes for newer kerberos headers. NB: the casts are needed because
11199 the 'length' field is signed on one version and unsigned on another
11200 with no (?) obvious way to tell the difference, without these VC++
11201 complains. Also the "definition" of FAR (blank) is no longer included
11202 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11203 some needed definitions.
11207 * Undo Cygwin change.
11211 * Added support for proxy certificates according to RFC 3820.
11212 Because they may be a security thread to unaware applications,
11213 they must be explicitly allowed in run-time. See
11214 docs/HOWTO/proxy_certificates.txt for further information.
11218 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11220 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11221 server and client random values. Previously
11222 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11223 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11225 This change has negligible security impact because:
11227 1. Server and client random values still have 24 bytes of pseudo random
11230 2. Server and client random values are sent in the clear in the initial
11233 3. The master secret is derived using the premaster secret (48 bytes in
11234 size for static RSA ciphersuites) as well as client server and random
11237 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11240 *Stephen Henson, reported by UK NISCC*
11242 * Use Windows randomness collection on Cygwin.
11246 * Fix hang in EGD/PRNGD query when communication socket is closed
11247 prematurely by EGD/PRNGD.
11249 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11251 * Prompt for pass phrases when appropriate for PKCS12 input format.
11255 * Back-port of selected performance improvements from development
11256 branch, as well as improved support for PowerPC platforms.
11260 * Add lots of checks for memory allocation failure, error codes to indicate
11261 failure and freeing up memory if a failure occurs.
11263 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11265 * Add new -passin argument to dgst.
11269 * Perform some character comparisons of different types in X509_NAME_cmp:
11270 this is needed for some certificates that re-encode DNs into UTF8Strings
11271 (in violation of RFC3280) and can't or won't issue name rollover
11276 * Make an explicit check during certificate validation to see that
11277 the CA setting in each certificate on the chain is correct. As a
11278 side effect always do the following basic checks on extensions,
11279 not just when there's an associated purpose to the check:
11281 - if there is an unhandled critical extension (unless the user
11282 has chosen to ignore this fault)
11283 - if the path length has been exceeded (if one is set at all)
11284 - that certain extensions fit the associated purpose (if one has
11289 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11291 * Avoid a race condition when CRLs are checked in a multi threaded
11292 environment. This would happen due to the reordering of the revoked
11293 entries during signature checking and serial number lookup. Now the
11294 encoding is cached and the serial number sort performed under a lock.
11295 Add new STACK function sk_is_sorted().
11299 * Add Delta CRL to the extension code.
11303 * Various fixes to s3_pkt.c so alerts are sent properly.
11305 *David Holmes <d.holmes@f5.com>*
11307 * Reduce the chances of duplicate issuer name and serial numbers (in
11308 violation of RFC3280) using the OpenSSL certificate creation utilities.
11309 This is done by creating a random 64 bit value for the initial serial
11310 number when a serial number file is created or when a self signed
11311 certificate is created using 'openssl req -x509'. The initial serial
11312 number file is created using 'openssl x509 -next_serial' in CA.pl
11313 rather than being initialized to 1.
11317 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11319 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11320 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11322 *Joe Orton, Steve Henson*
11324 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11327 *Joe Orton, Steve Henson*
11329 * Make it possible to have multiple active certificates with the same
11330 subject in the CA index file. This is done only if the keyword
11331 'unique_subject' is set to 'no' in the main CA section (default
11332 if 'CA_default') of the configuration file. The value is saved
11333 with the database itself in a separate index attribute file,
11334 named like the index file with '.attr' appended to the name.
11338 * X509 verify fixes. Disable broken certificate workarounds when
11339 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11340 keyUsage extension present. Don't accept CRLs with unhandled critical
11341 extensions: since verify currently doesn't process CRL extensions this
11342 rejects a CRL with *any* critical extensions. Add new verify error codes
11347 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11348 A clarification of RFC2560 will require the use of OCTET STRINGs and
11349 some implementations cannot handle the current raw format. Since OpenSSL
11350 copies and compares OCSP nonces as opaque blobs without any attempt at
11351 parsing them this should not create any compatibility issues.
11355 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11356 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11357 this HMAC (and other) operations are several times slower than OpenSSL
11362 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11364 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11366 * Use the correct content when signing type "other".
11370 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11372 * Fix various bugs revealed by running the NISCC test suite:
11374 Stop out of bounds reads in the ASN1 code when presented with
11375 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11377 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11379 If verify callback ignores invalid public key errors don't try to check
11380 certificate signature with the NULL public key.
11384 * New -ignore_err option in ocsp application to stop the server
11385 exiting on the first error in a request.
11389 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11390 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11395 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11396 extra data after the compression methods not only for TLS 1.0
11397 but also for SSL 3.0 (as required by the specification).
11399 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11401 * Change X509_certificate_type() to mark the key as exported/exportable
11402 when it's 512 *bits* long, not 512 bytes.
11406 * Change AES_cbc_encrypt() so it outputs exact multiple of
11407 blocks during encryption.
11411 * Various fixes to base64 BIO and non blocking I/O. On write
11412 flushes were not handled properly if the BIO retried. On read
11413 data was not being buffered properly and had various logic bugs.
11414 This also affects blocking I/O when the data being decoded is a
11419 * Various S/MIME bugfixes and compatibility changes:
11420 output correct application/pkcs7 MIME type if
11421 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11422 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11423 of files as .eml work). Correctly handle very long lines in MIME
11428 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11430 * Countermeasure against the Klima-Pokorny-Rosa extension of
11431 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11432 a protocol version number mismatch like a decryption error
11433 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11437 * Turn on RSA blinding by default in the default implementation
11438 to avoid a timing attack. Applications that don't want it can call
11439 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11440 They would be ill-advised to do so in most cases.
11442 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11444 * Change RSA blinding code so that it works when the PRNG is not
11445 seeded (in this case, the secret RSA exponent is abused as
11446 an unpredictable seed -- if it is not unpredictable, there
11447 is no point in blinding anyway). Make RSA blinding thread-safe
11448 by remembering the creator's thread ID in rsa->blinding and
11449 having all other threads use local one-time blinding factors
11450 (this requires more computation than sharing rsa->blinding, but
11451 avoids excessive locking; and if an RSA object is not shared
11452 between threads, blinding will still be very fast).
11456 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11457 ENGINE as defaults for all supported algorithms irrespective of
11458 the 'flags' parameter. 'flags' is now honoured, so applications
11459 should make sure they are passing it correctly.
11463 * Target "mingw" now allows native Windows code to be generated in
11464 the Cygwin environment as well as with the MinGW compiler.
11468 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11470 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11471 via timing by performing a MAC computation even if incorrect
11472 block cipher padding has been found. This is a countermeasure
11473 against active attacks where the attacker has to distinguish
11474 between bad padding and a MAC verification error. ([CVE-2003-0078])
11476 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11477 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11478 Martin Vuagnoux (EPFL, Ilion)*
11480 * Make the no-err option work as intended. The intention with no-err
11481 is not to have the whole error stack handling routines removed from
11482 libcrypto, it's only intended to remove all the function name and
11483 reason texts, thereby removing some of the footprint that may not
11484 be interesting if those errors aren't displayed anyway.
11486 NOTE: it's still possible for any application or module to have its
11487 own set of error texts inserted. The routines are there, just not
11488 used by default when no-err is given.
11492 * Add support for FreeBSD on IA64.
11494 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11496 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11497 Kerberos function mit_des_cbc_cksum(). Before this change,
11498 the value returned by DES_cbc_cksum() was like the one from
11499 mit_des_cbc_cksum(), except the bytes were swapped.
11501 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11503 * Allow an application to disable the automatic SSL chain building.
11504 Before this a rather primitive chain build was always performed in
11505 ssl3_output_cert_chain(): an application had no way to send the
11506 correct chain if the automatic operation produced an incorrect result.
11508 Now the chain builder is disabled if either:
11510 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11512 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11514 The reasoning behind this is that an application would not want the
11515 auto chain building to take place if extra chain certificates are
11516 present and it might also want a means of sending no additional
11517 certificates (for example the chain has two certificates and the
11522 * Add the possibility to build without the ENGINE framework.
11524 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11526 * Under Win32 gmtime() can return NULL: check return value in
11527 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11531 * DSA routines: under certain error conditions uninitialized BN objects
11532 could be freed. Solution: make sure initialization is performed early
11533 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11534 Nils Larsch <nla@trustcenter.de> via PR#459)
11538 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11539 checked on reconnect on the client side, therefore session resumption
11540 could still fail with a "ssl session id is different" error. This
11541 behaviour is masked when SSL_OP_ALL is used due to
11542 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11543 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11544 followup to PR #377.
11548 * IA-32 assembler support enhancements: unified ELF targets, support
11549 for SCO/Caldera platforms, fix for Cygwin shared build.
11553 * Add support for FreeBSD on sparc64. As a consequence, support for
11554 FreeBSD on non-x86 processors is separate from x86 processors on
11555 the config script, much like the NetBSD support.
11557 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11559 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11561 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11564 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11565 code (06) was taken as the first octet of the session ID and the last
11566 octet was ignored consequently. As a result SSLv2 client side session
11567 caching could not have worked due to the session ID mismatch between
11569 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11574 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11575 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11580 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11581 seems that in spite of existing for more than a year, many application
11582 author have done nothing to provide the necessary callbacks, which
11583 means that this particular engine will not work properly anywhere.
11584 This is a very unfortunate situation which forces us, in the name
11585 of usability, to give the hw_ncipher.c a static lock, which is part
11587 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11588 appear in 0.9.8 or later. We EXPECT application authors to have
11589 dealt properly with this when 0.9.8 is released (unless we actually
11590 make such changes in the libcrypto locking code that changes will
11591 have to be made anyway).
11595 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11596 octets have been read, EOF or an error occurs. Without this change
11597 some truncated ASN1 structures will not produce an error.
11601 * Disable Heimdal support, since it hasn't been fully implemented.
11602 Still give the possibility to force the use of Heimdal, but with
11603 warnings and a request that patches get sent to openssl-dev.
11607 * Add the VC-CE target, introduce the WINCE sysname, and add
11608 INSTALL.WCE and appropriate conditionals to make it build.
11610 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11612 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11613 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11614 edit numbers of the version.
11616 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11618 * Introduce safe string copy and catenation functions
11619 (BUF_strlcpy() and BUF_strlcat()).
11621 *Ben Laurie (CHATS) and Richard Levitte*
11623 * Avoid using fixed-size buffers for one-line DNs.
11625 *Ben Laurie (CHATS)*
11627 * Add BUF_MEM_grow_clean() to avoid information leakage when
11628 resizing buffers containing secrets, and use where appropriate.
11630 *Ben Laurie (CHATS)*
11632 * Avoid using fixed size buffers for configuration file location.
11634 *Ben Laurie (CHATS)*
11636 * Avoid filename truncation for various CA files.
11638 *Ben Laurie (CHATS)*
11640 * Use sizeof in preference to magic numbers.
11642 *Ben Laurie (CHATS)*
11644 * Avoid filename truncation in cert requests.
11646 *Ben Laurie (CHATS)*
11648 * Add assertions to check for (supposedly impossible) buffer
11651 *Ben Laurie (CHATS)*
11653 * Don't cache truncated DNS entries in the local cache (this could
11654 potentially lead to a spoofing attack).
11656 *Ben Laurie (CHATS)*
11658 * Fix various buffers to be large enough for hex/decimal
11659 representations in a platform independent manner.
11661 *Ben Laurie (CHATS)*
11663 * Add CRYPTO_realloc_clean() to avoid information leakage when
11664 resizing buffers containing secrets, and use where appropriate.
11666 *Ben Laurie (CHATS)*
11668 * Add BIO_indent() to avoid much slightly worrying code to do
11671 *Ben Laurie (CHATS)*
11673 * Convert sprintf()/BIO_puts() to BIO_printf().
11675 *Ben Laurie (CHATS)*
11677 * buffer_gets() could terminate with the buffer only half
11680 *Ben Laurie (CHATS)*
11682 * Add assertions to prevent user-supplied crypto functions from
11683 overflowing internal buffers by having large block sizes, etc.
11685 *Ben Laurie (CHATS)*
11687 * New OPENSSL_assert() macro (similar to assert(), but enabled
11690 *Ben Laurie (CHATS)*
11692 * Eliminate unused copy of key in RC4.
11694 *Ben Laurie (CHATS)*
11696 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11698 *Ben Laurie (CHATS)*
11700 * Fix off-by-one error in EGD path.
11702 *Ben Laurie (CHATS)*
11704 * If RANDFILE path is too long, ignore instead of truncating.
11706 *Ben Laurie (CHATS)*
11708 * Eliminate unused and incorrectly sized X.509 structure
11711 *Ben Laurie (CHATS)*
11713 * Eliminate unused and dangerous function knumber().
11715 *Ben Laurie (CHATS)*
11717 * Eliminate unused and dangerous structure, KSSL_ERR.
11719 *Ben Laurie (CHATS)*
11721 * Protect against overlong session ID context length in an encoded
11722 session object. Since these are local, this does not appear to be
11725 *Ben Laurie (CHATS)*
11727 * Change from security patch (see 0.9.6e below) that did not affect
11728 the 0.9.6 release series:
11730 Remote buffer overflow in SSL3 protocol - an attacker could
11731 supply an oversized master key in Kerberos-enabled versions.
11734 *Ben Laurie (CHATS)*
11736 * Change the SSL kerb5 codes to match RFC 2712.
11740 * Make -nameopt work fully for req and add -reqopt switch.
11742 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11744 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11746 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11748 * Make sure tests can be performed even if the corresponding algorithms
11749 have been removed entirely. This was also the last step to make
11750 OpenSSL compilable with DJGPP under all reasonable conditions.
11752 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11754 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11755 to allow version independent disabling of normally unselected ciphers,
11756 which may be activated as a side-effect of selecting a single cipher.
11758 (E.g., cipher list string "RSA" enables ciphersuites that are left
11759 out of "ALL" because they do not provide symmetric encryption.
11760 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11762 *Lutz Jaenicke, Bodo Moeller*
11764 * Add appropriate support for separate platform-dependent build
11765 directories. The recommended way to make a platform-dependent
11766 build directory is the following (tested on Linux), maybe with
11769 # Place yourself outside of the OpenSSL source tree. In
11770 # this example, the environment variable OPENSSL_SOURCE
11771 # is assumed to contain the absolute OpenSSL source directory.
11772 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11773 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11774 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11775 mkdir -p `dirname $F`
11776 ln -s $OPENSSL_SOURCE/$F $F
11779 To be absolutely sure not to disturb the source tree, a "make clean"
11780 is a good thing. If it isn't successful, don't worry about it,
11781 it probably means the source directory is very clean.
11785 * Make sure any ENGINE control commands make local copies of string
11786 pointers passed to them whenever necessary. Otherwise it is possible
11787 the caller may have overwritten (or deallocated) the original string
11788 data when a later ENGINE operation tries to use the stored values.
11790 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11792 * Improve diagnostics in file reading and command-line digests.
11794 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11796 * Add AES modes CFB and OFB to the object database. Correct an
11797 error in AES-CFB decryption.
11801 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11802 allows existing EVP_CIPHER_CTX structures to be reused after
11803 calling `EVP_*Final()`. This behaviour is used by encryption
11804 BIOs and some applications. This has the side effect that
11805 applications must explicitly clean up cipher contexts with
11806 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11810 * Check the values of dna and dnb in bn_mul_recursive before calling
11811 bn_mul_comba (a non zero value means the a or b arrays do not contain
11812 n2 elements) and fallback to bn_mul_normal if either is not zero.
11816 * Fix escaping of non-ASCII characters when using the -subj option
11817 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11821 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11822 form for "surname", serialNumber has no short form.
11823 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11824 therefore remove "mail" short name for "internet 7".
11825 The OID for unique identifiers in X509 certificates is
11826 x500UniqueIdentifier, not uniqueIdentifier.
11827 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11831 * Add an "init" command to the ENGINE config module and auto initialize
11832 ENGINEs. Without any "init" command the ENGINE will be initialized
11833 after all ctrl commands have been executed on it. If init=1 the
11834 ENGINE is initialized at that point (ctrls before that point are run
11835 on the uninitialized ENGINE and after on the initialized one). If
11836 init=0 then the ENGINE will not be initialized at all.
11840 * Fix the 'app_verify_callback' interface so that the user-defined
11841 argument is actually passed to the callback: In the
11842 SSL_CTX_set_cert_verify_callback() prototype, the callback
11843 declaration has been changed from
11846 int (*cb)(X509_STORE_CTX *,void *);
11847 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11848 i=s->ctx->app_verify_callback(&ctx)
11849 has been changed into
11850 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11852 To update applications using SSL_CTX_set_cert_verify_callback(),
11853 a dummy argument can be added to their callback functions.
11855 *D. K. Smetters <smetters@parc.xerox.com>*
11857 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11859 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11861 * Add and OPENSSL_LOAD_CONF define which will cause
11862 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11863 This allows older applications to transparently support certain
11864 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11865 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11866 load the config file and OPENSSL_add_all_algorithms_conf() which will
11867 always load it have also been added.
11871 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11872 Adjust NIDs and EVP layer.
11874 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11876 * Config modules support in openssl utility.
11878 Most commands now load modules from the config file,
11879 though in a few (such as version) this isn't done
11880 because it couldn't be used for anything.
11882 In the case of ca and req the config file used is
11883 the same as the utility itself: that is the -config
11884 command line option can be used to specify an
11889 * Move default behaviour from OPENSSL_config(). If appname is NULL
11890 use "openssl_conf" if filename is NULL use default openssl config file.
11894 * Add an argument to OPENSSL_config() to allow the use of an alternative
11895 config section name. Add a new flag to tolerate a missing config file
11896 and move code to CONF_modules_load_file().
11900 * Support for crypto accelerator cards from Accelerated Encryption
11901 Processing, www.aep.ie. (Use engine 'aep')
11902 The support was copied from 0.9.6c [engine] and adapted/corrected
11903 to work with the new engine framework.
11905 *AEP Inc. and Richard Levitte*
11907 * Support for SureWare crypto accelerator cards from Baltimore
11908 Technologies. (Use engine 'sureware')
11909 The support was copied from 0.9.6c [engine] and adapted
11910 to work with the new engine framework.
11914 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11915 make the newer ENGINE framework commands for the CHIL engine work.
11917 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11919 * Make it possible to produce shared libraries on ReliantUNIX.
11921 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11923 * Add the configuration target debug-linux-ppro.
11924 Make 'openssl rsa' use the general key loading routines
11925 implemented in `apps.c`, and make those routines able to
11926 handle the key format FORMAT_NETSCAPE and the variant
11929 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11931 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11933 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11935 * Add -keyform to rsautl, and document -engine.
11937 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11939 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11940 BIO_R_NO_SUCH_FILE error code rather than the generic
11941 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11945 * Add new functions
11946 ERR_peek_last_error
11947 ERR_peek_last_error_line
11948 ERR_peek_last_error_line_data.
11949 These are similar to
11951 ERR_peek_error_line
11952 ERR_peek_error_line_data,
11953 but report on the latest error recorded rather than the first one
11954 still in the error queue.
11956 *Ben Laurie, Bodo Moeller*
11958 * default_algorithms option in ENGINE config module. This allows things
11960 default_algorithms = ALL
11961 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11965 * Preliminary ENGINE config module.
11969 * New experimental application configuration code.
11973 * Change the AES code to follow the same name structure as all other
11974 symmetric ciphers, and behave the same way. Move everything to
11975 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11977 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11979 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11981 *Ben Laurie and Theo de Raadt*
11983 * Add option to output public keys in req command.
11985 *Massimiliano Pala madwolf@openca.org*
11987 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11988 (up to about 10% better than before for P-192 and P-224).
11992 * New functions/macros
11994 SSL_CTX_set_msg_callback(ctx, cb)
11995 SSL_CTX_set_msg_callback_arg(ctx, arg)
11996 SSL_set_msg_callback(ssl, cb)
11997 SSL_set_msg_callback_arg(ssl, arg)
11999 to request calling a callback function
12001 void cb(int write_p, int version, int content_type,
12002 const void *buf, size_t len, SSL *ssl, void *arg)
12004 whenever a protocol message has been completely received
12005 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12006 protocol version according to which the SSL library interprets
12007 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12008 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12009 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12010 specification (change_cipher_spec(20), alert(21), handshake(22)).
12011 'buf' and 'len' point to the actual message, 'ssl' to the
12012 SSL object, and 'arg' is the application-defined value set by
12013 SSL[_CTX]_set_msg_callback_arg().
12015 'openssl s_client' and 'openssl s_server' have new '-msg' options
12016 to enable a callback that displays all protocol messages.
12020 * Change the shared library support so shared libraries are built as
12021 soon as the corresponding static library is finished, and thereby get
12022 openssl and the test programs linked against the shared library.
12023 This still only happens when the keyword "shard" has been given to
12024 the configuration scripts.
12026 NOTE: shared library support is still an experimental thing, and
12027 backward binary compatibility is still not guaranteed.
12029 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12031 * Add support for Subject Information Access extension.
12033 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12035 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12036 additional bytes when new memory had to be allocated, not just
12037 when reusing an existing buffer.
12041 * New command line and configuration option 'utf8' for the req command.
12042 This allows field values to be specified as UTF8 strings.
12046 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12047 runs for the former and machine-readable output for the latter.
12051 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12052 of the e-mail address in the DN (i.e., it will go into a certificate
12053 extension only). The new configuration file option 'email_in_dn = no'
12054 has the same effect.
12056 *Massimiliano Pala madwolf@openca.org*
12058 * Change all functions with names starting with `des_` to be starting
12059 with `DES_` instead. Add wrappers that are compatible with libdes,
12060 but are named `_ossl_old_des_*`. Finally, add macros that map the
12061 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12062 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12063 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12066 Since we provide two compatibility mappings, the user needs to
12067 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12068 compatibility is desired. The default (i.e., when that macro
12069 isn't defined) is OpenSSL 0.9.6c compatibility.
12071 There are also macros that enable and disable the support of old
12072 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12073 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12074 are defined, the default will apply: to support the old des routines.
12076 In either case, one must include openssl/des.h to get the correct
12077 definitions. Do not try to just include openssl/des_old.h, that
12080 NOTE: This is a major break of an old API into a new one. Software
12081 authors are encouraged to switch to the `DES_` style functions. Some
12082 time in the future, des_old.h and the libdes compatibility functions
12083 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12084 default), and then completely removed.
12088 * Test for certificates which contain unsupported critical extensions.
12089 If such a certificate is found during a verify operation it is
12090 rejected by default: this behaviour can be overridden by either
12091 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12092 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12093 X509_supported_extension() has also been added which returns 1 if a
12094 particular extension is supported.
12098 * Modify the behaviour of EVP cipher functions in similar way to digests
12099 to retain compatibility with existing code.
12103 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12104 compatibility with existing code. In particular the 'ctx' parameter does
12105 not have to be to be initialized before the call to EVP_DigestInit() and
12106 it is tidied up after a call to EVP_DigestFinal(). New function
12107 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12108 EVP_MD_CTX_copy() changed to not require the destination to be
12109 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12110 requires the destination to be valid.
12112 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12113 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12117 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12118 so that complete 'Handshake' protocol structures are kept in memory
12119 instead of overwriting 'msg_type' and 'length' with 'body' data.
12123 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12125 *Massimo Santin via Richard Levitte*
12127 * Major restructuring to the underlying ENGINE code. This includes
12128 reduction of linker bloat, separation of pure "ENGINE" manipulation
12129 (initialisation, etc) from functionality dealing with implementations
12130 of specific crypto interfaces. This change also introduces integrated
12131 support for symmetric ciphers and digest implementations - so ENGINEs
12132 can now accelerate these by providing EVP_CIPHER and EVP_MD
12133 implementations of their own. This is detailed in
12134 [crypto/engine/README.md](crypto/engine/README.md)
12135 as it couldn't be adequately described here. However, there are a few
12136 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12137 were changed in the original introduction of ENGINE code have now
12138 reverted back - the hooking from this code to ENGINE is now a good
12139 deal more passive and at run-time, operations deal directly with
12140 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12141 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12142 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12143 they were not being used by the framework as there is no concept of a
12144 BIGNUM_METHOD and they could not be generalised to the new
12145 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12146 ENGINE_cpy() has been removed as it cannot be consistently defined in
12151 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12155 * Change mkdef.pl to sort symbols that get the same entry number,
12156 and make sure the automatically generated functions `ERR_load_*`
12157 become part of libeay.num as well.
12161 * New function SSL_renegotiate_pending(). This returns true once
12162 renegotiation has been requested (either SSL_renegotiate() call
12163 or HelloRequest/ClientHello received from the peer) and becomes
12164 false once a handshake has been completed.
12165 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12166 sends a HelloRequest, but does not ensure that a handshake takes
12167 place. SSL_renegotiate_pending() is useful for checking if the
12168 client has followed the request.)
12172 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12173 By default, clients may request session resumption even during
12174 renegotiation (if session ID contexts permit); with this option,
12175 session resumption is possible only in the first handshake.
12177 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12178 more bits available for options that should not be part of
12179 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12183 * Add some demos for certificate and certificate request creation.
12187 * Make maximum certificate chain size accepted from the peer application
12188 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12189 "Douglas E. Engert" <deengert@anl.gov>.
12193 * Add support for shared libraries for Unixware-7
12194 (Boyd Lynn Gerber <gerberb@zenez.com>).
12198 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12199 be done prior to destruction. Use this to unload error strings from
12200 ENGINEs that load their own error strings. NB: This adds two new API
12201 functions to "get" and "set" this destroy handler in an ENGINE.
12205 * Alter all existing ENGINE implementations (except "openssl" and
12206 "openbsd") to dynamically instantiate their own error strings. This
12207 makes them more flexible to be built both as statically-linked ENGINEs
12208 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12209 Also, add stub code to each that makes building them as self-contained
12210 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12214 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12215 implementations into applications that are completely implemented in
12216 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12217 commands that can be used to configure what shared-library to load and
12218 to control aspects of the way it is handled. Also, made an update to
12219 the [README-Engine.md](README-Engine.md) file
12220 that brings its information up-to-date and
12221 provides some information and instructions on the "dynamic" ENGINE
12222 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12226 * Make it possible to unload ranges of ERR strings with a new
12227 "ERR_unload_strings" function.
12231 * Add a copy() function to EVP_MD.
12235 * Make EVP_MD routines take a context pointer instead of just the
12236 md_data void pointer.
12240 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12241 that the digest can only process a single chunk of data
12242 (typically because it is provided by a piece of
12243 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12244 is only going to provide a single chunk of data, and hence the
12245 framework needn't accumulate the data for oneshot drivers.
12249 * As with "ERR", make it possible to replace the underlying "ex_data"
12250 functions. This change also alters the storage and management of global
12251 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12252 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12253 index counters. The API functions that use this state have been changed
12254 to take a "class_index" rather than pointers to the class's local STACK
12255 and counter, and there is now an API function to dynamically create new
12256 classes. This centralisation allows us to (a) plug a lot of the
12257 thread-safety problems that existed, and (b) makes it possible to clean
12258 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12259 such data would previously have always leaked in application code and
12260 workarounds were in place to make the memory debugging turn a blind eye
12261 to it. Application code that doesn't use this new function will still
12262 leak as before, but their memory debugging output will announce it now
12263 rather than letting it slide.
12265 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12266 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12267 has a return value to indicate success or failure.
12271 * Make it possible to replace the underlying "ERR" functions such that the
12272 global state (2 LHASH tables and 2 locks) is only used by the "default"
12273 implementation. This change also adds two functions to "get" and "set"
12274 the implementation prior to it being automatically set the first time
12275 any other ERR function takes place. Ie. an application can call "get",
12276 pass the return value to a module it has just loaded, and that module
12277 can call its own "set" function using that value. This means the
12278 module's "ERR" operations will use (and modify) the error state in the
12279 application and not in its own statically linked copy of OpenSSL code.
12283 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12284 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12285 the operation, and provides a more encapsulated way for external code
12286 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12287 to use these functions rather than manually incrementing the counts.
12289 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12293 * Add EVP test program.
12297 * Add symmetric cipher support to ENGINE. Expect the API to change!
12301 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12302 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12303 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12304 These allow a CRL to be built without having to access X509_CRL fields
12305 directly. Modify 'ca' application to use new functions.
12309 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12310 bug workarounds. Rollback attack detection is a security feature.
12311 The problem will only arise on OpenSSL servers when TLSv1 is not
12312 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12313 Software authors not wanting to support TLSv1 will have special reasons
12314 for their choice and can explicitly enable this option.
12316 *Bodo Moeller, Lutz Jaenicke*
12318 * Rationalise EVP so it can be extended: don't include a union of
12319 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12320 (similar to those existing for EVP_CIPHER_CTX).
12325 EVP_MD_CTX_init(&md); /* new function call */
12326 EVP_DigestInit(&md, EVP_sha1());
12327 EVP_DigestUpdate(&md, in, len);
12328 EVP_DigestFinal(&md, out, NULL);
12329 EVP_MD_CTX_cleanup(&md); /* new function call */
12333 * Make DES key schedule conform to the usual scheme, as well as
12334 correcting its structure. This means that calls to DES functions
12335 now have to pass a pointer to a des_key_schedule instead of a
12336 plain des_key_schedule (which was actually always a pointer
12339 des_key_schedule ks;
12341 des_set_key_checked(..., &ks);
12342 des_ncbc_encrypt(..., &ks, ...);
12344 (Note that a later change renames 'des_...' into 'DES_...'.)
12348 * Initial reduction of linker bloat: the use of some functions, such as
12349 PEM causes large amounts of unused functions to be linked in due to
12350 poor organisation. For example pem_all.c contains every PEM function
12351 which has a knock on effect of linking in large amounts of (unused)
12352 ASN1 code. Grouping together similar functions and splitting unrelated
12353 functions prevents this.
12357 * Cleanup of EVP macros.
12361 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12362 correct `_ecb suffix`.
12366 * Add initial OCSP responder support to ocsp application. The
12367 revocation information is handled using the text based index
12368 use by the ca application. The responder can either handle
12369 requests generated internally, supplied in files (for example
12370 via a CGI script) or using an internal minimal server.
12374 * Add configuration choices to get zlib compression for TLS.
12378 * Changes to Kerberos SSL for RFC 2712 compliance:
12379 1. Implemented real KerberosWrapper, instead of just using
12380 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12381 2. Implemented optional authenticator field of KerberosWrapper.
12383 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12384 and authenticator structs; see crypto/krb5/.
12386 Generalized Kerberos calls to support multiple Kerberos libraries.
12387 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12388 via Richard Levitte*
12390 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12391 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12392 values for each of the key sizes rather than having just
12393 parameters (and 'speed' generating keys each time).
12397 * Speed up EVP routines.
12400 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12401 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12402 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12403 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12405 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12406 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12407 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12410 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12412 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12416 * Added the OS2-EMX target.
12418 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12420 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12421 New functions to support `NCONF `routines in extension code.
12422 New function `CONF_set_nconf()`
12423 to allow functions which take an `NCONF` to also handle the old `LHASH`
12424 structure: this means that the old `CONF` compatible routines can be
12425 retained (in particular w.rt. extensions) without having to duplicate the
12426 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12430 * Enhance the general user interface with mechanisms for inner control
12431 and with possibilities to have yes/no kind of prompts.
12435 * Change all calls to low level digest routines in the library and
12436 applications to use EVP. Add missing calls to HMAC_cleanup() and
12437 don't assume HMAC_CTX can be copied using memcpy().
12439 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12441 * Add the possibility to control engines through control names but with
12442 arbitrary arguments instead of just a string.
12443 Change the key loaders to take a UI_METHOD instead of a callback
12444 function pointer. NOTE: this breaks binary compatibility with earlier
12445 versions of OpenSSL [engine].
12446 Adapt the nCipher code for these new conditions and add a card insertion
12451 * Enhance the general user interface with mechanisms to better support
12452 dialog box interfaces, application-defined prompts, the possibility
12453 to use defaults (for example default passwords from somewhere else)
12454 and interrupts/cancellations.
12458 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12459 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12463 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12464 tidy up some unnecessarily weird code in 'sk_new()').
12466 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12468 * Change the key loading routines for ENGINEs to use the same kind
12469 callback (pem_password_cb) as all other routines that need this
12474 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12475 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12476 than this minimum value is recommended.
12480 * New random seeder for OpenVMS, using the system process statistics
12481 that are easily reachable.
12485 * Windows apparently can't transparently handle global
12486 variables defined in DLLs. Initialisations such as:
12488 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12490 won't compile. This is used by the any applications that need to
12491 declare their own ASN1 modules. This was fixed by adding the option
12492 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12493 needed for static libraries under Win32.
12497 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12498 setting of purpose and trust fields. New X509_STORE trust and
12499 purpose functions and tidy up setting in other SSL functions.
12503 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12504 structure. These are inherited by X509_STORE_CTX when it is
12505 initialised. This allows various defaults to be set in the
12506 X509_STORE structure (such as flags for CRL checking and custom
12507 purpose or trust settings) for functions which only use X509_STORE_CTX
12508 internally such as S/MIME.
12510 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12511 trust settings if they are not set in X509_STORE. This allows X509_STORE
12512 purposes and trust (in S/MIME for example) to override any set by default.
12514 Add command line options for CRL checking to smime, s_client and s_server
12519 * Initial CRL based revocation checking. If the CRL checking flag(s)
12520 are set then the CRL is looked up in the X509_STORE structure and
12521 its validity and signature checked, then if the certificate is found
12522 in the CRL the verify fails with a revoked error.
12524 Various new CRL related callbacks added to X509_STORE_CTX structure.
12526 Command line options added to 'verify' application to support this.
12528 This needs some additional work, such as being able to handle multiple
12529 CRLs with different times, extension based lookup (rather than just
12530 by subject name) and ultimately more complete V2 CRL extension
12535 * Add a general user interface API (crypto/ui/). This is designed
12536 to replace things like des_read_password and friends (backward
12537 compatibility functions using this new API are provided).
12538 The purpose is to remove prompting functions from the DES code
12539 section as well as provide for prompting through dialog boxes in
12540 a window system and the like.
12544 * Add "ex_data" support to ENGINE so implementations can add state at a
12545 per-structure level rather than having to store it globally.
12549 * Make it possible for ENGINE structures to be copied when retrieved by
12550 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12551 This causes the "original" ENGINE structure to act like a template,
12552 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12553 operational state can be localised to each ENGINE structure, despite the
12554 fact they all share the same "methods". New ENGINE structures returned in
12555 this case have no functional references and the return value is the single
12556 structural reference. This matches the single structural reference returned
12557 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12562 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12563 needs to match any other type at all we need to manually clear the
12568 * Changes to the "openssl engine" utility to include;
12569 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12570 about an ENGINE's available control commands.
12571 - executing control commands from command line arguments using the
12572 '-pre' and '-post' switches. '-post' is only used if '-t' is
12573 specified and the ENGINE is successfully initialised. The syntax for
12574 the individual commands are colon-separated, for example;
12575 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12579 * New dynamic control command support for ENGINEs. ENGINEs can now
12580 declare their own commands (numbers), names (strings), descriptions,
12581 and input types for run-time discovery by calling applications. A
12582 subset of these commands are implicitly classed as "executable"
12583 depending on their input type, and only these can be invoked through
12584 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12585 can be based on user input, config files, etc). The distinction is
12586 that "executable" commands cannot return anything other than a boolean
12587 result and can only support numeric or string input, whereas some
12588 discoverable commands may only be for direct use through
12589 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12590 pointers, or other custom uses. The "executable" commands are to
12591 support parameterisations of ENGINE behaviour that can be
12592 unambiguously defined by ENGINEs and used consistently across any
12593 OpenSSL-based application. Commands have been added to all the
12594 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12595 control over shared-library paths without source code alterations.
12599 * Changed all ENGINE implementations to dynamically allocate their
12600 ENGINEs rather than declaring them statically. Apart from this being
12601 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12602 this also allows the implementations to compile without using the
12603 internal engine_int.h header.
12607 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12608 'const' value. Any code that should be able to modify a RAND_METHOD
12609 should already have non-const pointers to it (ie. they should only
12610 modify their own ones).
12614 * Made a variety of little tweaks to the ENGINE code.
12615 - "atalla" and "ubsec" string definitions were moved from header files
12616 to C code. "nuron" string definitions were placed in variables
12617 rather than hard-coded - allowing parameterisation of these values
12618 later on via ctrl() commands.
12619 - Removed unused "#if 0"'d code.
12620 - Fixed engine list iteration code so it uses ENGINE_free() to release
12621 structural references.
12622 - Constified the RAND_METHOD element of ENGINE structures.
12623 - Constified various get/set functions as appropriate and added
12624 missing functions (including a catch-all ENGINE_cpy that duplicates
12625 all ENGINE values onto a new ENGINE except reference counts/state).
12626 - Removed NULL parameter checks in get/set functions. Setting a method
12627 or function to NULL is a way of cancelling out a previously set
12628 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12629 and doesn't justify the extra error symbols and code.
12630 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12631 flags from engine_int.h to engine.h.
12632 - Changed prototypes for ENGINE handler functions (init(), finish(),
12633 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12637 * Implement binary inversion algorithm for BN_mod_inverse in addition
12638 to the algorithm using long division. The binary algorithm can be
12639 used only if the modulus is odd. On 32-bit systems, it is faster
12640 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12641 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12642 up to 450 bits. In 64-bit environments, the binary algorithm
12643 appears to be advantageous for much longer moduli; here we use it
12644 for moduli up to 2048 bits.
12648 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12649 could not support the combine flag in choice fields.
12653 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12654 extensions from a certificate request to the certificate.
12658 * Allow multiple 'certopt' and 'nameopt' options to be separated
12659 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12660 file: this allows the display of the certificate about to be
12661 signed to be customised, to allow certain fields to be included
12662 or excluded and extension details. The old system didn't display
12663 multicharacter strings properly, omitted fields not in the policy
12664 and couldn't display additional details such as extensions.
12668 * Function EC_POINTs_mul for multiple scalar multiplication
12669 of an arbitrary number of elliptic curve points
12670 \sum scalars[i]*points[i],
12671 optionally including the generator defined for the EC_GROUP:
12672 scalar*generator + \sum scalars[i]*points[i].
12674 EC_POINT_mul is a simple wrapper function for the typical case
12675 that the point list has just one item (besides the optional
12680 * First EC_METHODs for curves over GF(p):
12682 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12683 operations and provides various method functions that can also
12684 operate with faster implementations of modular arithmetic.
12686 EC_GFp_mont_method() reuses most functions that are part of
12687 EC_GFp_simple_method, but uses Montgomery arithmetic.
12689 *Bodo Moeller; point addition and point doubling
12690 implementation directly derived from source code provided by
12691 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12693 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12694 crypto/ec/ec_lib.c):
12696 Curves are EC_GROUP objects (with an optional group generator)
12697 based on EC_METHODs that are built into the library.
12699 Points are EC_POINT objects based on EC_GROUP objects.
12701 Most of the framework would be able to handle curves over arbitrary
12702 finite fields, but as there are no obvious types for fields other
12703 than GF(p), some functions are limited to that for now.
12707 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12708 that the file contains a complete HTTP response.
12712 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12713 change the def and num file printf format specifier from "%-40sXXX"
12714 to "%-39s XXX". The latter will always guarantee a space after the
12715 field while the former will cause them to run together if the field
12716 is 40 of more characters long.
12720 * Constify the cipher and digest 'method' functions and structures
12721 and modify related functions to take constant EVP_MD and EVP_CIPHER
12726 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12727 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12731 * Modify `EVP_Digest*()` routines so they now return values. Although the
12732 internal software routines can never fail additional hardware versions
12737 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12739 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12740 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12743 ERR_R_NESTED_ASN1_ERROR
12745 ERR_R_MISSING_ASN1_EOS
12746 were 4 .. 9, conflicting with
12747 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12749 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12750 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12752 Add new error code 'ERR_R_INTERNAL_ERROR'.
12756 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12761 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12762 sets the subject name for a new request or supersedes the
12763 subject name in a given request. Formats that can be parsed are
12764 'CN=Some Name, OU=myOU, C=IT'
12766 'CN=Some Name/OU=myOU/C=IT'.
12768 Add options '-batch' and '-verbose' to 'openssl req'.
12770 *Massimiliano Pala <madwolf@hackmasters.net>*
12772 * Introduce the possibility to access global variables through
12773 functions on platform were that's the best way to handle exporting
12774 global variables in shared libraries. To enable this functionality,
12775 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12776 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12777 is normally done by Configure or something similar).
12779 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12780 in the source file (foo.c) like this:
12782 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12783 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12785 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12786 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12788 OPENSSL_DECLARE_GLOBAL(int,foo);
12789 #define foo OPENSSL_GLOBAL_REF(foo)
12790 OPENSSL_DECLARE_GLOBAL(double,bar);
12791 #define bar OPENSSL_GLOBAL_REF(bar)
12793 The #defines are very important, and therefore so is including the
12794 header file everywhere where the defined globals are used.
12796 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12797 of ASN.1 items, but that structure is a bit different.
12799 The largest change is in util/mkdef.pl which has been enhanced with
12800 better and easier to understand logic to choose which symbols should
12801 go into the Windows .def files as well as a number of fixes and code
12802 cleanup (among others, algorithm keywords are now sorted
12803 lexicographically to avoid constant rewrites).
12807 * In BN_div() keep a copy of the sign of 'num' before writing the
12808 result to 'rm' because if rm==num the value will be overwritten
12809 and produce the wrong result if 'num' is negative: this caused
12810 problems with BN_mod() and BN_nnmod().
12814 * Function OCSP_request_verify(). This checks the signature on an
12815 OCSP request and verifies the signer certificate. The signer
12816 certificate is just checked for a generic purpose and OCSP request
12821 * Add OCSP_check_validity() function to check the validity of OCSP
12822 responses. OCSP responses are prepared in real time and may only
12823 be a few seconds old. Simply checking that the current time lies
12824 between thisUpdate and nextUpdate max reject otherwise valid responses
12825 caused by either OCSP responder or client clock inaccuracy. Instead
12826 we allow thisUpdate and nextUpdate to fall within a certain period of
12827 the current time. The age of the response can also optionally be
12828 checked. Two new options -validity_period and -status_age added to
12833 * If signature or public key algorithm is unrecognized print out its
12834 OID rather that just UNKNOWN.
12838 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12839 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12840 ID to be generated from the issuer certificate alone which can then be
12841 passed to OCSP_id_issuer_cmp().
12845 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12846 ASN1 modules to export functions returning ASN1_ITEM pointers
12847 instead of the ASN1_ITEM structures themselves. This adds several
12848 new macros which allow the underlying ASN1 function/structure to
12849 be accessed transparently. As a result code should not use ASN1_ITEM
12850 references directly (such as &X509_it) but instead use the relevant
12851 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12852 use of the new ASN1 code on platforms where exporting structures
12853 is problematical (for example in shared libraries) but exporting
12854 functions returning pointers to structures is not.
12858 * Add support for overriding the generation of SSL/TLS session IDs.
12859 These callbacks can be registered either in an SSL_CTX or per SSL.
12860 The purpose of this is to allow applications to control, if they wish,
12861 the arbitrary values chosen for use as session IDs, particularly as it
12862 can be useful for session caching in multiple-server environments. A
12863 command-line switch for testing this (and any client code that wishes
12864 to use such a feature) has been added to "s_server".
12866 *Geoff Thorpe, Lutz Jaenicke*
12868 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12869 of the form `#if defined(...) || defined(...) || ...` and
12870 `#if !defined(...) && !defined(...) && ...`. This also avoids
12871 the growing number of special cases it was previously handling.
12875 * Make all configuration macros available for application by making
12876 sure they are available in opensslconf.h, by giving them names starting
12877 with `OPENSSL_` to avoid conflicts with other packages and by making
12878 sure e_os2.h will cover all platform-specific cases together with
12880 Additionally, it is now possible to define configuration/platform-
12881 specific names (called "system identities"). In the C code, these
12882 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12883 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12884 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12889 * New option -set_serial to 'req' and 'x509' this allows the serial
12890 number to use to be specified on the command line. Previously self
12891 signed certificates were hard coded with serial number 0 and the
12892 CA options of 'x509' had to use a serial number in a file which was
12897 * New options to 'ca' utility to support V2 CRL entry extensions.
12898 Currently CRL reason, invalidity date and hold instruction are
12899 supported. Add new CRL extensions to V3 code and some new objects.
12903 * New function EVP_CIPHER_CTX_set_padding() this is used to
12904 disable standard block padding (aka PKCS#5 padding) in the EVP
12905 API, which was previously mandatory. This means that the data is
12906 not padded in any way and so the total length much be a multiple
12907 of the block size, otherwise an error occurs.
12911 * Initial (incomplete) OCSP SSL support.
12915 * New function OCSP_parse_url(). This splits up a URL into its host,
12916 port and path components: primarily to parse OCSP URLs. New -url
12917 option to ocsp utility.
12921 * New nonce behavior. The return value of OCSP_check_nonce() now
12922 reflects the various checks performed. Applications can decide
12923 whether to tolerate certain situations such as an absent nonce
12924 in a response when one was present in a request: the ocsp application
12925 just prints out a warning. New function OCSP_add1_basic_nonce()
12926 this is to allow responders to include a nonce in a response even if
12927 the request is nonce-less.
12931 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12932 skipped when using openssl x509 multiple times on a single input file,
12933 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12937 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12938 set string type: to handle setting ASN1_TIME structures. Fix ca
12939 utility to correctly initialize revocation date of CRLs.
12943 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12944 the clients preferred ciphersuites and rather use its own preferences.
12945 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12946 Internet Explorer by ensuring unchanged hash method during stepup.
12947 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12951 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12952 to aes and add a new 'exist' option to print out symbols that don't
12957 * Additional options to ocsp utility to allow flags to be set and
12958 additional certificates supplied.
12962 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12963 OCSP client a number of certificate to only verify the response
12968 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12969 handle the new API. Currently only ECB, CBC modes supported. Add new
12972 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12973 Encryption Standard (AES) Ciphersuites for Transport Layer
12974 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12975 not enabled by default and were not part of the "ALL" ciphersuite
12976 alias because they were not yet official; they could be
12977 explicitly requested by specifying the "AESdraft" ciphersuite
12978 group alias. In the final release of OpenSSL 0.9.7, the group
12979 alias is called "AES" and is part of "ALL".)
12981 *Ben Laurie, Steve Henson, Bodo Moeller*
12983 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12984 request to response.
12988 * Functions for OCSP responders. OCSP_request_onereq_count(),
12989 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12990 extract information from a certificate request. OCSP_response_create()
12991 creates a response and optionally adds a basic response structure.
12992 OCSP_basic_add1_status() adds a complete single response to a basic
12993 response and returns the OCSP_SINGLERESP structure just added (to allow
12994 extensions to be included for example). OCSP_basic_add1_cert() adds a
12995 certificate to a basic response and OCSP_basic_sign() signs a basic
12996 response with various flags. New helper functions ASN1_TIME_check()
12997 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12998 (converts ASN1_TIME to GeneralizedTime).
13002 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13003 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13004 structure from a certificate. X509_pubkey_digest() digests the public_key
13005 contents: this is used in various key identifiers.
13009 * Make sk_sort() tolerate a NULL argument.
13011 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13013 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13014 passed by the function are trusted implicitly. If any of them signed the
13015 response then it is assumed to be valid and is not verified.
13019 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13020 to data. This was previously part of the PKCS7 ASN1 code. This
13021 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13022 *Steve Henson, reported by Kenneth R. Robinette
13023 <support@securenetterm.com>*
13025 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13026 routines: without these tracing memory leaks is very painful.
13027 Fix leaks in PKCS12 and PKCS7 routines.
13031 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13032 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13033 effectively meant GeneralizedTime would never be used. Now it
13034 is initialised to -1 but X509_time_adj() now has to check the value
13035 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13036 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13037 *Steve Henson, reported by Kenneth R. Robinette
13038 <support@securenetterm.com>*
13040 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13041 result in a zero length in the ASN1_INTEGER structure which was
13042 not consistent with the structure when d2i_ASN1_INTEGER() was used
13043 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13044 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13045 where it did not print out a minus for negative ASN1_INTEGER.
13049 * Add summary printout to ocsp utility. The various functions which
13050 convert status values to strings have been renamed to:
13051 OCSP_response_status_str(), OCSP_cert_status_str() and
13052 OCSP_crl_reason_str() and are no longer static. New options
13053 to verify nonce values and to disable verification. OCSP response
13054 printout format cleaned up.
13058 * Add additional OCSP certificate checks. These are those specified
13059 in RFC2560. This consists of two separate checks: the CA of the
13060 certificate being checked must either be the OCSP signer certificate
13061 or the issuer of the OCSP signer certificate. In the latter case the
13062 OCSP signer certificate must contain the OCSP signing extended key
13063 usage. This check is performed by attempting to match the OCSP
13064 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13065 in the OCSP_CERTID structures of the response.
13069 * Initial OCSP certificate verification added to OCSP_basic_verify()
13070 and related routines. This uses the standard OpenSSL certificate
13071 verify routines to perform initial checks (just CA validity) and
13072 to obtain the certificate chain. Then additional checks will be
13073 performed on the chain. Currently the root CA is checked to see
13074 if it is explicitly trusted for OCSP signing. This is used to set
13075 a root CA as a global signing root: that is any certificate that
13076 chains to that CA is an acceptable OCSP signing certificate.
13080 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13081 extensions from a separate configuration file.
13082 As when reading extensions from the main configuration file,
13083 the '-extensions ...' option may be used for specifying the
13086 *Massimiliano Pala <madwolf@comune.modena.it>*
13088 * New OCSP utility. Allows OCSP requests to be generated or
13089 read. The request can be sent to a responder and the output
13090 parsed, outputted or printed in text form. Not complete yet:
13091 still needs to check the OCSP response validity.
13095 * New subcommands for 'openssl ca':
13096 `openssl ca -status <serial>` prints the status of the cert with
13097 the given serial number (according to the index file).
13098 `openssl ca -updatedb` updates the expiry status of certificates
13101 *Massimiliano Pala <madwolf@comune.modena.it>*
13103 * New '-newreq-nodes' command option to CA.pl. This is like
13104 '-newreq', but calls 'openssl req' with the '-nodes' option
13105 so that the resulting key is not encrypted.
13107 *Damien Miller <djm@mindrot.org>*
13109 * New configuration for the GNU Hurd.
13111 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13113 * Initial code to implement OCSP basic response verify. This
13114 is currently incomplete. Currently just finds the signer's
13115 certificate and verifies the signature on the response.
13119 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13120 value of OPENSSLDIR. This is available via the new '-d' option
13121 to 'openssl version', and is also included in 'openssl version -a'.
13125 * Allowing defining memory allocation callbacks that will be given
13126 file name and line number information in additional arguments
13127 (a `const char*` and an int). The basic functionality remains, as
13128 well as the original possibility to just replace malloc(),
13129 realloc() and free() by functions that do not know about these
13130 additional arguments. To register and find out the current
13131 settings for extended allocation functions, the following
13132 functions are provided:
13134 CRYPTO_set_mem_ex_functions
13135 CRYPTO_set_locked_mem_ex_functions
13136 CRYPTO_get_mem_ex_functions
13137 CRYPTO_get_locked_mem_ex_functions
13139 These work the same way as CRYPTO_set_mem_functions and friends.
13140 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13141 extended allocation function is enabled.
13142 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13143 a conventional allocation function is enabled.
13145 *Richard Levitte, Bodo Moeller*
13147 * Finish off removing the remaining LHASH function pointer casts.
13148 There should no longer be any prototype-casting required when using
13149 the LHASH abstraction, and any casts that remain are "bugs". See
13150 the callback types and macros at the head of lhash.h for details
13151 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13155 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13156 If /dev/[u]random devices are not available or do not return enough
13157 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13159 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13160 /etc/entropy will be queried once each in this sequence, querying stops
13161 when enough entropy was collected without querying more sockets.
13165 * Change the Unix RAND_poll() variant to be able to poll several
13166 random devices, as specified by DEVRANDOM, until a sufficient amount
13167 of data has been collected. We spend at most 10 ms on each file
13168 (select timeout) and read in non-blocking mode. DEVRANDOM now
13169 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13170 (previously it was just the string "/dev/urandom"), so on typical
13171 platforms the 10 ms delay will never occur.
13172 Also separate out the Unix variant to its own file, rand_unix.c.
13173 For VMS, there's a currently-empty rand_vms.c.
13177 * Move OCSP client related routines to ocsp_cl.c. These
13178 provide utility functions which an application needing
13179 to issue a request to an OCSP responder and analyse the
13180 response will typically need: as opposed to those which an
13181 OCSP responder itself would need which will be added later.
13183 OCSP_request_sign() signs an OCSP request with an API similar
13184 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13185 response. OCSP_response_get1_basic() extracts basic response
13186 from response. OCSP_resp_find_status(): finds and extracts status
13187 information from an OCSP_CERTID structure (which will be created
13188 when the request structure is built). These are built from lower
13189 level functions which work on OCSP_SINGLERESP structures but
13190 won't normally be used unless the application wishes to examine
13191 extensions in the OCSP response for example.
13193 Replace nonce routines with a pair of functions.
13194 OCSP_request_add1_nonce() adds a nonce value and optionally
13195 generates a random value. OCSP_check_nonce() checks the
13196 validity of the nonce in an OCSP response.
13200 * Change function OCSP_request_add() to OCSP_request_add0_id().
13201 This doesn't copy the supplied OCSP_CERTID and avoids the
13202 need to free up the newly created id. Change return type
13203 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13204 This can then be used to add extensions to the request.
13205 Deleted OCSP_request_new(), since most of its functionality
13206 is now in OCSP_REQUEST_new() (and the case insensitive name
13207 clash) apart from the ability to set the request name which
13208 will be added elsewhere.
13212 * Update OCSP API. Remove obsolete extensions argument from
13213 various functions. Extensions are now handled using the new
13214 OCSP extension code. New simple OCSP HTTP function which
13215 can be used to send requests and parse the response.
13219 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13220 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13221 uses the special reorder version of SET OF to sort the attributes
13222 and reorder them to match the encoded order. This resolves a long
13223 standing problem: a verify on a PKCS7 structure just after signing
13224 it used to fail because the attribute order did not match the
13225 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13226 it uses the received order. This is necessary to tolerate some broken
13227 software that does not order SET OF. This is handled by encoding
13228 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13229 to produce the required SET OF.
13233 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13234 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13235 files to get correct declarations of the ASN.1 item variables.
13239 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13240 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13241 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13242 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13243 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13244 ASN1_ITEM and no wrapper functions.
13248 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13249 replace the old function pointer based I/O routines. Change most of
13250 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13254 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13255 lines, recognize more "algorithms" that can be deselected, and make
13256 it complain about algorithm deselection that isn't recognised.
13260 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13261 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13262 to use new functions. Add NO_ASN1_OLD which can be set to remove
13263 some old style ASN1 functions: this can be used to determine if old
13264 code will still work when these eventually go away.
13268 * New extension functions for OCSP structures, these follow the
13269 same conventions as certificates and CRLs.
13273 * New function X509V3_add1_i2d(). This automatically encodes and
13274 adds an extension. Its behaviour can be customised with various
13275 flags to append, replace or delete. Various wrappers added for
13276 certificates and CRLs.
13280 * Fix to avoid calling the underlying ASN1 print routine when
13281 an extension cannot be parsed. Correct a typo in the
13282 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13286 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13287 entries for variables.
13291 * Add functionality to `apps/openssl.c` for detecting locking
13292 problems: As the program is single-threaded, all we have
13293 to do is register a locking callback using an array for
13294 storing which locks are currently held by the program.
13298 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13299 SSL_get_ex_data_X509_STORE_idx(), which is used in
13300 ssl_verify_cert_chain() and thus can be called at any time
13301 during TLS/SSL handshakes so that thread-safety is essential.
13302 Unfortunately, the ex_data design is not at all suited
13303 for multi-threaded use, so it probably should be abolished.
13307 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13309 *Broadcom, tweaked and integrated by Geoff Thorpe*
13311 * Move common extension printing code to new function
13312 X509V3_print_extensions(). Reorganise OCSP print routines and
13313 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13317 * New function X509_signature_print() to remove duplication in some
13322 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13323 set (this was treated exactly the same as SET OF previously). This
13324 is used to reorder the STACK representing the structure to match the
13325 encoding. This will be used to get round a problem where a PKCS7
13326 structure which was signed could not be verified because the STACK
13327 order did not reflect the encoded order.
13331 * Reimplement the OCSP ASN1 module using the new code.
13335 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13336 for its ASN1 operations. The old style function pointers still exist
13337 for now but they will eventually go away.
13341 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13342 completely replaces the old ASN1 functionality with a table driven
13343 encoder and decoder which interprets an ASN1_ITEM structure describing
13344 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13345 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13346 has also been converted to the new form.
13350 * Change BN_mod_exp_recp so that negative moduli are tolerated
13351 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13352 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13353 for negative moduli.
13357 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13358 of not touching the result's sign bit.
13362 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13367 * Changed the LHASH code to use prototypes for callbacks, and created
13368 macros to declare and implement thin (optionally static) functions
13369 that provide type-safety and avoid function pointer casting for the
13370 type-specific callbacks.
13374 * Added Kerberos Cipher Suites to be used with TLS, as written in
13376 *Veers Staats <staatsvr@asc.hpc.mil>,
13377 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13379 * Reformat the FAQ so the different questions and answers can be divided
13380 in sections depending on the subject.
13384 * Have the zlib compression code load ZLIB.DLL dynamically under
13389 * New function BN_mod_sqrt for computing square roots modulo a prime
13390 (using the probabilistic Tonelli-Shanks algorithm unless
13391 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13392 be handled deterministically).
13394 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13396 * Make BN_mod_inverse faster by explicitly handling small quotients
13397 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13398 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13402 * New function BN_kronecker.
13406 * Fix BN_gcd so that it works on negative inputs; the result is
13407 positive unless both parameters are zero.
13408 Previously something reasonably close to an infinite loop was
13409 possible because numbers could be growing instead of shrinking
13410 in the implementation of Euclid's algorithm.
13414 * Fix BN_is_word() and BN_is_one() macros to take into account the
13415 sign of the number in question.
13417 Fix BN_is_word(a,w) to work correctly for w == 0.
13419 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13420 because its test if the absolute value of 'a' equals 'w'.
13421 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13422 it exists mostly for use in the implementations of BN_is_zero(),
13423 BN_is_one(), and BN_is_word().
13427 * New function BN_swap.
13431 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13432 the exponentiation functions are more likely to produce reasonable
13433 results on negative inputs.
13437 * Change BN_mod_mul so that the result is always non-negative.
13438 Previously, it could be negative if one of the factors was negative;
13439 I don't think anyone really wanted that behaviour.
13443 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13444 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13445 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13446 and add new functions:
13455 BN_mod_lshift1_quick
13457 BN_mod_lshift_quick
13459 These functions always generate non-negative results.
13461 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13462 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13464 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13465 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13466 be reduced modulo `m`.
13468 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13471 The following entry accidentally appeared in the CHANGES file
13472 distributed with OpenSSL 0.9.7. The modifications described in
13473 it do *not* apply to OpenSSL 0.9.7.
13475 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13476 was actually never needed) and in BN_mul(). The removal in BN_mul()
13477 required a small change in bn_mul_part_recursive() and the addition
13478 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13479 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13480 bn_sub_words() and bn_add_words() except they take arrays with
13486 * In 'openssl passwd', verify passwords read from the terminal
13487 unless the '-salt' option is used (which usually means that
13488 verification would just waste user's time since the resulting
13489 hash is going to be compared with some given password hash)
13490 or the new '-noverify' option is used.
13492 This is an incompatible change, but it does not affect
13493 non-interactive use of 'openssl passwd' (passwords on the command
13494 line, '-stdin' option, '-in ...' option) and thus should not
13495 cause any problems.
13499 * Remove all references to RSAref, since there's no more need for it.
13503 * Make DSO load along a path given through an environment variable
13504 (SHLIB_PATH) with shl_load().
13508 * Constify the ENGINE code as a result of BIGNUM constification.
13509 Also constify the RSA code and most things related to it. In a
13510 few places, most notable in the depth of the ASN.1 code, ugly
13511 casts back to non-const were required (to be solved at a later
13516 * Make it so the openssl application has all engines loaded by default.
13520 * Constify the BIGNUM routines a little more.
13524 * Add the following functions:
13526 ENGINE_load_cswift()
13528 ENGINE_load_atalla()
13529 ENGINE_load_nuron()
13530 ENGINE_load_builtin_engines()
13532 That way, an application can itself choose if external engines that
13533 are built-in in OpenSSL shall ever be used or not. The benefit is
13534 that applications won't have to be linked with libdl or other dso
13535 libraries unless it's really needed.
13537 Changed 'openssl engine' to load all engines on demand.
13538 Changed the engine header files to avoid the duplication of some
13539 declarations (they differed!).
13543 * 'openssl engine' can now list capabilities.
13547 * Better error reporting in 'openssl engine'.
13551 * Never call load_dh_param(NULL) in s_server.
13555 * Add engine application. It can currently list engines by name and
13556 identity, and test if they are actually available.
13560 * Improve RPM specification file by forcing symbolic linking and making
13561 sure the installed documentation is also owned by root.root.
13563 *Damien Miller <djm@mindrot.org>*
13565 * Give the OpenSSL applications more possibilities to make use of
13566 keys (public as well as private) handled by engines.
13570 * Add OCSP code that comes from CertCo.
13574 * Add VMS support for the Rijndael code.
13578 * Added untested support for Nuron crypto accelerator.
13582 * Add support for external cryptographic devices. This code was
13583 previously distributed separately as the "engine" branch.
13585 *Geoff Thorpe, Richard Levitte*
13587 * Rework the filename-translation in the DSO code. It is now possible to
13588 have far greater control over how a "name" is turned into a filename
13589 depending on the operating environment and any oddities about the
13590 different shared library filenames on each system.
13594 * Support threads on FreeBSD-elf in Configure.
13598 * Fix for SHA1 assembly problem with MASM: it produces
13599 warnings about corrupt line number information when assembling
13600 with debugging information. This is caused by the overlapping
13603 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13606 NCONF_get_number() has no error checking at all. As a replacement,
13607 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13608 promoted strongly. The old NCONF_get_number is kept around for
13609 binary backward compatibility.
13610 Make it possible for methods to load from something other than a BIO,
13611 by providing a function pointer that is given a name instead of a BIO.
13612 For example, this could be used to load configuration data from an
13617 * Fix for non blocking accept BIOs. Added new I/O special reason
13618 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13619 with non blocking I/O was not possible because no retry code was
13620 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13625 * Added the beginnings of Rijndael support.
13629 * Fix for bug in DirectoryString mask setting. Add support for
13630 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13631 to allow certificate printing to more controllable, additional
13632 'certopt' option to 'x509' to allow new printing options to be
13637 * Clean old EAY MD5 hack from e_os.h.
13641 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13643 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13644 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13646 *Joe Orton, Steve Henson*
13648 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13650 * Fix additional bug revealed by the NISCC test suite:
13652 Stop bug triggering large recursion when presented with
13653 certain ASN.1 tags ([CVE-2003-0851])
13657 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13659 * Fix various bugs revealed by running the NISCC test suite:
13661 Stop out of bounds reads in the ASN1 code when presented with
13662 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13664 If verify callback ignores invalid public key errors don't try to check
13665 certificate signature with the NULL public key.
13669 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13670 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13675 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13676 extra data after the compression methods not only for TLS 1.0
13677 but also for SSL 3.0 (as required by the specification).
13679 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13681 * Change X509_certificate_type() to mark the key as exported/exportable
13682 when it's 512 *bits* long, not 512 bytes.
13686 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13688 * Countermeasure against the Klima-Pokorny-Rosa extension of
13689 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13690 a protocol version number mismatch like a decryption error
13691 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13695 * Turn on RSA blinding by default in the default implementation
13696 to avoid a timing attack. Applications that don't want it can call
13697 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13698 They would be ill-advised to do so in most cases.
13700 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13702 * Change RSA blinding code so that it works when the PRNG is not
13703 seeded (in this case, the secret RSA exponent is abused as
13704 an unpredictable seed -- if it is not unpredictable, there
13705 is no point in blinding anyway). Make RSA blinding thread-safe
13706 by remembering the creator's thread ID in rsa->blinding and
13707 having all other threads use local one-time blinding factors
13708 (this requires more computation than sharing rsa->blinding, but
13709 avoids excessive locking; and if an RSA object is not shared
13710 between threads, blinding will still be very fast).
13714 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13716 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13717 via timing by performing a MAC computation even if incorrect
13718 block cipher padding has been found. This is a countermeasure
13719 against active attacks where the attacker has to distinguish
13720 between bad padding and a MAC verification error. ([CVE-2003-0078])
13722 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13723 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13724 Martin Vuagnoux (EPFL, Ilion)*
13726 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13728 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13729 memory from its contents. This is done with a counter that will
13730 place alternating values in each byte. This can be used to solve
13731 two issues: 1) the removal of calls to memset() by highly optimizing
13732 compilers, and 2) cleansing with other values than 0, since those can
13733 be read through on certain media, for example a swap space on disk.
13737 * Bugfix: client side session caching did not work with external caching,
13738 because the session->cipher setting was not restored when reloading
13739 from the external cache. This problem was masked, when
13740 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13741 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13745 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13746 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13748 *Zeev Lieber <zeev-l@yahoo.com>*
13750 * Undo an undocumented change introduced in 0.9.6e which caused
13751 repeated calls to OpenSSL_add_all_ciphers() and
13752 OpenSSL_add_all_digests() to be ignored, even after calling
13757 * Change the default configuration reader to deal with last line not
13758 being properly terminated.
13762 * Change X509_NAME_cmp() so it applies the special rules on handling
13763 DN values that are of type PrintableString, as well as RDNs of type
13764 emailAddress where the value has the type ia5String.
13766 *stefank@valicert.com via Richard Levitte*
13768 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13769 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13770 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13771 the bitwise-OR of the two for use by the majority of applications
13772 wanting this behaviour, and update the docs. The documented
13773 behaviour and actual behaviour were inconsistent and had been
13774 changing anyway, so this is more a bug-fix than a behavioural
13777 *Geoff Thorpe, diagnosed by Nadav Har'El*
13779 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13780 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13784 * Fix initialization code race conditions in
13785 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13786 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13787 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13788 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13789 ssl2_get_cipher_by_char(),
13790 ssl3_get_cipher_by_char().
13792 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13794 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13795 the cached sessions are flushed, as the remove_cb() might use ex_data
13796 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13797 (see [openssl.org #212]).
13799 *Geoff Thorpe, Lutz Jaenicke*
13801 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13802 length, instead of the encoding length to d2i_ASN1_OBJECT.
13806 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13808 * [In 0.9.6g-engine release:]
13809 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13811 *Lynn Gazis <lgazis@rainbow.com>*
13813 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13815 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13816 and get fix the header length calculation.
13817 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13818 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13820 * Use proper error handling instead of 'assertions' in buffer
13821 overflow checks added in 0.9.6e. This prevents DoS (the
13822 assertions could call abort()).
13824 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13826 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13828 * Add various sanity checks to asn1_get_length() to reject
13829 the ASN1 length bytes if they exceed sizeof(long), will appear
13830 negative or the content length exceeds the length of the
13833 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13835 * Fix cipher selection routines: ciphers without encryption had no flags
13836 for the cipher strength set and where therefore not handled correctly
13837 by the selection routines (PR #130).
13841 * Fix EVP_dsa_sha macro.
13846 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13847 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13848 that was added in OpenSSL 0.9.6d.
13850 As the countermeasure turned out to be incompatible with some
13851 broken SSL implementations, the new option is part of SSL_OP_ALL.
13852 SSL_OP_ALL is usually employed when compatibility with weird SSL
13853 implementations is desired (e.g. '-bugs' option to 's_client' and
13854 's_server'), so the new option is automatically set in many
13859 * Changes in security patch:
13861 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13862 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13863 Air Force Materiel Command, USAF, under agreement number
13866 * Add various sanity checks to asn1_get_length() to reject
13867 the ASN1 length bytes if they exceed sizeof(long), will appear
13868 negative or the content length exceeds the length of the
13869 supplied buffer. ([CVE-2002-0659])
13871 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13873 * Assertions for various potential buffer overflows, not known to
13874 happen in practice.
13876 *Ben Laurie (CHATS)*
13878 * Various temporary buffers to hold ASCII versions of integers were
13879 too small for 64 bit platforms. ([CVE-2002-0655])
13880 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13882 * Remote buffer overflow in SSL3 protocol - an attacker could
13883 supply an oversized session ID to a client. ([CVE-2002-0656])
13885 *Ben Laurie (CHATS)*
13887 * Remote buffer overflow in SSL2 protocol - an attacker could
13888 supply an oversized client master key. ([CVE-2002-0656])
13890 *Ben Laurie (CHATS)*
13892 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13894 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13895 encoded as NULL) with id-dsa-with-sha1.
13897 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13899 * Check various `X509_...()` return values in `apps/req.c`.
13901 *Nils Larsch <nla@trustcenter.de>*
13903 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13904 an end-of-file condition would erroneously be flagged, when the CRLF
13905 was just at the end of a processed block. The bug was discovered when
13906 processing data through a buffering memory BIO handing the data to a
13907 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13908 <ptsekov@syntrex.com> and Nedelcho Stanev.
13912 * Implement a countermeasure against a vulnerability recently found
13913 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13914 before application data chunks to avoid the use of known IVs
13915 with data potentially chosen by the attacker.
13919 * Fix length checks in ssl3_get_client_hello().
13923 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13924 to prevent ssl3_read_internal() from incorrectly assuming that
13925 ssl3_read_bytes() found application data while handshake
13926 processing was enabled when in fact s->s3->in_read_app_data was
13927 merely automatically cleared during the initial handshake.
13929 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13931 * Fix object definitions for Private and Enterprise: they were not
13932 recognized in their shortname (=lowercase) representation. Extend
13933 obj_dat.pl to issue an error when using undefined keywords instead
13934 of silently ignoring the problem (Svenning Sorensen
13935 <sss@sss.dnsalias.net>).
13939 * Fix DH_generate_parameters() so that it works for 'non-standard'
13940 generators, i.e. generators other than 2 and 5. (Previously, the
13941 code did not properly initialise the 'add' and 'rem' values to
13942 BN_generate_prime().)
13944 In the new general case, we do not insist that 'generator' is
13945 actually a primitive root: This requirement is rather pointless;
13946 a generator of the order-q subgroup is just as good, if not
13951 * Map new X509 verification errors to alerts. Discovered and submitted by
13952 Tom Wu <tom@arcot.com>.
13956 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13957 returning non-zero before the data has been completely received
13958 when using non-blocking I/O.
13960 *Bodo Moeller; problem pointed out by John Hughes*
13962 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13964 *Ben Laurie, Lutz Jaenicke*
13966 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13967 Yoram Zahavi <YoramZ@gilian.com>).
13971 * Add information about CygWin 1.3 and on, and preserve proper
13972 configuration for the versions before that.
13974 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13976 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13977 check whether we deal with a copy of a session and do not delete from
13978 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13979 <izhar@checkpoint.com>.
13983 * Do not store session data into the internal session cache, if it
13984 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13985 flag is set). Proposed by Aslam <aslam@funk.com>.
13989 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13994 * [In 0.9.6d-engine release:]
13995 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13997 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13999 * Add the configuration target linux-s390x.
14001 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14003 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14004 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14005 variable as an indication that a ClientHello message has been
14006 received. As the flag value will be lost between multiple
14007 invocations of ssl3_accept when using non-blocking I/O, the
14008 function may not be aware that a handshake has actually taken
14009 place, thus preventing a new session from being added to the
14012 To avoid this problem, we now set s->new_session to 2 instead of
14013 using a local variable.
14015 *Lutz Jaenicke, Bodo Moeller*
14017 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14018 if the SSL_R_LENGTH_MISMATCH error is detected.
14020 *Geoff Thorpe, Bodo Moeller*
14022 * New 'shared_ldflag' column in Configure platform table.
14026 * Fix EVP_CIPHER_mode macro.
14028 *"Dan S. Camper" <dan@bti.net>*
14030 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14031 type, we must throw them away by setting rr->length to 0.
14033 *D P Chang <dpc@qualys.com>*
14035 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14037 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14038 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14039 worked incorrectly for those cases where range = `10..._2` and
14040 `3*range` is two bits longer than range.)
14044 * Only add signing time to PKCS7 structures if it is not already
14049 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14050 OBJ_ld_ce should be OBJ_id_ce.
14051 Also some ip-pda OIDs in crypto/objects/objects.txt were
14052 incorrect (cf. RFC 3039).
14054 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14056 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14057 returns early because it has nothing to do.
14059 *Andy Schneider <andy.schneider@bjss.co.uk>*
14061 * [In 0.9.6c-engine release:]
14062 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14064 *Andy Schneider <andy.schneider@bjss.co.uk>*
14066 * [In 0.9.6c-engine release:]
14067 Add support for Cryptographic Appliance's keyserver technology.
14068 (Use engine 'keyclient')
14070 *Cryptographic Appliances and Geoff Thorpe*
14072 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14073 is called via tools/c89.sh because arguments have to be
14074 rearranged (all '-L' options must appear before the first object
14077 *Richard Shapiro <rshapiro@abinitio.com>*
14079 * [In 0.9.6c-engine release:]
14080 Add support for Broadcom crypto accelerator cards, backported
14083 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14085 * [In 0.9.6c-engine release:]
14086 Add support for SureWare crypto accelerator cards from
14087 Baltimore Technologies. (Use engine 'sureware')
14089 *Baltimore Technologies and Mark Cox*
14091 * [In 0.9.6c-engine release:]
14092 Add support for crypto accelerator cards from Accelerated
14093 Encryption Processing, www.aep.ie. (Use engine 'aep')
14095 *AEP Inc. and Mark Cox*
14097 * Add a configuration entry for gcc on UnixWare.
14099 *Gary Benson <gbenson@redhat.com>*
14101 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14102 messages are stored in a single piece (fixed-length part and
14103 variable-length part combined) and fix various bugs found on the way.
14107 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14108 instead. BIO_gethostbyname() does not know what timeouts are
14109 appropriate, so entries would stay in cache even when they have
14111 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14113 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14114 faced with a pathologically small ClientHello fragment that does
14115 not contain client_version: Instead of aborting with an error,
14116 simply choose the highest available protocol version (i.e.,
14117 TLS 1.0 unless it is disabled). In practice, ClientHello
14118 messages are never sent like this, but this change gives us
14119 strictly correct behaviour at least for TLS.
14123 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14124 never resets s->method to s->ctx->method when called from within
14125 one of the SSL handshake functions.
14127 *Bodo Moeller; problem pointed out by Niko Baric*
14129 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14130 (sent using the client's version number) if client_version is
14131 smaller than the protocol version in use. Also change
14132 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14133 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14134 the client will at least see that alert.
14138 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14143 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14144 client receives HelloRequest while in a handshake.
14146 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14148 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14149 should end in 'break', not 'goto end' which circumvents various
14150 cleanups done in state SSL_ST_OK. But session related stuff
14151 must be disabled for SSL_ST_OK in the case that we just sent a
14154 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14155 before just sending a HelloRequest.
14157 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14159 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14160 reveal whether illegal block cipher padding was found or a MAC
14161 verification error occurred. (Neither SSLerr() codes nor alerts
14162 are directly visible to potential attackers, but the information
14163 may leak via logfiles.)
14165 Similar changes are not required for the SSL 2.0 implementation
14166 because the number of padding bytes is sent in clear for SSL 2.0,
14167 and the extra bytes are just ignored. However ssl/s2_pkt.c
14168 failed to verify that the purported number of padding bytes is in
14173 * Add OpenUNIX-8 support including shared libraries
14174 (Boyd Lynn Gerber <gerberb@zenez.com>).
14178 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14179 'wristwatch attack' using huge encoding parameters (cf.
14180 James H. Manger's CRYPTO 2001 paper). Note that the
14181 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14182 encoding parameters and hence was not vulnerable.
14186 * BN_sqr() bug fix.
14188 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14190 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14191 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14192 followed by modular reduction.
14194 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14196 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14197 equivalent based on BN_pseudo_rand() instead of BN_rand().
14201 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14202 This function was broken, as the check for a new client hello message
14203 to handle SGC did not allow these large messages.
14204 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14208 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14212 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14213 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14217 * Rework the configuration and shared library support for Tru64 Unix.
14218 The configuration part makes use of modern compiler features and
14219 still retains old compiler behavior for those that run older versions
14220 of the OS. The shared library support part includes a variant that
14221 uses the RPATH feature, and is available through the special
14222 configuration target "alpha-cc-rpath", which will never be selected
14225 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14227 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14228 with the same message size as in ssl3_get_certificate_request().
14229 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14230 messages might inadvertently be reject as too long.
14232 *Petr Lampa <lampa@fee.vutbr.cz>*
14234 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14238 * Modified SSL library such that the verify_callback that has been set
14239 specifically for an SSL object with SSL_set_verify() is actually being
14240 used. Before the change, a verify_callback set with this function was
14241 ignored and the verify_callback() set in the SSL_CTX at the time of
14242 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14243 to allow the necessary settings.
14247 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14248 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14249 done automatically (in contradiction to the requirements of the C
14250 standard). This made problems when used from OpenSSH.
14254 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14255 dh->length and always used
14257 BN_rand_range(priv_key, dh->p).
14259 BN_rand_range() is not necessary for Diffie-Hellman, and this
14260 specific range makes Diffie-Hellman unnecessarily inefficient if
14261 dh->length (recommended exponent length) is much smaller than the
14262 length of dh->p. We could use BN_rand_range() if the order of
14263 the subgroup was stored in the DH structure, but we only have
14268 BN_rand(priv_key, l, ...)
14270 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14277 RSA_eay_public_encrypt
14278 RSA_eay_private_decrypt
14279 RSA_eay_private_encrypt (signing)
14280 RSA_eay_public_decrypt (signature verification)
14282 (default implementations for RSA_public_encrypt,
14283 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14284 always reject numbers >= n.
14288 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14289 to synchronize access to 'locking_thread'. This is necessary on
14290 systems where access to 'locking_thread' (an 'unsigned long'
14291 variable) is not atomic.
14295 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14296 *before* setting the 'crypto_lock_rand' flag. The previous code had
14297 a race condition if 0 is a valid thread ID.
14299 *Travis Vitek <vitek@roguewave.com>*
14301 * Add support for shared libraries under Irix.
14303 *Albert Chin-A-Young <china@thewrittenword.com>*
14305 * Add configuration option to build on Linux on both big-endian and
14306 little-endian MIPS.
14308 *Ralf Baechle <ralf@uni-koblenz.de>*
14310 * Add the possibility to create shared libraries on HP-UX.
14314 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14316 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14317 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14318 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14319 PRNG state recovery was possible based on the output of
14320 one PRNG request appropriately sized to gain knowledge on
14321 'md' followed by enough consecutive 1-byte PRNG requests
14322 to traverse all of 'state'.
14324 1. When updating 'md_local' (the current thread's copy of 'md')
14325 during PRNG output generation, hash all of the previous
14326 'md_local' value, not just the half used for PRNG output.
14328 2. Make the number of bytes from 'state' included into the hash
14329 independent from the number of PRNG bytes requested.
14331 The first measure alone would be sufficient to avoid
14332 Markku-Juhani's attack. (Actually it had never occurred
14333 to me that the half of 'md_local' used for chaining was the
14334 half from which PRNG output bytes were taken -- I had always
14335 assumed that the secret half would be used.) The second
14336 measure makes sure that additional data from 'state' is never
14337 mixed into 'md_local' in small portions; this heuristically
14338 further strengthens the PRNG.
14342 * Fix crypto/bn/asm/mips3.s.
14346 * When only the key is given to "enc", the IV is undefined. Print out
14347 an error message in this case.
14351 * Handle special case when X509_NAME is empty in X509 printing routines.
14355 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14356 positive and less than q.
14360 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14361 used: it isn't thread safe and the add_lock_callback should handle
14364 *Paul Rose <Paul.Rose@bridge.com>*
14366 * Verify that incoming data obeys the block size in
14367 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14373 *Ulf Möller, Bodo Möller*
14375 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14376 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14377 when fixing the server behaviour for backwards-compatible 'client
14378 hello' messages. (Note that the attack is impractical against
14379 SSL 3.0 and TLS 1.0 anyway because length and version checking
14380 means that the probability of guessing a valid ciphertext is
14381 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14384 Before 0.9.5, the countermeasure (hide the error by generating a
14385 random 'decryption result') did not work properly because
14386 ERR_clear_error() was missing, meaning that SSL_get_error() would
14387 detect the supposedly ignored error.
14389 Both problems are now fixed.
14393 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14394 (previously it was 1024).
14398 * Fix for compatibility mode trust settings: ignore trust settings
14399 unless some valid trust or reject settings are present.
14403 * Fix for blowfish EVP: its a variable length cipher.
14407 * Fix various bugs related to DSA S/MIME verification. Handle missing
14408 parameters in DSA public key structures and return an error in the
14409 DSA routines if parameters are absent.
14413 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14414 in the current directory if neither $RANDFILE nor $HOME was set.
14415 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14416 caused some confusion to Windows users who haven't defined $HOME.
14417 Thus RAND_file_name() is changed again: e_os.h can define a
14418 DEFAULT_HOME, which will be used if $HOME is not set.
14419 For Windows, we use "C:"; on other platforms, we still require
14420 environment variables.
14422 * Move 'if (!initialized) RAND_poll()' into regions protected by
14423 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14424 having multiple threads call RAND_poll() concurrently.
14428 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14429 combination of a flag and a thread ID variable.
14430 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14431 flag), *other* threads can enter ssleay_add_bytes without obeying
14432 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14433 that they do not hold after the first thread unsets add_do_not_lock).
14437 * Change bctest again: '-x' expressions are not available in all
14438 versions of 'test'.
14442 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14444 * Fix a couple of memory leaks in PKCS7_dataDecode()
14446 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14448 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14449 the default extension for executables, if any. Also, make the perl
14450 scripts that use symlink() to test if it really exists and use "cp"
14451 if it doesn't. All this made OpenSSL compilable and installable in
14456 * Fix for asn1_GetSequence() for indefinite length constructed data.
14457 If SEQUENCE is length is indefinite just set c->slen to the total
14458 amount of data available.
14460 *Steve Henson, reported by shige@FreeBSD.org*
14462 *This change does not apply to 0.9.7.*
14464 * Change bctest to avoid here-documents inside command substitution
14465 (workaround for FreeBSD /bin/sh bug).
14466 For compatibility with Ultrix, avoid shell functions (introduced
14467 in the bctest version that searches along $PATH).
14471 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14472 with des_encrypt() defined on some operating systems, like Solaris
14477 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14478 On the Importance of Eliminating Errors in Cryptographic
14479 Computations, J. Cryptology 14 (2001) 2, 101-119,
14480 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14484 * MIPS assembler BIGNUM division bug fix.
14488 * Disabled incorrect Alpha assembler code.
14492 * Fix PKCS#7 decode routines so they correctly update the length
14493 after reading an EOC for the EXPLICIT tag.
14497 *This change does not apply to 0.9.7.*
14499 * Fix bug in PKCS#12 key generation routines. This was triggered
14500 if a 3DES key was generated with a 0 initial byte. Include
14501 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14502 (but broken) behaviour.
14506 * Enhance bctest to search for a working bc along $PATH and print
14509 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14511 * Fix memory leaks in err.c: free err_data string if necessary;
14512 don't write to the wrong index in ERR_set_error_data.
14516 * Implement ssl23_peek (analogous to ssl23_read), which previously
14521 * Replace rdtsc with `_emit` statements for VC++ version 5.
14523 *Jeremy Cooper <jeremy@baymoo.org>*
14525 * Make it possible to reuse SSLv2 sessions.
14529 * In copy_email() check for >= 0 as a return value for
14530 X509_NAME_get_index_by_NID() since 0 is a valid index.
14532 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14534 * Avoid coredump with unsupported or invalid public keys by checking if
14535 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14536 PKCS7_verify() fails with non detached data.
14540 * Don't use getenv in library functions when run as setuid/setgid.
14541 New function OPENSSL_issetugid().
14545 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14546 due to incorrect handling of multi-threading:
14548 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14550 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14552 3. Count how many times MemCheck_off() has been called so that
14553 nested use can be treated correctly. This also avoids
14554 inband-signalling in the previous code (which relied on the
14555 assumption that thread ID 0 is impossible).
14559 * Add "-rand" option also to s_client and s_server.
14563 * Fix CPU detection on Irix 6.x.
14564 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14565 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14567 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14572 *This change does not apply to 0.9.7.*
14574 * Use the cached encoding of an X509_NAME structure rather than
14575 copying it. This is apparently the reason for the libsafe "errors"
14576 but the code is actually correct.
14580 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14581 Bleichenbacher's DSA attack.
14582 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14583 to be set and top=0 forces the highest bit to be set; top=-1 is new
14584 and leaves the highest bit random.
14586 *Ulf Moeller, Bodo Moeller*
14588 * In the `NCONF_...`-based implementations for `CONF_...` queries
14589 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14590 a temporary CONF structure with the data component set to NULL
14591 (which gives segmentation faults in lh_retrieve).
14592 Instead, use NULL for the CONF pointer in CONF_get_string and
14593 CONF_get_number (which may use environment variables) and directly
14594 return NULL from CONF_get_section.
14598 * Fix potential buffer overrun for EBCDIC.
14602 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14603 keyUsage if basicConstraints absent for a CA.
14607 * Make SMIME_write_PKCS7() write mail header values with a format that
14608 is more generally accepted (no spaces before the semicolon), since
14609 some programs can't parse those values properly otherwise. Also make
14610 sure BIO's that break lines after each write do not create invalid
14615 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14616 macros previously used would not encode an empty SEQUENCE OF
14617 and break the signature.
14621 *This change does not apply to 0.9.7.*
14623 * Zero the premaster secret after deriving the master secret in
14628 * Add some EVP_add_digest_alias registrations (as found in
14629 OpenSSL_add_all_digests()) to SSL_library_init()
14630 aka OpenSSL_add_ssl_algorithms(). This provides improved
14631 compatibility with peers using X.509 certificates
14632 with unconventional AlgorithmIdentifier OIDs.
14636 * Fix for Irix with NO_ASM.
14638 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14640 * ./config script fixes.
14642 *Ulf Moeller, Richard Levitte*
14644 * Fix 'openssl passwd -1'.
14648 * Change PKCS12_key_gen_asc() so it can cope with non null
14649 terminated strings whose length is passed in the passlen
14650 parameter, for example from PEM callbacks. This was done
14651 by adding an extra length parameter to asc2uni().
14653 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14655 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14656 call failed, free the DSA structure.
14660 * Fix to uni2asc() to cope with zero length Unicode strings.
14661 These are present in some PKCS#12 files.
14665 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14666 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14667 when writing a 32767 byte record.
14669 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14671 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14672 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14674 (RSA objects have a reference count access to which is protected
14675 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14676 so they are meant to be shared between threads.)
14677 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14678 "Reddie, Steven" <Steven.Reddie@ca.com>*
14680 * Fix a deadlock in CRYPTO_mem_leaks().
14684 * Use better test patterns in bntest.
14688 * rand_win.c fix for Borland C.
14692 * BN_rshift bugfix for n == 0.
14696 * Add a 'bctest' script that checks for some known 'bc' bugs
14697 so that 'make test' does not abort just because 'bc' is broken.
14701 * Store verify_result within SSL_SESSION also for client side to
14702 avoid potential security hole. (Re-used sessions on the client side
14703 always resulted in verify_result==X509_V_OK, not using the original
14704 result of the server certificate verification.)
14708 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14709 SSL3_RT_APPLICATION_DATA, return 0.
14710 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14715 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14716 releases, have been re-implemented by renaming the previous
14717 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14718 and ssl3_read_internal, respectively, and adding 'peek' parameters
14719 to them. The new ssl[23]_{read,peek} functions are calls to
14720 ssl[23]_read_internal with the 'peek' flag set appropriately.
14721 A 'peek' parameter has also been added to ssl3_read_bytes, which
14722 does the actual work for ssl3_read_internal.
14726 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14727 the method-specific "init()" handler. Also clean up ex_data after
14728 calling the method-specific "finish()" handler. Previously, this was
14729 happening the other way round.
14733 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14734 The previous value, 12, was not always sufficient for BN_mod_exp().
14738 * Make sure that shared libraries get the internal name engine with
14739 the full version number and not just 0. This should mark the
14740 shared libraries as not backward compatible. Of course, this should
14741 be changed again when we can guarantee backward binary compatibility.
14745 * Fix typo in get_cert_by_subject() in by_dir.c
14747 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14749 * Rework the system to generate shared libraries:
14751 - Make note of the expected extension for the shared libraries and
14752 if there is a need for symbolic links from for example libcrypto.so.0
14753 to libcrypto.so.0.9.7. There is extended info in Configure for
14756 - Make as few rebuilds of the shared libraries as possible.
14758 - Still avoid linking the OpenSSL programs with the shared libraries.
14760 - When installing, install the shared libraries separately from the
14765 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14767 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14768 and not in SSL_clear because the latter is also used by the
14769 accept/connect functions; previously, the settings made by
14770 SSL_set_read_ahead would be lost during the handshake.
14772 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14774 * Correct util/mkdef.pl to be selective about disabled algorithms.
14775 Previously, it would create entries for disabled algorithms no
14780 * Added several new manual pages for SSL_* function.
14784 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14786 * In ssl23_get_client_hello, generate an error message when faced
14787 with an initial SSL 3.0/TLS record that is too small to contain the
14788 first two bytes of the ClientHello message, i.e. client_version.
14789 (Note that this is a pathologic case that probably has never happened
14790 in real life.) The previous approach was to use the version number
14791 from the record header as a substitute; but our protocol choice
14792 should not depend on that one because it is not authenticated
14793 by the Finished messages.
14797 * More robust randomness gathering functions for Windows.
14799 *Jeffrey Altman <jaltman@columbia.edu>*
14801 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14802 not set then we don't setup the error code for issuer check errors
14803 to avoid possibly overwriting other errors which the callback does
14804 handle. If an application does set the flag then we assume it knows
14805 what it is doing and can handle the new informational codes
14810 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14811 a general "ANY" type, as such it should be able to decode anything
14812 including tagged types. However it didn't check the class so it would
14813 wrongly interpret tagged types in the same way as their universal
14814 counterpart and unknown types were just rejected. Changed so that the
14815 tagged and unknown types are handled in the same way as a SEQUENCE:
14816 that is the encoding is stored intact. There is also a new type
14817 "V_ASN1_OTHER" which is used when the class is not universal, in this
14818 case we have no idea what the actual type is so we just lump them all
14823 * On VMS, stdout may very well lead to a file that is written to
14824 in a record-oriented fashion. That means that every write() will
14825 write a separate record, which will be read separately by the
14826 programs trying to read from it. This can be very confusing.
14828 The solution is to put a BIO filter in the way that will buffer
14829 text until a linefeed is reached, and then write everything a
14830 line at a time, so every record written will be an actual line,
14831 not chunks of lines and not (usually doesn't happen, but I've
14832 seen it once) several lines in one record. BIO_f_linebuffer() is
14835 Currently, it's a VMS-only method, because that's where it has
14836 been tested well enough.
14840 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14841 it can return incorrect results.
14842 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14843 but it was in 0.9.6-beta[12].)
14847 * Disable the check for content being present when verifying detached
14848 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14849 include zero length content when signing messages.
14853 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14854 BIO_ctrl (for BIO pairs).
14858 * Add DSO method for VMS.
14862 * Bug fix: Montgomery multiplication could produce results with the
14867 * Add RPM specification openssl.spec and modify it to build three
14868 packages. The default package contains applications, application
14869 documentation and run-time libraries. The devel package contains
14870 include files, static libraries and function documentation. The
14871 doc package contains the contents of the doc directory. The original
14872 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14876 * Add a large number of documentation files for many SSL routines.
14878 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14880 * Add a configuration entry for Sony News 4.
14882 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14884 * Don't set the two most significant bits to one when generating a
14885 random number < q in the DSA library.
14889 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14890 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14891 the underlying transport is blocking) if a handshake took place.
14892 (The default behaviour is needed by applications such as s_client
14893 and s_server that use select() to determine when to use SSL_read;
14894 but for applications that know in advance when to expect data, it
14895 just makes things more complicated.)
14899 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14904 * Add a few more EBCDIC conditionals that make `req` and `x509`
14905 work better on such systems.
14907 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14909 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14910 Update PKCS12_parse() so it copies the friendlyName and the
14911 keyid to the certificates aux info.
14915 * Fix bug in PKCS7_verify() which caused an infinite loop
14916 if there was more than one signature.
14918 *Sven Uszpelkat <su@celocom.de>*
14920 * Major change in util/mkdef.pl to include extra information
14921 about each symbol, as well as presenting variables as well
14922 as functions. This change means that there's n more need
14923 to rebuild the .num files when some algorithms are excluded.
14927 * Allow the verify time to be set by an application,
14928 rather than always using the current time.
14932 * Phase 2 verify code reorganisation. The certificate
14933 verify code now looks up an issuer certificate by a
14934 number of criteria: subject name, authority key id
14935 and key usage. It also verifies self signed certificates
14936 by the same criteria. The main comparison function is
14937 X509_check_issued() which performs these checks.
14939 Lot of changes were necessary in order to support this
14940 without completely rewriting the lookup code.
14942 Authority and subject key identifier are now cached.
14944 The LHASH 'certs' is X509_STORE has now been replaced
14945 by a STACK_OF(X509_OBJECT). This is mainly because an
14946 LHASH can't store or retrieve multiple objects with
14947 the same hash value.
14949 As a result various functions (which were all internal
14950 use only) have changed to handle the new X509_STORE
14951 structure. This will break anything that messed round
14952 with X509_STORE internally.
14954 The functions X509_STORE_add_cert() now checks for an
14955 exact match, rather than just subject name.
14957 The X509_STORE API doesn't directly support the retrieval
14958 of multiple certificates matching a given criteria, however
14959 this can be worked round by performing a lookup first
14960 (which will fill the cache with candidate certificates)
14961 and then examining the cache for matches. This is probably
14962 the best we can do without throwing out X509_LOOKUP
14963 entirely (maybe later...).
14965 The X509_VERIFY_CTX structure has been enhanced considerably.
14967 All certificate lookup operations now go via a get_issuer()
14968 callback. Although this currently uses an X509_STORE it
14969 can be replaced by custom lookups. This is a simple way
14970 to bypass the X509_STORE hackery necessary to make this
14971 work and makes it possible to use more efficient techniques
14972 in future. A very simple version which uses a simple
14973 STACK for its trusted certificate store is also provided
14974 using X509_STORE_CTX_trusted_stack().
14976 The verify_cb() and verify() callbacks now have equivalents
14977 in the X509_STORE_CTX structure.
14979 X509_STORE_CTX also has a 'flags' field which can be used
14980 to customise the verify behaviour.
14984 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14985 excludes S/MIME capabilities.
14989 * When a certificate request is read in keep a copy of the
14990 original encoding of the signed data and use it when outputting
14991 again. Signatures then use the original encoding rather than
14992 a decoded, encoded version which may cause problems if the
14993 request is improperly encoded.
14997 * For consistency with other BIO_puts implementations, call
14998 buffer_write(b, ...) directly in buffer_puts instead of calling
15001 In BIO_puts, increment b->num_write as in BIO_write.
15003 *Peter.Sylvester@EdelWeb.fr*
15005 * Fix BN_mul_word for the case where the word is 0. (We have to use
15006 BN_zero, we may not return a BIGNUM with an array consisting of
15007 words set to zero.)
15011 * Avoid calling abort() from within the library when problems are
15012 detected, except if preprocessor symbols have been defined
15013 (such as REF_CHECK, BN_DEBUG etc.).
15017 * New openssl application 'rsautl'. This utility can be
15018 used for low level RSA operations. DER public key
15019 BIO/fp routines also added.
15023 * New Configure entry and patches for compiling on QNX 4.
15025 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15027 * A demo state-machine implementation was sponsored by
15028 Nuron (<http://www.nuron.com/>) and is now available in
15029 demos/state_machine.
15033 * New options added to the 'dgst' utility for signature
15034 generation and verification.
15038 * Unrecognized PKCS#7 content types are now handled via a
15039 catch all ASN1_TYPE structure. This allows unsupported
15040 types to be stored as a "blob" and an application can
15041 encode and decode it manually.
15045 * Fix various signed/unsigned issues to make a_strex.c
15046 compile under VC++.
15048 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15050 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15051 length if passed a buffer. ASN1_INTEGER_to_BN failed
15052 if passed a NULL BN and its argument was negative.
15054 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15056 * Modification to PKCS#7 encoding routines to output definite
15057 length encoding. Since currently the whole structures are in
15058 memory there's not real point in using indefinite length
15059 constructed encoding. However if OpenSSL is compiled with
15060 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15064 * Added BIO_vprintf() and BIO_vsnprintf().
15068 * Added more prefixes to parse for in the strings written
15069 through a logging bio, to cover all the levels that are available
15070 through syslog. The prefixes are now:
15072 PANIC, EMERG, EMR => LOG_EMERG
15073 ALERT, ALR => LOG_ALERT
15074 CRIT, CRI => LOG_CRIT
15075 ERROR, ERR => LOG_ERR
15076 WARNING, WARN, WAR => LOG_WARNING
15077 NOTICE, NOTE, NOT => LOG_NOTICE
15078 INFO, INF => LOG_INFO
15079 DEBUG, DBG => LOG_DEBUG
15081 and as before, if none of those prefixes are present at the
15082 beginning of the string, LOG_ERR is chosen.
15084 On Win32, the `LOG_*` levels are mapped according to this:
15086 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15087 LOG_WARNING => EVENTLOG_WARNING_TYPE
15088 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15092 * Made it possible to reconfigure with just the configuration
15093 argument "reconf" or "reconfigure". The command line arguments
15094 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15095 and are retrieved from there when reconfiguring.
15101 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15103 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15107 * The obj_dat.pl script was messing up the sorting of object
15108 names. The reason was that it compared the quoted version
15109 of strings as a result "OCSP" > "OCSP Signing" because
15110 " > SPACE. Changed script to store unquoted versions of
15111 names and add quotes on output. It was also omitting some
15112 names from the lookup table if they were given a default
15113 value (that is if SN is missing it is given the same
15114 value as LN and vice versa), these are now added on the
15115 grounds that if an object has a name we should be able to
15116 look it up. Finally added warning output when duplicate
15117 short or long names are found.
15121 * Changes needed for Tandem NSK.
15123 *Scott Uroff <scott@xypro.com>*
15125 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15126 RSA_padding_check_SSLv23(), special padding was never detected
15127 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15128 version rollback attacks was not effective.
15130 In s23_clnt.c, don't use special rollback-attack detection padding
15131 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15132 client; similarly, in s23_srvr.c, don't do the rollback check if
15133 SSL 2.0 is the only protocol enabled in the server.
15137 * Make it possible to get hexdumps of unprintable data with 'openssl
15138 asn1parse'. By implication, the functions ASN1_parse_dump() and
15139 BIO_dump_indent() are added.
15143 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15144 these print out strings and name structures based on various
15145 flags including RFC2253 support and proper handling of
15146 multibyte characters. Added options to the 'x509' utility
15147 to allow the various flags to be set.
15151 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15152 Also change the functions X509_cmp_current_time() and
15153 X509_gmtime_adj() work with an ASN1_TIME structure,
15154 this will enable certificates using GeneralizedTime in validity
15155 dates to be checked.
15159 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15160 negative public key encodings) on by default,
15161 NO_NEG_PUBKEY_BUG can be set to disable it.
15165 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15166 content octets. An i2c_ASN1_OBJECT is unnecessary because
15167 the encoding can be trivially obtained from the structure.
15171 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15172 not read locks (`CRYPTO_r_[un]lock`).
15176 * A first attempt at creating official support for shared
15177 libraries through configuration. I've kept it so the
15178 default is static libraries only, and the OpenSSL programs
15179 are always statically linked for now, but there are
15180 preparations for dynamic linking in place.
15181 This has been tested on Linux and Tru64.
15185 * Randomness polling function for Win9x, as described in:
15186 Peter Gutmann, Software Generation of Practically Strong
15191 * Fix so PRNG is seeded in req if using an already existing
15196 * New options to smime application. -inform and -outform
15197 allow alternative formats for the S/MIME message including
15198 PEM and DER. The -content option allows the content to be
15199 specified separately. This should allow things like Netscape
15200 form signing output easier to verify.
15204 * Fix the ASN1 encoding of tags using the 'long form'.
15208 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15209 STRING types. These convert content octets to and from the
15210 underlying type. The actual tag and length octets are
15211 already assumed to have been read in and checked. These
15212 are needed because all other string types have virtually
15213 identical handling apart from the tag. By having versions
15214 of the ASN1 functions that just operate on content octets
15215 IMPLICIT tagging can be handled properly. It also allows
15216 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15217 and ASN1_INTEGER are identical apart from the tag.
15221 * Change the handling of OID objects as follows:
15223 - New object identifiers are inserted in objects.txt, following
15224 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15225 - objects.pl is used to process obj_mac.num and create a new
15227 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15230 This is currently kind of a hack, and the perl code in objects.pl
15231 isn't very elegant, but it works as I intended. The simplest way
15232 to check that it worked correctly is to look in obj_dat.h and
15233 check the array nid_objs and make sure the objects haven't moved
15234 around (this is important!). Additions are OK, as well as
15235 consistent name changes.
15239 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15243 * Addition of the command line parameter '-rand file' to 'openssl req'.
15244 The given file adds to whatever has already been seeded into the
15245 random pool through the RANDFILE configuration file option or
15246 environment variable, or the default random state file.
15250 * mkstack.pl now sorts each macro group into lexical order.
15251 Previously the output order depended on the order the files
15252 appeared in the directory, resulting in needless rewriting
15257 * Patches to make OpenSSL compile under Win32 again. Mostly
15258 work arounds for the VC++ problem that it treats func() as
15259 func(void). Also stripped out the parts of mkdef.pl that
15260 added extra typesafe functions: these no longer exist.
15264 * Reorganisation of the stack code. The macros are now all
15265 collected in safestack.h . Each macro is defined in terms of
15266 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15267 DEBUG_SAFESTACK is now handled in terms of function casts,
15268 this has the advantage of retaining type safety without the
15269 use of additional functions. If DEBUG_SAFESTACK is not defined
15270 then the non typesafe macros are used instead. Also modified the
15271 mkstack.pl script to handle the new form. Needs testing to see
15272 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15273 the default if no major problems. Similar behaviour for ASN1_SET_OF
15274 and PKCS12_STACK_OF.
15278 * When some versions of IIS use the 'NET' form of private key the
15279 key derivation algorithm is different. Normally MD5(password) is
15280 used as a 128 bit RC4 key. In the modified case
15281 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15282 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15283 as the old Netscape_RSA functions except they have an additional
15284 'sgckey' parameter which uses the modified algorithm. Also added
15285 an -sgckey command line option to the rsa utility. Thanks to
15286 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15287 algorithm to openssl-dev.
15291 * The evp_local.h macros were using 'c.##kname' which resulted in
15292 invalid expansion on some systems (SCO 5.0.5 for example).
15293 Corrected to 'c.kname'.
15295 *Phillip Porch <root@theporch.com>*
15297 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15298 a STACK of email addresses from a certificate or request, these look
15299 in the subject name and the subject alternative name extensions and
15300 omit any duplicate addresses.
15304 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15305 This makes DSA verification about 2 % faster.
15309 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15310 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15311 plus overhead for 1024 bit moduli).
15312 This makes exponentiations about 0.5 % faster for 1024 bit
15313 exponents (as measured by "openssl speed rsa2048").
15317 * Rename memory handling macros to avoid conflicts with other
15319 Malloc => OPENSSL_malloc
15320 Malloc_locked => OPENSSL_malloc_locked
15321 Realloc => OPENSSL_realloc
15322 Free => OPENSSL_free
15326 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15327 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15331 * CygWin32 support.
15333 *John Jarvie <jjarvie@newsguy.com>*
15335 * The type-safe stack code has been rejigged. It is now only compiled
15336 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15337 by default all type-specific stack functions are "#define"d back to
15338 standard stack functions. This results in more streamlined output
15339 but retains the type-safety checking possibilities of the original
15344 * The STACK code has been cleaned up, and certain type declarations
15345 that didn't make a lot of sense have been brought in line. This has
15346 also involved a cleanup of sorts in safestack.h to more correctly
15347 map type-safe stack functions onto their plain stack counterparts.
15348 This work has also resulted in a variety of "const"ifications of
15349 lots of the code, especially `_cmp` operations which should normally
15350 be prototyped with "const" parameters anyway.
15354 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15355 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15356 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15357 where all of 'md' is used each time the PRNG is used, but 'state'
15358 is used only indexed by a cyclic counter. As entropy may not be
15359 well distributed from the beginning, 'md' is important as a
15360 chaining variable. However, the output function chains only half
15361 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15362 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15363 in all of 'state' being rewritten, with the new values depending
15364 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15368 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15369 the handshake is continued after ssl_verify_cert_chain();
15370 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15371 can lead to 'unexplainable' connection aborts later.
15373 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15375 * Major EVP API cipher revision.
15376 Add hooks for extra EVP features. This allows various cipher
15377 parameters to be set in the EVP interface. Support added for variable
15378 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15379 setting of RC2 and RC5 parameters.
15381 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15384 Remove lots of duplicated code from the EVP library. For example *every*
15385 cipher init() function handles the 'iv' in the same way according to the
15386 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15387 for CFB and OFB modes they zero ctx->num.
15389 New functionality allows removal of S/MIME code RC2 hack.
15391 Most of the routines have the same form and so can be declared in terms
15394 By shifting this to the top level EVP_CipherInit() it can be removed from
15395 all individual ciphers. If the cipher wants to handle IVs or keys
15396 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15399 Change lots of functions like EVP_EncryptUpdate() to now return a
15400 value: although software versions of the algorithms cannot fail
15401 any installed hardware versions can.
15405 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15406 this option is set, tolerate broken clients that send the negotiated
15407 protocol version number instead of the requested protocol version
15412 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15413 i.e. non-zero for export ciphersuites, zero otherwise.
15414 Previous versions had this flag inverted, inconsistent with
15415 rsa_tmp_cb (..._TMP_RSA_CB).
15417 *Bodo Moeller; problem reported by Amit Chopra*
15419 * Add missing DSA library text string. Work around for some IIS
15420 key files with invalid SEQUENCE encoding.
15424 * Add a document (doc/standards.txt) that list all kinds of standards
15425 and so on that are implemented in OpenSSL.
15429 * Enhance c_rehash script. Old version would mishandle certificates
15430 with the same subject name hash and wouldn't handle CRLs at all.
15431 Added -fingerprint option to crl utility, to support new c_rehash
15436 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15440 * Fix for SSL server purpose checking. Server checking was
15441 rejecting certificates which had extended key usage present
15442 but no ssl client purpose.
15444 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15446 * Make PKCS#12 code work with no password. The PKCS#12 spec
15447 is a little unclear about how a blank password is handled.
15448 Since the password in encoded as a BMPString with terminating
15449 double NULL a zero length password would end up as just the
15450 double NULL. However no password at all is different and is
15451 handled differently in the PKCS#12 key generation code. NS
15452 treats a blank password as zero length. MSIE treats it as no
15453 password on export: but it will try both on import. We now do
15454 the same: PKCS12_parse() tries zero length and no password if
15455 the password is set to "" or NULL (NULL is now a valid password:
15456 it wasn't before) as does the pkcs12 application.
15460 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15461 perror when PEM_read_bio_X509_REQ fails, the error message must
15462 be obtained from the error queue.
15466 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15467 it in ERR_remove_state if appropriate, and change ERR_get_state
15468 accordingly to avoid race conditions (this is necessary because
15469 thread_hash is no longer constant once set).
15473 * Bugfix for linux-elf makefile.one.
15477 * RSA_get_default_method() will now cause a default
15478 RSA_METHOD to be chosen if one doesn't exist already.
15479 Previously this was only set during a call to RSA_new()
15480 or RSA_new_method(NULL) meaning it was possible for
15481 RSA_get_default_method() to return NULL.
15485 * Added native name translation to the existing DSO code
15486 that will convert (if the flag to do so is set) filenames
15487 that are sufficiently small and have no path information
15488 into a canonical native form. Eg. "blah" converted to
15489 "libblah.so" or "blah.dll" etc.
15493 * New function ERR_error_string_n(e, buf, len) which is like
15494 ERR_error_string(e, buf), but writes at most 'len' bytes
15495 including the 0 terminator. For ERR_error_string_n, 'buf'
15498 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15500 * CONF library reworked to become more general. A new CONF
15501 configuration file reader "class" is implemented as well as a
15502 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15503 old `CONF_*` functions are still there, but are reimplemented to
15504 work in terms of the new functions. Also, a set of functions
15505 to handle the internal storage of the configuration data is
15506 provided to make it easier to write new configuration file
15507 reader "classes" (I can definitely see something reading a
15508 configuration file in XML format, for example), called `_CONF_*`,
15509 or "the configuration storage API"...
15511 The new configuration file reading functions are:
15513 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15514 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15516 NCONF_default, NCONF_WIN32
15518 NCONF_dump_fp, NCONF_dump_bio
15520 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15521 NCONF_new creates a new CONF object. This works in the same way
15522 as other interfaces in OpenSSL, like the BIO interface.
15523 `NCONF_dump_*` dump the internal storage of the configuration file,
15524 which is useful for debugging. All other functions take the same
15525 arguments as the old `CONF_*` functions with the exception of the
15526 first that must be a `CONF *` instead of a `LHASH *`.
15528 To make it easier to use the new classes with the old `CONF_*` functions,
15529 the function CONF_set_default_method is provided.
15533 * Add '-tls1' option to 'openssl ciphers', which was already
15534 mentioned in the documentation but had not been implemented.
15535 (This option is not yet really useful because even the additional
15536 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15540 * Initial DSO code added into libcrypto for letting OpenSSL (and
15541 OpenSSL-based applications) load shared libraries and bind to
15542 them in a portable way.
15544 *Geoff Thorpe, with contributions from Richard Levitte*
15546 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15548 * Make sure _lrotl and _lrotr are only used with MSVC.
15550 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15551 (the default implementation of RAND_status).
15553 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15554 to '-clrext' (= clear extensions), as intended and documented.
15555 *Bodo Moeller; inconsistency pointed out by Michael Attili
15556 <attili@amaxo.com>*
15558 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15559 was larger than the MD block size.
15561 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15563 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15564 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15565 using the passed key: if the passed key was a private key the result
15566 of X509_print(), for example, would be to print out all the private key
15571 * des_quad_cksum() byte order bug fix.
15572 *Ulf Möller, using the problem description in krb4-0.9.7, where
15573 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15575 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15578 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15580 * For easily testing in shell scripts whether some command
15581 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15582 returns with exit code 0 iff no command of the given name is available.
15583 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15584 the output goes to stdout and nothing is printed to stderr.
15585 Additional arguments are always ignored.
15587 Since for each cipher there is a command of the same name,
15588 the 'no-cipher' compilation switches can be tested this way.
15590 ('openssl no-XXX' is not able to detect pseudo-commands such
15591 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15595 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15599 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15600 is set; it will be thrown away anyway because each handshake creates
15602 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15603 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15604 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15605 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15609 * New s_client option -ign_eof: EOF at stdin is ignored, and
15610 'Q' and 'R' lose their special meanings (quit/renegotiate).
15611 This is part of what -quiet does; unlike -quiet, -ign_eof
15612 does not suppress any output.
15616 * Add compatibility options to the purpose and trust code. The
15617 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15618 accepts a certificate or CA, this was the previous behaviour,
15619 with all the associated security issues.
15621 X509_TRUST_COMPAT is the old trust behaviour: only and
15622 automatically trust self signed roots in certificate store. A
15623 new trust setting X509_TRUST_DEFAULT is used to specify that
15624 a purpose has no associated trust setting and it should instead
15625 use the value in the default purpose.
15629 * Fix the PKCS#8 DSA private key code so it decodes keys again
15630 and fix a memory leak.
15634 * In util/mkerr.pl (which implements 'make errors'), preserve
15635 reason strings from the previous version of the .c file, as
15636 the default to have only downcase letters (and digits) in
15637 automatically generated reasons codes is not always appropriate.
15641 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15642 using strerror. Previously, ERR_reason_error_string() returned
15643 library names as reason strings for SYSerr; but SYSerr is a special
15644 case where small numbers are errno values, not library numbers.
15648 * Add '-dsaparam' option to 'openssl dhparam' application. This
15649 converts DSA parameters into DH parameters. (When creating parameters,
15650 DSA_generate_parameters is used.)
15654 * Include 'length' (recommended exponent length) in C code generated
15655 by 'openssl dhparam -C'.
15659 * The second argument to set_label in perlasm was already being used
15660 so couldn't be used as a "file scope" flag. Moved to third argument
15665 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15666 instead of RAND_bytes for encryption IVs and salts.
15670 * Include RAND_status() into RAND_METHOD instead of implementing
15671 it only for md_rand.c Otherwise replacing the PRNG by calling
15672 RAND_set_rand_method would be impossible.
15676 * Don't let DSA_generate_key() enter an infinite loop if the random
15677 number generation fails.
15681 * New 'rand' application for creating pseudo-random output.
15685 * Added configuration support for Linux/IA64
15687 *Rolf Haberrecker <rolf@suse.de>*
15689 * Assembler module support for Mingw32.
15693 * Shared library support for HPUX (in shlib/).
15695 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15697 * Shared library support for Solaris gcc.
15699 *Lutz Behnke <behnke@trustcenter.de>*
15701 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15703 * PKCS7_encrypt() was adding text MIME headers twice because they
15704 were added manually and by SMIME_crlf_copy().
15708 * In bntest.c don't call BN_rand with zero bits argument.
15710 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15712 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15713 case was implemented. This caused BN_div_recp() to fail occasionally.
15717 * Add an optional second argument to the set_label() in the perl
15718 assembly language builder. If this argument exists and is set
15719 to 1 it signals that the assembler should use a symbol whose
15720 scope is the entire file, not just the current function. This
15721 is needed with MASM which uses the format label:: for this scope.
15723 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15725 * Change the ASN1 types so they are typedefs by default. Before
15726 almost all types were #define'd to ASN1_STRING which was causing
15727 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15732 * Change names of new functions to the new get1/get0 naming
15733 convention: After 'get1', the caller owns a reference count
15734 and has to call `..._free`; 'get0' returns a pointer to some
15735 data structure without incrementing reference counters.
15736 (Some of the existing 'get' functions increment a reference
15737 counter, some don't.)
15738 Similarly, 'set1' and 'add1' functions increase reference
15739 counters or duplicate objects.
15743 * Allow for the possibility of temp RSA key generation failure:
15744 the code used to assume it always worked and crashed on failure.
15748 * Fix potential buffer overrun problem in BIO_printf().
15749 *Ulf Möller, using public domain code by Patrick Powell; problem
15750 pointed out by David Sacerdote <das33@cornell.edu>*
15752 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15753 RAND_egd() and RAND_status(). In the command line application,
15754 the EGD socket can be specified like a seed file using RANDFILE
15759 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15760 Some CAs (e.g. Verisign) distribute certificates in this form.
15764 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15765 list to exclude them. This means that no special compilation option
15766 is needed to use anonymous DH: it just needs to be included in the
15771 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15772 EVP_MD_type. The old functionality is available in a new macro called
15773 EVP_MD_md(). Change code that uses it and update docs.
15777 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15778 where the `void *` argument is replaced by a function pointer argument.
15779 Previously `void *` was abused to point to functions, which works on
15780 many platforms, but is not correct. As these functions are usually
15781 called by macros defined in OpenSSL header files, most source code
15782 should work without changes.
15786 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15787 sections with information on -D... compiler switches used for
15788 compiling the library so that applications can see them. To enable
15789 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15790 must be defined. E.g.,
15791 #define OPENSSL_ALGORITHM_DEFINES
15792 #include <openssl/opensslconf.h>
15793 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15795 *Richard Levitte, Ulf and Bodo Möller*
15797 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15802 * Change the 'other' type in certificate aux info to a STACK_OF
15803 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15804 the required ASN1 format: arbitrary types determined by an OID.
15808 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15809 argument to 'req'. This is not because the function is newer or
15810 better than others it just uses the work 'NEW' in the certificate
15811 request header lines. Some software needs this.
15815 * Reorganise password command line arguments: now passwords can be
15816 obtained from various sources. Delete the PEM_cb function and make
15817 it the default behaviour: i.e. if the callback is NULL and the
15818 usrdata argument is not NULL interpret it as a null terminated pass
15819 phrase. If usrdata and the callback are NULL then the pass phrase
15820 is prompted for as usual.
15824 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15825 the support is automatically enabled. The resulting binaries will
15826 autodetect the card and use it if present.
15828 *Ben Laurie and Compaq Inc.*
15830 * Work around for Netscape hang bug. This sends certificate request
15831 and server done in one record. Since this is perfectly legal in the
15832 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15833 the bugs/SSLv3 entry for more info.
15837 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15841 * Add -rand argument to smime and pkcs12 applications and read/write
15846 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15850 * Add command line password options to the remaining applications.
15854 * Bug fix for BN_div_recp() for numerators with an even number of
15859 * More tests in bntest.c, and changed test_bn output.
15863 * ./config recognizes MacOS X now.
15867 * Bug fix for BN_div() when the first words of num and divisor are
15868 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15872 * Add support for various broken PKCS#8 formats, and command line
15873 options to produce them.
15877 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15878 get temporary BIGNUMs from a BN_CTX.
15882 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15887 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15888 include a #define from the old name to the new. The original intent
15889 was that statically linked binaries could for example just call
15890 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15891 link with digests. This never worked because SSLeay_add_all_digests()
15892 and SSLeay_add_all_ciphers() were in the same source file so calling
15893 one would link with the other. They are now in separate source files.
15897 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15901 * Use a less unusual form of the Miller-Rabin primality test (it used
15902 a binary algorithm for exponentiation integrated into the Miller-Rabin
15903 loop, our standard modexp algorithms are faster).
15907 * Support for the EBCDIC character set completed.
15909 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15911 * Source code cleanups: use const where appropriate, eliminate casts,
15912 use `void *` instead of `char *` in lhash.
15916 * Bugfix: ssl3_send_server_key_exchange was not restartable
15917 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15918 this the server could overwrite ephemeral keys that the client
15923 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15924 using 50 iterations of the Rabin-Miller test.
15926 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15927 iterations of the Rabin-Miller test as required by the appendix
15928 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15929 As BN_is_prime_fasttest includes trial division, DSA parameter
15930 generation becomes much faster.
15932 This implies a change for the callback functions in DSA_is_prime
15933 and DSA_generate_parameters: The callback function is called once
15934 for each positive witness in the Rabin-Miller test, not just
15935 occasionally in the inner loop; and the parameters to the
15936 callback function now provide an iteration count for the outer
15937 loop rather than for the current invocation of the inner loop.
15938 DSA_generate_parameters additionally can call the callback
15939 function with an 'iteration count' of -1, meaning that a
15940 candidate has passed the trial division test (when q is generated
15941 from an application-provided seed, trial division is skipped).
15945 * New function BN_is_prime_fasttest that optionally does trial
15946 division before starting the Rabin-Miller test and has
15947 an additional BN_CTX * argument (whereas BN_is_prime always
15948 has to allocate at least one BN_CTX).
15949 'callback(1, -1, cb_arg)' is called when a number has passed the
15950 trial division stage.
15954 * Fix for bug in CRL encoding. The validity dates weren't being handled
15959 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15963 * New function BN_pseudo_rand().
15967 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15968 bignum version of BN_from_montgomery() with the working code from
15969 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15974 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15975 made it impossible to use the same SSL_SESSION data structure in
15976 SSL2 clients in multiple threads.
15980 * The return value of RAND_load_file() no longer counts bytes obtained
15981 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15982 to seed the PRNG (previously an explicit byte count was required).
15984 *Ulf Möller, Bodo Möller*
15986 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15987 used `char *` instead of `void *` and had casts all over the place.
15991 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15995 * Retain source code compatibility for BN_prime_checks macro:
15996 BN_is_prime(..., BN_prime_checks, ...) now uses
15997 BN_prime_checks_for_size to determine the appropriate number of
15998 Rabin-Miller iterations.
16002 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16003 DH_CHECK_P_NOT_SAFE_PRIME.
16004 (Check if this is true? OpenPGP calls them "strong".)
16008 * Merge the functionality of "dh" and "gendh" programs into a new program
16009 "dhparam". The old programs are retained for now but will handle DH keys
16010 (instead of parameters) in future.
16014 * Make the ciphers, s_server and s_client programs check the return values
16015 when a new cipher list is set.
16019 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16020 ciphers. Before when the 56bit ciphers were enabled the sorting was
16023 The syntax for the cipher sorting has been extended to support sorting by
16024 cipher-strength (using the strength_bits hard coded in the tables).
16025 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16027 Fix a bug in the cipher-command parser: when supplying a cipher command
16028 string with an "undefined" symbol (neither command nor alphanumeric
16029 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16030 an error is flagged.
16032 Due to the strength-sorting extension, the code of the
16033 ssl_create_cipher_list() function was completely rearranged. I hope that
16034 the readability was also increased :-)
16036 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16038 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16039 for the first serial number and places 2 in the serial number file. This
16040 avoids problems when the root CA is created with serial number zero and
16041 the first user certificate has the same issuer name and serial number
16046 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16047 the new code. Add documentation for this stuff.
16051 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16052 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16053 structures and behave in an analogous way to the X509v3 functions:
16054 they shouldn't be called directly but wrapper functions should be used
16057 So we also now have some wrapper functions that call the X509at functions
16058 when passed certificate requests. (TO DO: similar things can be done with
16059 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16060 things. Some of these need some d2i or i2d and print functionality
16061 because they handle more complex structures.)
16065 * Add missing #ifndefs that caused missing symbols when building libssl
16066 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16067 NO_RSA in `ssl/s2*.c`.
16069 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16071 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16072 has a return value which indicates the quality of the random data
16073 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16074 error queue. New function RAND_pseudo_bytes() generates output that is
16075 guaranteed to be unique but not unpredictable. RAND_add is like
16076 RAND_seed, but takes an extra argument for an entropy estimate
16077 (RAND_seed always assumes full entropy).
16081 * Do more iterations of Rabin-Miller probable prime test (specifically,
16082 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16083 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16084 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16085 false-positive rate of at most 2^-80 for random input.
16089 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16093 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16094 in the 0.9.5 release), this returns the chain
16095 from an X509_CTX structure with a dup of the stack and all
16096 the X509 reference counts upped: so the stack will exist
16097 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16100 Also make SSL_SESSION_print() print out the verify return
16105 * Add manpage for the pkcs12 command. Also change the default
16106 behaviour so MAC iteration counts are used unless the new
16107 -nomaciter option is used. This improves file security and
16108 only older versions of MSIE (4.0 for example) need it.
16112 * Honor the no-xxx Configure options when creating .DEF files.
16116 * Add PKCS#10 attributes to field table: challengePassword,
16117 unstructuredName and unstructuredAddress. These are taken from
16118 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16119 international characters are used.
16121 More changes to X509_ATTRIBUTE code: allow the setting of types
16122 based on strings. Remove the 'loc' parameter when adding
16123 attributes because these will be a SET OF encoding which is sorted
16128 * Initial changes to the 'req' utility to allow request generation
16129 automation. This will allow an application to just generate a template
16130 file containing all the field values and have req construct the
16133 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16134 used all over the place including certificate requests and PKCS#7
16135 structures. They are currently handled manually where necessary with
16136 some primitive wrappers for PKCS#7. The new functions behave in a
16137 manner analogous to the X509 extension functions: they allow
16138 attributes to be looked up by NID and added.
16140 Later something similar to the X509V3 code would be desirable to
16141 automatically handle the encoding, decoding and printing of the
16142 more complex types. The string types like challengePassword can
16143 be handled by the string table functions.
16145 Also modified the multi byte string table handling. Now there is
16146 a 'global mask' which masks out certain types. The table itself
16147 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16148 is useful when for example there is only one permissible type
16149 (as in countryName) and using the mask might result in no valid
16154 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16155 SSL_get_peer_finished to allow applications to obtain the latest
16156 Finished messages sent to the peer or expected from the peer,
16157 respectively. (SSL_get_peer_finished is usually the Finished message
16158 actually received from the peer, otherwise the protocol will be aborted.)
16160 As the Finished message are message digests of the complete handshake
16161 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16162 be used for external authentication procedures when the authentication
16163 provided by SSL/TLS is not desired or is not enough.
16167 * Enhanced support for Alpha Linux is added. Now ./config checks if
16168 the host supports BWX extension and if Compaq C is present on the
16169 $PATH. Just exploiting of the BWX extension results in 20-30%
16170 performance kick for some algorithms, e.g. DES and RC4 to mention
16171 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16176 * Add support for MS "fast SGC". This is arguably a violation of the
16177 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16178 weak crypto and after checking the certificate is SGC a second one
16179 with strong crypto. MS SGC stops the first handshake after receiving
16180 the server certificate message and sends a second client hello. Since
16181 a server will typically do all the time consuming operations before
16182 expecting any further messages from the client (server key exchange
16183 is the most expensive) there is little difference between the two.
16185 To get OpenSSL to support MS SGC we have to permit a second client
16186 hello message after we have sent server done. In addition we have to
16187 reset the MAC if we do get this second client hello.
16191 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16192 if a DER encoded private key is RSA or DSA traditional format. Changed
16193 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16194 format DER encoded private key. Newer code should use PKCS#8 format which
16195 has the key type encoded in the ASN1 structure. Added DER private key
16196 support to pkcs8 application.
16200 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16201 ciphersuites has been selected (as required by the SSL 3/TLS 1
16202 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16203 is set, we interpret this as a request to violate the specification
16204 (the worst that can happen is a handshake failure, and 'correct'
16205 behaviour would result in a handshake failure anyway).
16209 * In SSL_CTX_add_session, take into account that there might be multiple
16210 SSL_SESSION structures with the same session ID (e.g. when two threads
16211 concurrently obtain them from an external cache).
16212 The internal cache can handle only one SSL_SESSION with a given ID,
16213 so if there's a conflict, we now throw out the old one to achieve
16218 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16219 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16220 some routines that use cipher OIDs: some ciphers do not have OIDs
16221 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16226 * Simplify the trust setting structure and code. Now we just have
16227 two sequences of OIDs for trusted and rejected settings. These will
16228 typically have values the same as the extended key usage extension
16229 and any application specific purposes.
16231 The trust checking code now has a default behaviour: it will just
16232 check for an object with the same NID as the passed id. Functions can
16233 be provided to override either the default behaviour or the behaviour
16234 for a given id. SSL client, server and email already have functions
16235 in place for compatibility: they check the NID and also return "trusted"
16236 if the certificate is self signed.
16240 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16241 traditional format into an EVP_PKEY structure.
16245 * Add a password callback function PEM_cb() which either prompts for
16246 a password if usr_data is NULL or otherwise assumes it is a null
16247 terminated password. Allow passwords to be passed on command line
16248 environment or config files in a few more utilities.
16252 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16253 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16254 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16255 Update documentation.
16259 * Support for ASN1 "NULL" type. This could be handled before by using
16260 ASN1_TYPE but there wasn't any function that would try to read a NULL
16261 and produce an error if it couldn't. For compatibility we also have
16262 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16263 don't allocate anything because they don't need to.
16267 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16270 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16272 * Rebuild of the memory allocation routines used by OpenSSL code and
16273 possibly others as well. The purpose is to make an interface that
16274 provide hooks so anyone can build a separate set of allocation and
16275 deallocation routines to be used by OpenSSL, for example memory
16276 pool implementations, or something else, which was previously hard
16277 since Malloc(), Realloc() and Free() were defined as macros having
16278 the values malloc, realloc and free, respectively (except for Win32
16279 compilations). The same is provided for memory debugging code.
16280 OpenSSL already comes with functionality to find memory leaks, but
16281 this gives people a chance to debug other memory problems.
16283 With these changes, a new set of functions and macros have appeared:
16285 CRYPTO_set_mem_debug_functions() [F]
16286 CRYPTO_get_mem_debug_functions() [F]
16287 CRYPTO_dbg_set_options() [F]
16288 CRYPTO_dbg_get_options() [F]
16289 CRYPTO_malloc_debug_init() [M]
16291 The memory debug functions are NULL by default, unless the library
16292 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16293 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16294 gives the standard debugging functions that come with OpenSSL) or
16295 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16296 provided by the library user) must be used. When the standard
16297 debugging functions are used, CRYPTO_dbg_set_options can be used to
16298 request additional information:
16299 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16300 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16302 Also, things like CRYPTO_set_mem_functions will always give the
16303 expected result (the new set of functions is used for allocation
16304 and deallocation) at all times, regardless of platform and compiler
16307 To finish it up, some functions that were never use in any other
16308 way than through macros have a new API and new semantic:
16310 CRYPTO_dbg_malloc()
16311 CRYPTO_dbg_realloc()
16314 All macros of value have retained their old syntax.
16316 *Richard Levitte and Bodo Moeller*
16318 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16319 ordering of SMIMECapabilities wasn't in "strength order" and there
16320 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16325 * Some ASN1 types with illegal zero length encoding (INTEGER,
16326 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16328 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16330 * Merge in my S/MIME library for OpenSSL. This provides a simple
16331 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16332 functionality to handle multipart/signed properly) and a utility
16333 called 'smime' to call all this stuff. This is based on code I
16334 originally wrote for Celo who have kindly allowed it to be
16335 included in OpenSSL.
16339 * Add variants des_set_key_checked and des_set_key_unchecked of
16340 des_set_key (aka des_key_sched). Global variable des_check_key
16341 decides which of these is called by des_set_key; this way
16342 des_check_key behaves as it always did, but applications and
16343 the library itself, which was buggy for des_check_key == 1,
16344 have a cleaner way to pick the version they need.
16348 * New function PKCS12_newpass() which changes the password of a
16353 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16354 dynamic mix. In both cases the ids can be used as an index into the
16355 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16356 functions so they accept a list of the field values and the
16357 application doesn't need to directly manipulate the X509_TRUST
16362 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16367 * Modify the way the V3 extension code looks up extensions. This now
16368 works in a similar way to the object code: we have some "standard"
16369 extensions in a static table which is searched with OBJ_bsearch()
16370 and the application can add dynamic ones if needed. The file
16371 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16372 updated whenever a new extension is added to the core code and kept
16373 in ext_nid order. There is a simple program 'tabtest.c' which checks
16374 this. New extensions are not added too often so this file can readily
16375 be maintained manually.
16377 There are two big advantages in doing things this way. The extensions
16378 can be looked up immediately and no longer need to be "added" using
16379 X509V3_add_standard_extensions(): this function now does nothing.
16380 Side note: I get *lots* of email saying the extension code doesn't
16381 work because people forget to call this function.
16382 Also no dynamic allocation is done unless new extensions are added:
16383 so if we don't add custom extensions there is no need to call
16384 X509V3_EXT_cleanup().
16388 * Modify enc utility's salting as follows: make salting the default. Add a
16389 magic header, so unsalted files fail gracefully instead of just decrypting
16390 to garbage. This is because not salting is a big security hole, so people
16391 should be discouraged from doing it.
16395 * Fixes and enhancements to the 'x509' utility. It allowed a message
16396 digest to be passed on the command line but it only used this
16397 parameter when signing a certificate. Modified so all relevant
16398 operations are affected by the digest parameter including the
16399 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16400 DSA key was used because it didn't fix the digest.
16404 * Initial certificate chain verify code. Currently tests the untrusted
16405 certificates for consistency with the verify purpose (which is set
16406 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16408 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16409 this is because it will reject chains with invalid extensions whereas
16410 every previous version of OpenSSL and SSLeay made no checks at all.
16412 Trust code: checks the root CA for the relevant trust settings. Trust
16413 settings have an initial value consistent with the verify purpose: e.g.
16414 if the verify purpose is for SSL client use it expects the CA to be
16415 trusted for SSL client use. However the default value can be changed to
16416 permit custom trust settings: one example of this would be to only trust
16417 certificates from a specific "secure" set of CAs.
16419 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16420 which should be used for version portability: especially since the
16421 verify structure is likely to change more often now.
16423 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16424 to set them. If not set then assume SSL clients will verify SSL servers
16427 Two new options to the verify program: -untrusted allows a set of
16428 untrusted certificates to be passed in and -purpose which sets the
16429 intended purpose of the certificate. If a purpose is set then the
16430 new chain verify code is used to check extension consistency.
16434 * Support for the authority information access extension.
16438 * Modify RSA and DSA PEM read routines to transparently handle
16439 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16440 public keys in a format compatible with certificate
16441 SubjectPublicKeyInfo structures. Unfortunately there were already
16442 functions called *_PublicKey_* which used various odd formats so
16443 these are retained for compatibility: however the DSA variants were
16444 never in a public release so they have been deleted. Changed dsa/rsa
16445 utilities to handle the new format: note no releases ever handled public
16446 keys so we should be OK.
16448 The primary motivation for this change is to avoid the same fiasco
16449 that dogs private keys: there are several incompatible private key
16450 formats some of which are standard and some OpenSSL specific and
16451 require various evil hacks to allow partial transparent handling and
16452 even then it doesn't work with DER formats. Given the option anything
16453 other than PKCS#8 should be dumped: but the other formats have to
16454 stay in the name of compatibility.
16456 With public keys and the benefit of hindsight one standard format
16457 is used which works with EVP_PKEY, RSA or DSA structures: though
16458 it clearly returns an error if you try to read the wrong kind of key.
16460 Added a -pubkey option to the 'x509' utility to output the public key.
16461 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16462 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16463 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16464 that do the same as the `EVP_PKEY_assign_*()` except they up the
16465 reference count of the added key (they don't "swallow" the
16470 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16471 CRLs would fail if the file contained no certificates or no CRLs:
16472 added a new function to read in both types and return the number
16473 read: this means that if none are read it will be an error. The
16474 DER versions of the certificate and CRL reader would always fail
16475 because it isn't possible to mix certificates and CRLs in DER format
16476 without choking one or the other routine. Changed this to just read
16477 a certificate: this is the best we can do. Also modified the code
16478 in `apps/verify.c` to take notice of return codes: it was previously
16479 attempting to read in certificates from NULL pointers and ignoring
16480 any errors: this is one reason why the cert and CRL reader seemed
16481 to work. It doesn't check return codes from the default certificate
16482 routines: these may well fail if the certificates aren't installed.
16486 * Code to support otherName option in GeneralName.
16490 * First update to verify code. Change the verify utility
16491 so it warns if it is passed a self signed certificate:
16492 for consistency with the normal behaviour. X509_verify
16493 has been modified to it will now verify a self signed
16494 certificate if *exactly* the same certificate appears
16495 in the store: it was previously impossible to trust a
16496 single self signed certificate. This means that:
16497 openssl verify ss.pem
16498 now gives a warning about a self signed certificate but
16499 openssl verify -CAfile ss.pem ss.pem
16504 * For servers, store verify_result in SSL_SESSION data structure
16505 (and add it to external session representation).
16506 This is needed when client certificate verifications fails,
16507 but an application-provided verification callback (set by
16508 SSL_CTX_set_cert_verify_callback) allows accepting the session
16509 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16510 but returns 1): When the session is reused, we have to set
16511 ssl->verify_result to the appropriate error code to avoid
16514 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16516 * Fix a bug in the new PKCS#7 code: it didn't consider the
16517 case in PKCS7_dataInit() where the signed PKCS7 structure
16518 didn't contain any existing data because it was being created.
16520 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16522 * Add a salt to the key derivation routines in enc.c. This
16523 forms the first 8 bytes of the encrypted file. Also add a
16524 -S option to allow a salt to be input on the command line.
16528 * New function X509_cmp(). Oddly enough there wasn't a function
16529 to compare two certificates. We do this by working out the SHA1
16530 hash and comparing that. X509_cmp() will be needed by the trust
16535 * SSL_get1_session() is like SSL_get_session(), but increments
16536 the reference count in the SSL_SESSION returned.
16538 *Geoff Thorpe <geoff@eu.c2.net>*
16540 * Fix for 'req': it was adding a null to request attributes.
16541 Also change the X509_LOOKUP and X509_INFO code to handle
16542 certificate auxiliary information.
16546 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16551 * Add the possibility to add extra information to the memory leak
16552 detecting output, to form tracebacks, showing from where each
16553 allocation was originated: CRYPTO_push_info("constant string") adds
16554 the string plus current file name and line number to a per-thread
16555 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16556 is like calling CYRPTO_pop_info() until the stack is empty.
16557 Also updated memory leak detection code to be multi-thread-safe.
16561 * Add options -text and -noout to pkcs7 utility and delete the
16562 encryption options which never did anything. Update docs.
16566 * Add options to some of the utilities to allow the pass phrase
16567 to be included on either the command line (not recommended on
16568 OSes like Unix) or read from the environment. Update the
16569 manpages and fix a few bugs.
16573 * Add a few manpages for some of the openssl commands.
16577 * Fix the -revoke option in ca. It was freeing up memory twice,
16578 leaking and not finding already revoked certificates.
16582 * Extensive changes to support certificate auxiliary information.
16583 This involves the use of X509_CERT_AUX structure and X509_AUX
16584 functions. An X509_AUX function such as PEM_read_X509_AUX()
16585 can still read in a certificate file in the usual way but it
16586 will also read in any additional "auxiliary information". By
16587 doing things this way a fair degree of compatibility can be
16588 retained: existing certificates can have this information added
16589 using the new 'x509' options.
16591 Current auxiliary information includes an "alias" and some trust
16592 settings. The trust settings will ultimately be used in enhanced
16593 certificate chain verification routines: currently a certificate
16594 can only be trusted if it is self signed and then it is trusted
16599 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16600 The problem was that one of the replacement routines had not been working
16601 since SSLeay releases. For now the offending routine has been replaced
16602 with non-optimised assembler. Even so, this now gives around 95%
16603 performance improvement for 1024 bit RSA signs.
16607 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16608 handling. Most clients have the effective key size in bits equal to
16609 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16610 A few however don't do this and instead use the size of the decrypted key
16611 to determine the RC2 key length and the AlgorithmIdentifier to determine
16612 the effective key length. In this case the effective key length can still
16613 be 40 bits but the key length can be 168 bits for example. This is fixed
16614 by manually forcing an RC2 key into the EVP_PKEY structure because the
16615 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16616 the key length and effective key length are equal.
16620 * Add a bunch of functions that should simplify the creation of
16621 X509_NAME structures. Now you should be able to do:
16622 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16623 and have it automatically work out the correct field type and fill in
16624 the structures. The more adventurous can try:
16625 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16626 and it will (hopefully) work out the correct multibyte encoding.
16630 * Change the 'req' utility to use the new field handling and multibyte
16631 copy routines. Before the DN field creation was handled in an ad hoc
16632 way in req, ca, and x509 which was rather broken and didn't support
16633 BMPStrings or UTF8Strings. Since some software doesn't implement
16634 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16635 using the dirstring_type option. See the new comment in the default
16636 openssl.cnf for more info.
16640 * Make crypto/rand/md_rand.c more robust:
16641 - Assure unique random numbers after fork().
16642 - Make sure that concurrent threads access the global counter and
16643 md serializably so that we never lose entropy in them
16644 or use exactly the same state in multiple threads.
16645 Access to the large state is not always serializable because
16646 the additional locking could be a performance killer, and
16647 md should be large enough anyway.
16651 * New file `apps/app_rand.c` with commonly needed functionality
16652 for handling the random seed file.
16654 Use the random seed file in some applications that previously did not:
16656 dsaparam -genkey (which also ignored its '-rand' option),
16659 x509 (when signing).
16660 Except on systems with /dev/urandom, it is crucial to have a random
16661 seed file at least for key creation, DSA signing, and for DH exchanges;
16662 for RSA signatures we could do without one.
16664 gendh and gendsa (unlike genrsa) used to read only the first byte
16665 of each file listed in the '-rand' option. The function as previously
16666 found in genrsa is now in app_rand.c and is used by all programs
16667 that support '-rand'.
16671 * In RAND_write_file, use mode 0600 for creating files;
16672 don't just chmod when it may be too late.
16676 * Report an error from X509_STORE_load_locations
16677 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16681 * New function ASN1_mbstring_copy() this copies a string in either
16682 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16683 into an ASN1_STRING type. A mask of permissible types is passed
16684 and it chooses the "minimal" type to use or an error if not type
16689 * Add function equivalents to the various macros in asn1.h. The old
16690 macros are retained with an `M_` prefix. Code inside the library can
16691 use the `M_` macros. External code (including the openssl utility)
16692 should *NOT* in order to be "shared library friendly".
16696 * Add various functions that can check a certificate's extensions
16697 to see if it usable for various purposes such as SSL client,
16698 server or S/MIME and CAs of these types. This is currently
16699 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16700 verification. Also added a -purpose flag to x509 utility to
16701 print out all the purposes.
16705 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16710 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16711 for, obtain and decode and extension and obtain its critical flag.
16712 This allows all the necessary extension code to be handled in a
16713 single function call.
16717 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16718 platforms. See crypto/rc4/rc4_enc.c for further details.
16722 * New -noout option to asn1parse. This causes no output to be produced
16723 its main use is when combined with -strparse and -out to extract data
16724 from a file (which may not be in ASN.1 format).
16728 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16729 when producing the local key id.
16731 *Richard Levitte <levitte@stacken.kth.se>*
16733 * New option -dhparam in s_server. This allows a DH parameter file to be
16734 stated explicitly. If it is not stated then it tries the first server
16735 certificate file. The previous behaviour hard coded the filename
16740 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16741 a public key to be input or output. For example:
16742 openssl rsa -in key.pem -pubout -out pubkey.pem
16743 Also added necessary DSA public key functions to handle this.
16747 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16748 in the message. This was handled by allowing
16749 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16751 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16753 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16754 to the end of the strings whereas this didn't. This would cause problems
16755 if strings read with d2i_ASN1_bytes() were later modified.
16757 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16759 * Fix for base64 decode bug. When a base64 bio reads only one line of
16760 data and it contains EOF it will end up returning an error. This is
16761 caused by input 46 bytes long. The cause is due to the way base64
16762 BIOs find the start of base64 encoded data. They do this by trying a
16763 trial decode on each line until they find one that works. When they
16764 do a flag is set and it starts again knowing it can pass all the
16765 data directly through the decoder. Unfortunately it doesn't reset
16766 the context it uses. This means that if EOF is reached an attempt
16767 is made to pass two EOFs through the context and this causes the
16768 resulting error. This can also cause other problems as well. As is
16769 usual with these problems it takes *ages* to find and the fix is
16770 trivial: move one line.
16772 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16774 * Ugly workaround to get s_client and s_server working under Windows. The
16775 old code wouldn't work because it needed to select() on sockets and the
16776 tty (for keypresses and to see if data could be written). Win32 only
16777 supports select() on sockets so we select() with a 1s timeout on the
16778 sockets and then see if any characters are waiting to be read, if none
16779 are present then we retry, we also assume we can always write data to
16780 the tty. This isn't nice because the code then blocks until we've
16781 received a complete line of data and it is effectively polling the
16782 keyboard at 1s intervals: however it's quite a bit better than not
16783 working at all :-) A dedicated Windows application might handle this
16784 with an event loop for example.
16788 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16789 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16790 will be called when RSA_sign() and RSA_verify() are used. This is useful
16791 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16792 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16793 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16794 This necessitated the support of an extra signature type NID_md5_sha1
16795 for SSL signatures and modifications to the SSL library to use it instead
16796 of calling RSA_public_decrypt() and RSA_private_encrypt().
16800 * Add new -verify -CAfile and -CApath options to the crl program, these
16801 will lookup a CRL issuers certificate and verify the signature in a
16802 similar way to the verify program. Tidy up the crl program so it
16803 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16804 less strict. It will now permit CRL extensions even if it is not
16805 a V2 CRL: this will allow it to tolerate some broken CRLs.
16809 * Initialize all non-automatic variables each time one of the openssl
16810 sub-programs is started (this is necessary as they may be started
16811 multiple times from the "OpenSSL>" prompt).
16813 *Lennart Bang, Bodo Moeller*
16815 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16816 removing all other RSA functionality (this is what NO_RSA does). This
16817 is so (for example) those in the US can disable those operations covered
16818 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16823 * Non-copying interface to BIO pairs.
16824 (still largely untested)
16828 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16829 ASCII string. This was handled independently in various places before.
16833 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16834 UTF8 strings a character at a time.
16838 * Use client_version from client hello to select the protocol
16839 (s23_srvr.c) and for RSA client key exchange verification
16840 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16844 * Add various utility functions to handle SPKACs, these were previously
16845 handled by poking round in the structure internals. Added new function
16846 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16847 print, verify and generate SPKACs. Based on an original idea from
16848 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16852 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16856 * Allow the config file extension section to be overwritten on the
16857 command line. Based on an original idea from Massimiliano Pala
16858 <madwolf@comune.modena.it>. The new option is called -extensions
16859 and can be applied to ca, req and x509. Also -reqexts to override
16860 the request extensions in req and -crlexts to override the crl extensions
16865 * Add new feature to the SPKAC handling in ca. Now you can include
16866 the same field multiple times by preceding it by "XXXX." for example:
16869 this is the same syntax as used in the req config file.
16873 * Allow certificate extensions to be added to certificate requests. These
16874 are specified in a 'req_extensions' option of the req section of the
16875 config file. They can be printed out with the -text option to req but
16876 are otherwise ignored at present.
16880 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16881 data read consists of only the final block it would not decrypted because
16882 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16883 A misplaced 'break' also meant the decrypted final block might not be
16884 copied until the next read.
16888 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16889 a few extra parameters to the DH structure: these will be useful if
16890 for example we want the value of 'q' or implement X9.42 DH.
16894 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16895 provides hooks that allow the default DSA functions or functions on a
16896 "per key" basis to be replaced. This allows hardware acceleration and
16897 hardware key storage to be handled without major modification to the
16898 library. Also added low level modexp hooks and CRYPTO_EX structure and
16899 associated functions.
16903 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16904 as "read only": it can't be written to and the buffer it points to will
16905 not be freed. Reading from a read only BIO is much more efficient than
16906 a normal memory BIO. This was added because there are several times when
16907 an area of memory needs to be read from a BIO. The previous method was
16908 to create a memory BIO and write the data to it, this results in two
16909 copies of the data and an O(n^2) reading algorithm. There is a new
16910 function BIO_new_mem_buf() which creates a read only memory BIO from
16911 an area of memory. Also modified the PKCS#7 routines to use read only
16916 * Bugfix: ssl23_get_client_hello did not work properly when called in
16917 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16918 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16919 but a retry condition occurred while trying to read the rest.
16923 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16924 NID_pkcs7_encrypted by default: this was wrong since this should almost
16925 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16926 the encrypted data type: this is a more sensible place to put it and it
16927 allows the PKCS#12 code to be tidied up that duplicated this
16932 * Changed obj_dat.pl script so it takes its input and output files on
16933 the command line. This should avoid shell escape redirection problems
16938 * Initial support for certificate extension requests, these are included
16939 in things like Xenroll certificate requests. Included functions to allow
16940 extensions to be obtained and added.
16944 * -crlf option to s_client and s_server for sending newlines as
16945 CRLF (as required by many protocols).
16949 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16951 * Install libRSAglue.a when OpenSSL is built with RSAref.
16953 *Ralf S. Engelschall*
16955 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16957 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16959 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16964 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16965 DH parameters/keys (q is lost during that conversion, but the resulting
16966 DH parameters contain its length).
16968 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16969 much faster than DH_generate_parameters (which creates parameters
16970 where `p = 2*q + 1`), and also the smaller q makes DH computations
16971 much more efficient (160-bit exponentiation instead of 1024-bit
16972 exponentiation); so this provides a convenient way to support DHE
16973 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16974 utter importance to use
16975 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16977 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16978 when such DH parameters are used, because otherwise small subgroup
16979 attacks may become possible!
16983 * Avoid memory leak in i2d_DHparams.
16987 * Allow the -k option to be used more than once in the enc program:
16988 this allows the same encrypted message to be read by multiple recipients.
16992 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16993 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16994 it will always use the numerical form of the OID, even if it has a short
16999 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17000 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17001 otherwise bn_mod_exp was called. In the case of hardware keys for example
17002 no private key components need be present and it might store extra data
17003 in the RSA structure, which cannot be accessed from bn_mod_exp.
17004 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17005 private key operations.
17009 * Added support for SPARC Linux.
17013 * pem_password_cb function type incompatibly changed from
17014 typedef int pem_password_cb(char *buf, int size, int rwflag);
17016 ....(char *buf, int size, int rwflag, void *userdata);
17017 so that applications can pass data to their callbacks:
17018 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17019 additional void * argument, which is just handed through whenever
17020 the password callback is called.
17022 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17024 New function SSL_CTX_set_default_passwd_cb_userdata.
17026 Compatibility note: As many C implementations push function arguments
17027 onto the stack in reverse order, the new library version is likely to
17028 interoperate with programs that have been compiled with the old
17029 pem_password_cb definition (PEM_whatever takes some data that
17030 happens to be on the stack as its last argument, and the callback
17031 just ignores this garbage); but there is no guarantee whatsoever that
17034 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17035 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17036 problems not only on Windows, but also on some Unix platforms.
17037 To avoid problematic command lines, these definitions are now in an
17038 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17039 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17043 * MIPS III/IV assembler module is reimplemented.
17047 * More DES library cleanups: remove references to srand/rand and
17048 delete an unused file.
17052 * Add support for the free Netwide assembler (NASM) under Win32,
17053 since not many people have MASM (ml) and it can be hard to obtain.
17054 This is currently experimental but it seems to work OK and pass all
17055 the tests. Check out INSTALL.W32 for info.
17059 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17060 without temporary keys kept an extra copy of the server key,
17061 and connections with temporary keys did not free everything in case
17066 * New function RSA_check_key and new openssl rsa option -check
17067 for verifying the consistency of RSA keys.
17069 *Ulf Moeller, Bodo Moeller*
17071 * Various changes to make Win32 compile work:
17072 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17073 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17074 comparison" warnings.
17075 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17079 * Add a debugging option to PKCS#5 v2 key generation function: when
17080 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17081 derived keys are printed to stderr.
17085 * Copy the flags in ASN1_STRING_dup().
17087 *Roman E. Pavlov <pre@mo.msk.ru>*
17089 * The x509 application mishandled signing requests containing DSA
17090 keys when the signing key was also DSA and the parameters didn't match.
17092 It was supposed to omit the parameters when they matched the signing key:
17093 the verifying software was then supposed to automatically use the CA's
17094 parameters if they were absent from the end user certificate.
17096 Omitting parameters is no longer recommended. The test was also
17097 the wrong way round! This was probably due to unusual behaviour in
17098 EVP_cmp_parameters() which returns 1 if the parameters match.
17099 This meant that parameters were omitted when they *didn't* match and
17100 the certificate was useless. Certificates signed with 'ca' didn't have
17103 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17105 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17106 The interface is as follows:
17107 Applications can use
17108 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17109 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17110 "off" is now the default.
17111 The library internally uses
17112 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17113 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17114 to disable memory-checking temporarily.
17116 Some inconsistent states that previously were possible (and were
17117 even the default) are now avoided.
17119 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17120 with each memory chunk allocated; this is occasionally more helpful
17121 than just having a counter.
17123 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17125 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17130 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17131 which largely parallels "options", but is for changing API behaviour,
17132 whereas "options" are about protocol behaviour.
17133 Initial "mode" flags are:
17135 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17136 a single record has been written.
17137 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17138 retries use the same buffer location.
17139 (But all of the contents must be
17144 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17147 * Fix problems with no-hmac etc.
17149 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17151 * New functions RSA_get_default_method(), RSA_set_method() and
17152 RSA_get_method(). These allows replacement of RSA_METHODs without having
17153 to mess around with the internals of an RSA structure.
17157 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17158 Also really enable memory leak checks in openssl.c and in some
17161 *Chad C. Mulligan, Bodo Moeller*
17163 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17164 up the length of negative integers. This has now been simplified to just
17165 store the length when it is first determined and use it later, rather
17166 than trying to keep track of where data is copied and updating it to
17168 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17170 * Add a new function PKCS7_signatureVerify. This allows the verification
17171 of a PKCS#7 signature but with the signing certificate passed to the
17172 function itself. This contrasts with PKCS7_dataVerify which assumes the
17173 certificate is present in the PKCS#7 structure. This isn't always the
17174 case: certificates can be omitted from a PKCS#7 structure and be
17175 distributed by "out of band" means (such as a certificate database).
17179 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17180 function prototypes in pem.h, also change util/mkdef.pl to add the
17181 necessary function names.
17185 * mk1mf.pl (used by Windows builds) did not properly read the
17186 options set by Configure in the top level Makefile, and Configure
17187 was not even able to write more than one option correctly.
17188 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17192 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17193 file to be loaded from a BIO or FILE pointer. The BIO version will
17194 for example allow memory BIOs to contain config info.
17198 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17199 Whoever hopes to achieve shared-library compatibility across versions
17200 must use this, not the compile-time macro.
17201 (Exercise 0.9.4: Which is the minimum library version required by
17203 Note: All this applies only to multi-threaded programs, others don't
17208 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17209 through a BIO pair triggered the default case, i.e.
17210 SSLerr(...,SSL_R_UNKNOWN_STATE).
17214 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17215 can use the SSL library even if none of the specific BIOs is
17220 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17221 for the encoded length.
17223 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17225 * Add initial documentation of the X509V3 functions.
17229 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17230 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17231 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17232 secure PKCS#8 private key format with a high iteration count.
17236 * Fix determination of Perl interpreter: A perl or perl5
17237 *directory* in $PATH was also accepted as the interpreter.
17239 *Ralf S. Engelschall*
17241 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17242 wrong with it but it was very old and did things like calling
17243 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17244 unusual formatting.
17248 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17249 to use the new extension code.
17253 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17254 with macros. This should make it easier to change their form, add extra
17255 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17260 * Add to configuration table a new entry that can specify an alternative
17261 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17262 according to Mark Crispin <MRC@Panda.COM>.
17266 * DES CBC did not update the IV. Weird.
17270 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17271 Changing the behaviour of the former might break existing programs --
17272 where IV updating is needed, des_ncbc_encrypt can be used.
17275 * When bntest is run from "make test" it drives bc to check its
17276 calculations, as well as internally checking them. If an internal check
17277 fails, it needs to cause bc to give a non-zero result or make test carries
17278 on without noticing the failure. Fixed.
17282 * DES library cleanups.
17286 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17287 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17288 ciphers. NOTE: although the key derivation function has been verified
17289 against some published test vectors it has not been extensively tested
17290 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17295 * Instead of "mkdir -p", which is not fully portable, use new
17296 Perl script "util/mkdir-p.pl".
17300 * Rewrite the way password based encryption (PBE) is handled. It used to
17301 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17302 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17303 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17304 the 'parameter' field of the AlgorithmIdentifier is passed to the
17305 underlying key generation function so it must do its own ASN1 parsing.
17306 This has also changed the EVP_PBE_CipherInit() function which now has a
17307 'parameter' argument instead of literal salt and iteration count values
17308 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17312 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17313 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17314 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17315 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17316 value was just used as a "magic string" and not used directly its
17317 value doesn't matter.
17321 * Introduce some semblance of const correctness to BN. Shame C doesn't
17326 * "linux-sparc64" configuration (ultrapenguin).
17328 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17329 "linux-sparc" configuration.
17331 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17333 * config now generates no-xxx options for missing ciphers.
17337 * Support the EBCDIC character set (work in progress).
17338 File ebcdic.c not yet included because it has a different license.
17340 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17342 * Support BS2000/OSD-POSIX.
17344 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17346 * Make callbacks for key generation use `void *` instead of `char *`.
17350 * Make S/MIME samples compile (not yet tested).
17354 * Additional typesafe stacks.
17358 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17362 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17364 * New configuration variant "sco5-gcc".
17366 * Updated some demos.
17368 *Sean O Riordain, Wade Scholine*
17370 * Add missing BIO_free at exit of pkcs12 application.
17374 * Fix memory leak in conf.c.
17378 * Updates for Win32 to assembler version of MD5.
17382 * Set #! path to perl in `apps/der_chop` to where we found it
17383 instead of using a fixed path.
17387 * SHA library changes for irix64-mips4-cc.
17391 * Improvements for VMS support.
17395 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17397 * Bignum library bug fix. IRIX 6 passes "make test" now!
17398 This also avoids the problems with SC4.2 and unpatched SC5.
17400 *Andy Polyakov <appro@fy.chalmers.se>*
17402 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17403 These are required because of the typesafe stack would otherwise break
17404 existing code. If old code used a structure member which used to be STACK
17405 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17406 sk_num or sk_value it would produce an error because the num, data members
17407 are not present in STACK_OF. Now it just produces a warning. sk_set
17408 replaces the old method of assigning a value to sk_value
17409 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17410 that does this will no longer work (and should use sk_set instead) but
17411 this could be regarded as a "questionable" behaviour anyway.
17415 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17416 correctly handle encrypted S/MIME data.
17420 * Change type of various DES function arguments from des_cblock
17421 (which means, in function argument declarations, pointer to char)
17422 to des_cblock * (meaning pointer to array with 8 char elements),
17423 which allows the compiler to do more typechecking; it was like
17424 that back in SSLeay, but with lots of ugly casts.
17426 Introduce new type const_des_cblock.
17430 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17431 problems: find RecipientInfo structure that matches recipient certificate
17432 and initialise the ASN1 structures properly based on passed cipher.
17436 * Belatedly make the BN tests actually check the results.
17440 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17441 to and from BNs: it was completely broken. New compilation option
17442 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17443 key elements as negative integers.
17447 * Reorganize and speed up MD5.
17449 *Andy Polyakov <appro@fy.chalmers.se>*
17453 *Richard Levitte <richard@levitte.org>*
17455 * New option -out to asn1parse to allow the parsed structure to be
17456 output to a file. This is most useful when combined with the -strparse
17457 option to examine the output of things like OCTET STRINGS.
17461 * Make SSL library a little more fool-proof by not requiring any longer
17462 that `SSL_set_{accept,connect}_state` be called before
17463 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17464 in many applications because usually everything *appeared* to work as
17465 intended anyway -- now it really works as intended).
17469 * Move openssl.cnf out of lib/.
17473 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17474 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17475 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17477 *Ralf S. Engelschall*
17479 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17480 handle PKCS#7 enveloped data properly.
17482 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17484 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17485 copying pointers. The cert_st handling is changed by this in
17486 various ways (and thus what used to be known as ctx->default_cert
17487 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17488 any longer when s->cert does not give us what we need).
17489 ssl_cert_instantiate becomes obsolete by this change.
17490 As soon as we've got the new code right (possibly it already is?),
17491 we have solved a couple of bugs of the earlier code where s->cert
17492 was used as if it could not have been shared with other SSL structures.
17494 Note that using the SSL API in certain dirty ways now will result
17495 in different behaviour than observed with earlier library versions:
17496 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17497 does not influence s as it used to.
17499 In order to clean up things more thoroughly, inside SSL_SESSION
17500 we don't use CERT any longer, but a new structure SESS_CERT
17501 that holds per-session data (if available); currently, this is
17502 the peer's certificate chain and, for clients, the server's certificate
17503 and temporary key. CERT holds only those values that can have
17504 meaningful defaults in an SSL_CTX.
17508 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17509 from the internal representation. Various PKCS#7 fixes: remove some
17510 evil casts and set the enc_dig_alg field properly based on the signing
17515 * Allow PKCS#12 password to be set from the command line or the
17516 environment. Let 'ca' get its config file name from the environment
17517 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17522 * Allow certificate policies extension to use an IA5STRING for the
17523 organization field. This is contrary to the PKIX definition but
17524 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17529 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17530 without disallowing inline assembler and the like for non-pedantic builds.
17534 * Support Borland C++ builder.
17536 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17542 * SHA-1 cleanups and performance enhancements.
17544 *Andy Polyakov <appro@fy.chalmers.se>*
17546 * Sparc v8plus assembler for the bignum library.
17548 *Andy Polyakov <appro@fy.chalmers.se>*
17550 * Accept any -xxx and +xxx compiler options in Configure.
17554 * Update HPUX configuration.
17558 * Add missing `sk_<type>_unshift()` function to safestack.h
17560 *Ralf S. Engelschall*
17562 * New function SSL_CTX_use_certificate_chain_file that sets the
17563 "extra_cert"s in addition to the certificate. (This makes sense
17564 only for "PEM" format files, as chains as a whole are not
17569 * Support verify_depth from the SSL API.
17570 x509_vfy.c had what can be considered an off-by-one-error:
17571 Its depth (which was not part of the external interface)
17572 was actually counting the number of certificates in a chain;
17573 now it really counts the depth.
17577 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17578 instead of X509err, which often resulted in confusing error
17579 messages since the error codes are not globally unique
17580 (e.g. an alleged error in ssl3_accept when a certificate
17581 didn't match the private key).
17583 * New function SSL_CTX_set_session_id_context that allows to set a default
17584 value (so that you don't need SSL_set_session_id_context for each
17585 connection using the SSL_CTX).
17589 * OAEP decoding bug fix.
17593 * Support INSTALL_PREFIX for package builders, as proposed by
17598 * New Configure options "threads" and "no-threads". For systems
17599 where the proper compiler options are known (currently Solaris
17600 and Linux), "threads" is the default.
17604 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17608 * Install various scripts to $(OPENSSLDIR)/misc, not to
17609 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17610 such as /usr/local/bin.
17614 * "make linux-shared" to build shared libraries.
17616 *Niels Poppe <niels@netbox.org>*
17618 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17622 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17623 extension adding in x509 utility.
17627 * Remove NOPROTO sections and error code comments.
17631 * Partial rewrite of the DEF file generator to now parse the ANSI
17636 * New Configure options --prefix=DIR and --openssldir=DIR.
17640 * Complete rewrite of the error code script(s). It is all now handled
17641 by one script at the top level which handles error code gathering,
17642 header rewriting and C source file generation. It should be much better
17643 than the old method: it now uses a modified version of Ulf's parser to
17644 read the ANSI prototypes in all header files (thus the old K&R definitions
17645 aren't needed for error creation any more) and do a better job of
17646 translating function codes into names. The old 'ASN1 error code embedded
17647 in a comment' is no longer necessary and it doesn't use .err files which
17648 have now been deleted. Also the error code call doesn't have to appear all
17649 on one line (which resulted in some large lines...).
17653 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17657 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17658 0 (which usually indicates a closed connection), but continue reading.
17662 * Fix some race conditions.
17666 * Add support for CRL distribution points extension. Add Certificate
17667 Policies and CRL distribution points documentation.
17671 * Move the autogenerated header file parts to crypto/opensslconf.h.
17675 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17676 8 of keying material. Merlin has also confirmed interop with this fix
17677 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17679 *Merlin Hughes <merlin@baltimore.ie>*
17681 * Fix lots of warnings.
17683 *Richard Levitte <levitte@stacken.kth.se>*
17685 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17686 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17688 *Richard Levitte <levitte@stacken.kth.se>*
17690 * Fix problems with sizeof(long) == 8.
17692 *Andy Polyakov <appro@fy.chalmers.se>*
17694 * Change functions to ANSI C.
17698 * Fix typos in error codes.
17700 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17702 * Remove defunct assembler files from Configure.
17706 * SPARC v8 assembler BIGNUM implementation.
17708 *Andy Polyakov <appro@fy.chalmers.se>*
17710 * Support for Certificate Policies extension: both print and set.
17711 Various additions to support the r2i method this uses.
17715 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17716 return a const string when you are expecting an allocated buffer.
17720 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17721 types DirectoryString and DisplayText.
17725 * Add code to allow r2i extensions to access the configuration database,
17726 add an LHASH database driver and add several ctx helper functions.
17730 * Fix an evil bug in bn_expand2() which caused various BN functions to
17731 fail when they extended the size of a BIGNUM.
17735 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17736 support typesafe stack.
17740 * Fix typo in SSL_[gs]et_options().
17742 *Nils Frostberg <nils@medcom.se>*
17744 * Delete various functions and files that belonged to the (now obsolete)
17745 old X509V3 handling code.
17749 * New Configure option "rsaref".
17753 * Don't auto-generate pem.h.
17757 * Introduce type-safe ASN.1 SETs.
17761 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17763 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17765 * Introduce type-safe STACKs. This will almost certainly break lots of code
17766 that links with OpenSSL (well at least cause lots of warnings), but fear
17767 not: the conversion is trivial, and it eliminates loads of evil casts. A
17768 few STACKed things have been converted already. Feel free to convert more.
17769 In the fullness of time, I'll do away with the STACK type altogether.
17773 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17774 specified in `<certfile>` by updating the entry in the index.txt file.
17775 This way one no longer has to edit the index.txt file manually for
17776 revoking a certificate. The -revoke option does the gory details now.
17778 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17780 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17781 `-text` option at all and this way the `-noout -text` combination was
17782 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17784 *Ralf S. Engelschall*
17786 * Make sure a corresponding plain text error message exists for the
17787 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17788 verify callback function determined that a certificate was revoked.
17790 *Ralf S. Engelschall*
17792 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17793 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17794 all available ciphers including rc5, which was forgotten until now.
17795 In order to let the testing shell script know which algorithms
17796 are available, a new (up to now undocumented) command
17797 `openssl list-cipher-commands` is used.
17801 * Bugfix: s_client occasionally would sleep in select() when
17802 it should have checked SSL_pending() first.
17806 * New functions DSA_do_sign and DSA_do_verify to provide access to
17807 the raw DSA values prior to ASN.1 encoding.
17811 * Tweaks to Configure
17813 *Niels Poppe <niels@netbox.org>*
17815 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17820 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17824 * New config option to avoid instructions that are illegal on the 80386.
17825 The default code is faster, but requires at least a 486.
17829 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17830 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17831 same as SSL2_VERSION anyway.
17835 * New "-showcerts" option for s_client.
17839 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17840 application. Various cleanups and fixes.
17844 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17845 modify error routines to work internally. Add error codes and PBE init
17846 to library startup routines.
17850 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17851 packing functions to asn1 and evp. Changed function names and error
17852 codes along the way.
17856 * PKCS12 integration: and so it begins... First of several patches to
17857 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17858 objects to objects.h
17862 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17863 and display support for Thawte strong extranet extension.
17867 * Add LinuxPPC support.
17869 *Jeff Dubrule <igor@pobox.org>*
17871 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17872 bn_div_words in alpha.s.
17874 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17876 * Make sure the RSA OAEP test is skipped under -DRSAref because
17877 OAEP isn't supported when OpenSSL is built with RSAref.
17879 *Ulf Moeller <ulf@fitug.de>*
17881 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17882 so they no longer are missing under -DNOPROTO.
17884 *Soren S. Jorvang <soren@t.dk>*
17886 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17888 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17889 doesn't work when the session is reused. Coming soon!
17893 * Fix a security hole, that allows sessions to be reused in the wrong
17894 context thus bypassing client cert protection! All software that uses
17895 client certs and session caches in multiple contexts NEEDS PATCHING to
17896 allow session reuse! A fuller solution is in the works.
17898 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17900 * Some more source tree cleanups (removed obsolete files
17901 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17902 permission on "config" script to be executable) and a fix for the INSTALL
17905 *Ulf Moeller <ulf@fitug.de>*
17907 * Remove some legacy and erroneous uses of malloc, free instead of
17910 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17912 * Make rsa_oaep_test return non-zero on error.
17914 *Ulf Moeller <ulf@fitug.de>*
17916 * Add support for native Solaris shared libraries. Configure
17917 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17918 if someone would make that last step automatic.
17920 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17922 * ctx_size was not built with the right compiler during "make links". Fixed.
17926 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17927 except NULL ciphers". This means the default cipher list will no longer
17928 enable NULL ciphers. They need to be specifically enabled e.g. with
17929 the string "DEFAULT:eNULL".
17933 * Fix to RSA private encryption routines: if p < q then it would
17934 occasionally produce an invalid result. This will only happen with
17935 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17939 * Be less restrictive and allow also `perl util/perlpath.pl
17940 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17941 because this way one can also use an interpreter named `perl5` (which is
17942 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17943 installed as `perl`).
17945 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17947 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17949 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17951 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17952 advapi32.lib to Win32 build and change the pem test comparison
17953 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17954 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17955 and crypto/des/ede_cbcm_enc.c.
17959 * DES quad checksum was broken on big-endian architectures. Fixed.
17963 * Comment out two functions in bio.h that aren't implemented. Fix up the
17964 Win32 test batch file so it (might) work again. The Win32 test batch file
17965 is horrible: I feel ill....
17969 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17970 in e_os.h. Audit of header files to check ANSI and non ANSI
17971 sections: 10 functions were absent from non ANSI section and not exported
17972 from Windows DLLs. Fixed up libeay.num for new functions.
17976 * Make `openssl version` output lines consistent.
17978 *Ralf S. Engelschall*
17980 * Fix Win32 symbol export lists for BIO functions: Added
17981 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17982 to ms/libeay{16,32}.def.
17984 *Ralf S. Engelschall*
17986 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17987 fine under Unix and passes some trivial tests I've now added. But the
17988 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17989 added to make sure no one expects that this stuff really works in the
17990 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17991 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17994 *Ralf S. Engelschall*
17996 * Fix the generation of two part addresses in perl.
17998 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18000 * Add config entry for Linux on MIPS.
18002 *John Tobey <jtobey@channel1.com>*
18004 * Make links whenever Configure is run, unless we are on Windoze.
18008 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18009 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18014 * Add a useful kludge to allow package maintainers to specify compiler and
18015 other platforms details on the command line without having to patch the
18016 Configure script every time: One now can use
18017 `perl Configure <id>:<details>`,
18018 i.e. platform ids are allowed to have details appended
18019 to them (separated by colons). This is treated as there would be a static
18020 pre-configured entry in Configure's %table under key `<id>` with value
18021 `<details>` and `perl Configure <id>` is called. So, when you want to
18022 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18023 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18024 now, which overrides the FreeBSD-elf entry on-the-fly.
18026 *Ralf S. Engelschall*
18028 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18032 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18033 on the `perl Configure ...` command line. This way one can compile
18034 OpenSSL libraries with Position Independent Code (PIC) which is needed
18035 for linking it into DSOs.
18037 *Ralf S. Engelschall*
18039 * Remarkably, export ciphers were totally broken and no-one had noticed!
18044 * Cleaned up the LICENSE document: The official contact for any license
18045 questions now is the OpenSSL core team under openssl-core@openssl.org.
18046 And add a paragraph about the dual-license situation to make sure people
18047 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18048 to the OpenSSL toolkit.
18050 *Ralf S. Engelschall*
18052 * General source tree makefile cleanups: Made `making xxx in yyy...`
18053 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18054 Additionally cleaned up the `make links` target: Remove unnecessary
18055 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18056 to speed processing and no longer clutter the display with confusing
18057 stuff. Instead only the actually done links are displayed.
18059 *Ralf S. Engelschall*
18061 * Permit null encryption ciphersuites, used for authentication only. It used
18062 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18063 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18068 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18069 signed attributes when verifying signatures (this would break them),
18070 the detached data encoding was wrong and public keys obtained using
18071 X509_get_pubkey() weren't freed.
18075 * Add text documentation for the BUFFER functions. Also added a work around
18076 to a Win95 console bug. This was triggered by the password read stuff: the
18077 last character typed gets carried over to the next fread(). If you were
18078 generating a new cert request using 'req' for example then the last
18079 character of the passphrase would be CR which would then enter the first
18084 * Added the new 'Includes OpenSSL Cryptography Software' button as
18085 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18086 button and can be used by applications based on OpenSSL to show the
18087 relationship to the OpenSSL project.
18089 *Ralf S. Engelschall*
18091 * Remove confusing variables in function signatures in files
18092 ssl/ssl_lib.c and ssl/ssl.h.
18094 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18096 * Don't install bss_file.c under PREFIX/include/
18098 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18100 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18101 functions that return function pointers and has support for NT specific
18102 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18103 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18104 unsigned to signed types: this was killing the Win32 compile.
18108 * Add new certificate file to stack functions,
18109 SSL_add_dir_cert_subjects_to_stack() and
18110 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18111 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18112 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18113 This means that Apache-SSL and similar packages don't have to mess around
18114 to add as many CAs as they want to the preferred list.
18118 * Experiment with doxygen documentation. Currently only partially applied to
18120 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18121 openssl.doxy as the configuration file.
18125 * Get rid of remaining C++-style comments which strict C compilers hate.
18127 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18129 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18130 compiled in by default: it has problems with large keys.
18134 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18135 DH private keys and/or callback functions which directly correspond to
18136 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18137 is needed for applications which have to configure certificates on a
18138 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18140 For the RSA certificate situation is makes no difference, but
18141 for the DSA certificate situation this fixes the "no shared cipher"
18142 problem where the OpenSSL cipher selection procedure failed because the
18143 temporary keys were not overtaken from the context and the API provided
18144 no way to reconfigure them.
18145 The new functions now let applications reconfigure the stuff and they
18146 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18147 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18148 non-public-API function ssl_cert_instantiate() is used as a helper
18149 function and also to reduce code redundancy inside ssl_rsa.c.
18151 *Ralf S. Engelschall*
18153 * Move s_server -dcert and -dkey options out of the undocumented feature
18154 area because they are useful for the DSA situation and should be
18155 recognized by the users.
18157 *Ralf S. Engelschall*
18159 * Fix the cipher decision scheme for export ciphers: the export bits are
18160 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18161 SSL_EXP_MASK. So, the original variable has to be used instead of the
18162 already masked variable.
18164 *Richard Levitte <levitte@stacken.kth.se>*
18166 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18168 *Richard Levitte <levitte@stacken.kth.se>*
18170 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18171 from `int` to `unsigned int` because it is a length and initialized by
18172 EVP_DigestFinal() which expects an `unsigned int *`.
18174 *Richard Levitte <levitte@stacken.kth.se>*
18176 * Don't hard-code path to Perl interpreter on shebang line of Configure
18177 script. Instead use the usual Shell->Perl transition trick.
18179 *Ralf S. Engelschall*
18181 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18182 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18183 -noout -modulus` as it's already the case for `openssl rsa -noout
18184 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18185 currently the public key is printed (a decision which was already done by
18186 `openssl dsa -modulus` in the past) which serves a similar purpose.
18187 Additionally the NO_RSA no longer completely removes the whole -modulus
18188 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18191 *Ralf S. Engelschall*
18193 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18194 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18196 *Arne Ansper <arne@ats.cyber.ee>*
18198 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18199 to be added. Now both 'req' and 'ca' can use new objects defined in the
18204 * Add cool BIO that does syslog (or event log on NT).
18206 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18208 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18209 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18210 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18211 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18215 * Add preliminary config info for new extension code.
18219 * Make RSA_NO_PADDING really use no padding.
18221 *Ulf Moeller <ulf@fitug.de>*
18223 * Generate errors when private/public key check is done.
18227 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18228 for some CRL extensions and new objects added.
18232 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18233 key usage extension and fuller support for authority key id.
18237 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18238 padding method for RSA, which is recommended for new applications in PKCS
18239 #1 v2.0 (RFC 2437, October 1998).
18240 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18241 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18242 against Bleichbacher's attack on RSA.
18243 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18246 * Updates to the new SSL compression code
18248 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18250 * Fix so that the version number in the master secret, when passed
18251 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18252 (because the server will not accept higher), that the version number
18253 is 0x03,0x01, not 0x03,0x00
18255 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18257 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18258 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18259 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18263 * Support for RAW extensions where an arbitrary extension can be
18264 created by including its DER encoding. See `apps/openssl.cnf` for
18269 * Make sure latest Perl versions don't interpret some generated C array
18270 code as Perl array code in the crypto/err/err_genc.pl script.
18272 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18274 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18275 not many people have the assembler. Various Win32 compilation fixes and
18276 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18277 build instructions.
18281 * Modify configure script 'Configure' to automatically create crypto/date.h
18282 file under Win32 and also build pem.h from pem.org. New script
18283 util/mkfiles.pl to create the MINFO file on environments that can't do a
18284 'make files': perl util/mkfiles.pl >MINFO should work.
18288 * Major rework of DES function declarations, in the pursuit of correctness
18289 and purity. As a result, many evil casts evaporated, and some weirdness,
18290 too. You may find this causes warnings in your code. Zapping your evil
18291 casts will probably fix them. Mostly.
18295 * Fix for a typo in asn1.h. Bug fix to object creation script
18296 obj_dat.pl. It considered a zero in an object definition to mean
18297 "end of object": none of the objects in objects.h have any zeros
18298 so it wasn't spotted.
18300 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18302 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18303 Masking (CBCM). In the absence of test vectors, the best I have been able
18304 to do is check that the decrypt undoes the encrypt, so far. Send me test
18305 vectors if you have them.
18309 * Correct calculation of key length for export ciphers (too much space was
18310 allocated for null ciphers). This has not been tested!
18314 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18315 message is now correct (it understands "crypto" and "ssl" on its
18316 command line). There is also now an "update" option. This will update
18317 the util/ssleay.num and util/libeay.num files with any new functions.
18319 perl util/mkdef.pl crypto ssl update
18320 it will update them.
18324 * Overhauled the Perl interface:
18325 - ported BN stuff to OpenSSL's different BN library
18326 - made the perl/ source tree CVS-aware
18327 - renamed the package from SSLeay to OpenSSL (the files still contain
18328 their history because I've copied them in the repository)
18329 - removed obsolete files (the test scripts will be replaced
18330 by better Test::Harness variants in the future)
18332 *Ralf S. Engelschall*
18334 * First cut for a very conservative source tree cleanup:
18335 1. merge various obsolete readme texts into doc/ssleay.txt
18336 where we collect the old documents and readme texts.
18337 2. remove the first part of files where I'm already sure that we no
18338 longer need them because of three reasons: either they are just temporary
18339 files which were left by Eric or they are preserved original files where
18340 I've verified that the diff is also available in the CVS via "cvs diff
18341 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18342 the crypto/md/ stuff).
18344 *Ralf S. Engelschall*
18346 * More extension code. Incomplete support for subject and issuer alt
18347 name, issuer and authority key id. Change the i2v function parameters
18348 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18349 what that's for :-) Fix to ASN1 macro which messed up
18350 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18354 * Preliminary support for ENUMERATED type. This is largely copied from the
18359 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18361 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18363 * Make sure `make rehash` target really finds the `openssl` program.
18365 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18367 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18368 like to hear about it if this slows down other processors.
18372 * Add CygWin32 platform information to Configure script.
18374 *Alan Batie <batie@aahz.jf.intel.com>*
18376 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18378 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18380 * New program nseq to manipulate netscape certificate sequences
18384 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18389 * Fixes to BN code. Previously the default was to define BN_RECURSION
18390 but the BN code had some problems that would cause failures when
18391 doing certificate verification and some other functions.
18393 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18395 * Add ASN1 and PEM code to support netscape certificate sequences.
18399 * Add ASN1 and PEM code to support netscape certificate sequences.
18403 * Add several PKIX and private extended key usage OIDs.
18407 * Modify the 'ca' program to handle the new extension code. Modify
18408 openssl.cnf for new extension format, add comments.
18412 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18413 and add a sample to openssl.cnf so req -x509 now adds appropriate
18418 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18419 error code, add initial support to X509_print() and x509 application.
18423 * Takes a deep breath and start adding X509 V3 extension support code. Add
18424 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18425 stuff is currently isolated and isn't even compiled yet.
18429 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18430 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18431 Removed the versions check from X509 routines when loading extensions:
18432 this allows certain broken certificates that don't set the version
18433 properly to be processed.
18437 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18438 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18439 can still be regenerated with "make depend".
18443 * Spelling mistake in C version of CAST-128.
18445 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18447 * Changes to the error generation code. The perl script err-code.pl
18448 now reads in the old error codes and retains the old numbers, only
18449 adding new ones if necessary. It also only changes the .err files if new
18450 codes are added. The makefiles have been modified to only insert errors
18451 when needed (to avoid needlessly modifying header files). This is done
18452 by only inserting errors if the .err file is newer than the auto generated
18453 C file. To rebuild all the error codes from scratch (the old behaviour)
18454 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18455 or delete all the .err files.
18459 * CAST-128 was incorrectly implemented for short keys. The C version has
18460 been fixed, but is untested. The assembler versions are also fixed, but
18461 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18462 to regenerate it if needed.
18463 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18464 Hagino <itojun@kame.net>*
18466 * File was opened incorrectly in randfile.c.
18468 *Ulf Möller <ulf@fitug.de>*
18470 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18471 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18472 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18473 al: it's just almost always a UTCTime. Note this patch adds new error
18474 codes so do a "make errors" if there are problems.
18478 * Correct Linux 1 recognition in config.
18480 *Ulf Möller <ulf@fitug.de>*
18482 * Remove pointless MD5 hash when using DSA keys in ca.
18484 *Anonymous <nobody@replay.com>*
18486 * Generate an error if given an empty string as a cert directory. Also
18487 generate an error if handed NULL (previously returned 0 to indicate an
18488 error, but didn't set one).
18490 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18492 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18496 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18497 parameters. This was causing a warning which killed off the Win32 compile.
18501 * Remove C++ style comments from crypto/bn/bn_local.h.
18503 *Neil Costigan <neil.costigan@celocom.com>*
18505 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18506 based on a text string, looking up short and long names and finally
18507 "dot" format. The "dot" format stuff didn't work. Added new function
18508 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18509 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18510 OID is not part of the table.
18514 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18515 X509_LOOKUP_by_alias().
18519 * Sort openssl functions by name.
18523 * Get the `gendsa` command working and add it to the `list` command. Remove
18524 encryption from sample DSA keys (in case anyone is interested the password
18529 * Make *all* `*_free` functions accept a NULL pointer.
18531 *Frans Heymans <fheymans@isaserver.be>*
18533 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18536 *Anonymous <nobody@replay.com>*
18538 * s_server should send the CAfile as acceptable CAs, not its own cert.
18540 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18542 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18544 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18546 * Temp key "for export" tests were wrong in s3_srvr.c.
18548 *Anonymous <nobody@replay.com>*
18550 * Add prototype for temp key callback functions
18551 SSL_CTX_set_tmp_{rsa,dh}_callback().
18555 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18556 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18560 * X509_name_add_entry() freed the wrong thing after an error.
18562 *Arne Ansper <arne@ats.cyber.ee>*
18564 * rsa_eay.c would attempt to free a NULL context.
18566 *Arne Ansper <arne@ats.cyber.ee>*
18568 * BIO_s_socket() had a broken should_retry() on Windoze.
18570 *Arne Ansper <arne@ats.cyber.ee>*
18572 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18574 *Arne Ansper <arne@ats.cyber.ee>*
18576 * Make sure the already existing X509_STORE->depth variable is initialized
18577 in X509_STORE_new(), but document the fact that this variable is still
18578 unused in the certificate verification process.
18580 *Ralf S. Engelschall*
18582 * Fix the various library and `apps/` files to free up pkeys obtained from
18583 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18587 * Fix reference counting in X509_PUBKEY_get(). This makes
18588 demos/maurice/example2.c work, amongst others, probably.
18590 *Steve Henson and Ben Laurie*
18592 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18593 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18594 are no longer created. This way we have a single and consistent command
18595 line interface `openssl <command>`, similar to `cvs <command>`.
18597 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18599 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18600 BIT STRING wrapper always have zero unused bits.
18604 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18608 * Make the top-level INSTALL documentation easier to understand.
18612 * Makefiles updated to exit if an error occurs in a sub-directory
18613 make (including if user presses ^C) [Paul Sutton]
18615 * Make Montgomery context stuff explicit in RSA data structure.
18619 * Fix build order of pem and err to allow for generated pem.h.
18623 * Fix renumbering bug in X509_NAME_delete_entry().
18627 * Enhanced the err-ins.pl script so it makes the error library number
18628 global and can add a library name. This is needed for external ASN1 and
18629 other error libraries.
18633 * Fixed sk_insert which never worked properly.
18637 * Fix ASN1 macros so they can handle indefinite length constructed
18638 EXPLICIT tags. Some non standard certificates use these: they can now
18643 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18644 into a single doc/ssleay.txt bundle. This way the information is still
18645 preserved but no longer messes up this directory. Now it's new room for
18646 the new set of documentation files.
18648 *Ralf S. Engelschall*
18650 * SETs were incorrectly DER encoded. This was a major pain, because they
18651 shared code with SEQUENCEs, which aren't coded the same. This means that
18652 almost everything to do with SETs or SEQUENCEs has either changed name or
18653 number of arguments.
18655 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18657 * Fix test data to work with the above.
18661 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18662 was already fixed by Eric for 0.9.1 it seems.
18664 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18666 * Autodetect FreeBSD3.
18670 * Fix various bugs in Configure. This affects the following platforms:
18674 unixware-2.0-pentium
18679 * Eliminate generated files from CVS. Reorder tests to regenerate files
18680 before they are needed.
18684 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18688 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18690 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18691 changed SSLeay to OpenSSL in version strings.
18693 *Ralf S. Engelschall*
18695 * Some fixups to the top-level documents.
18699 * Fixed the nasty bug where rsaref.h was not found under compile-time
18700 because the symlink to include/ was missing.
18702 *Ralf S. Engelschall*
18704 * Incorporated the popular no-RSA/DSA-only patches
18705 which allow to compile a RSA-free SSLeay.
18707 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18709 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18710 when "ssleay" is still not found.
18712 *Ralf S. Engelschall*
18714 * Added more platforms to Configure: Cray T3E, HPUX 11,
18716 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18718 * Updated the README file.
18720 *Ralf S. Engelschall*
18722 * Added various .cvsignore files in the CVS repository subdirs
18723 to make a "cvs update" really silent.
18725 *Ralf S. Engelschall*
18727 * Recompiled the error-definition header files and added
18728 missing symbols to the Win32 linker tables.
18730 *Ralf S. Engelschall*
18732 * Cleaned up the top-level documents;
18733 o new files: CHANGES and LICENSE
18734 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18735 o merged COPYRIGHT into LICENSE
18736 o removed obsolete TODO file
18737 o renamed MICROSOFT to INSTALL.W32
18739 *Ralf S. Engelschall*
18741 * Removed dummy files from the 0.9.1b source tree:
18743 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18744 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18745 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18746 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18747 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18750 *Ralf S. Engelschall*
18752 * Added various platform portability fixes.
18756 * The Genesis of the OpenSSL rpject:
18757 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18758 Young and Tim J. Hudson created while they were working for C2Net until
18761 *The OpenSSL Project*
18763 ### Changes between 0.9.0b and 0.9.1b [not released]
18765 * Updated a few CA certificates under certs/
18769 * Changed some BIGNUM api stuff.
18773 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18774 DGUX x86, Linux Alpha, etc.
18778 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18779 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18784 * Add -strparse option to asn1pars program which parses nested
18787 *Dr Stephen Henson <shenson@bigfoot.com>*
18789 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18793 * DSA fix for "ca" program.
18797 * Added "-genkey" option to "dsaparam" program.
18801 * Added RIPE MD160 (rmd160) message digest.
18805 * Added -a (all) option to "ssleay version" command.
18809 * Added PLATFORM define which is the id given to Configure.
18813 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18817 * Extended the ASN.1 parser routines.
18821 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18825 * Added a BN_CTX to the BN library.
18829 * Fixed the weak key values in DES library
18833 * Changed API in EVP library for cipher aliases.
18837 * Added support for RC2/64bit cipher.
18841 * Converted the lhash library to the crypto/mem.c functions.
18845 * Added more recognized ASN.1 object ids.
18849 * Added more RSA padding checks for SSL/TLS.
18853 * Added BIO proxy/filter functionality.
18857 * Added extra_certs to SSL_CTX which can be used
18858 send extra CA certificates to the client in the CA cert chain sending
18859 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18863 * Now Fortezza is denied in the authentication phase because
18864 this is key exchange mechanism is not supported by SSLeay at all.
18868 * Additional PKCS1 checks.
18872 * Support the string "TLSv1" for all TLS v1 ciphers.
18876 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18877 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18881 * Fixed a few memory leaks.
18885 * Fixed various code and comment typos.
18889 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18890 bytes sent in the client random.
18892 *Edward Bishop <ebishop@spyglass.com>*
18896 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18897 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18898 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18899 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18900 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18901 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18902 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18903 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18904 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18905 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18906 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18907 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18908 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18909 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18910 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18911 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18912 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18913 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18914 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18915 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18916 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18917 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18918 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18919 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18920 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18921 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18922 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18923 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18924 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18925 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18926 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18927 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18928 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18929 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18930 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18931 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18932 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18933 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18934 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18935 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18936 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18937 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18938 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18939 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18940 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18941 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18942 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18943 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18944 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18945 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18946 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18947 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18948 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18949 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18950 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18951 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18952 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18953 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18954 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18955 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18956 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18957 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18958 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18959 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18960 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18961 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18962 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18963 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18964 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18965 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18966 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18967 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18968 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18969 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18970 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18971 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18972 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18973 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18974 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18975 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18976 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18977 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18978 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18979 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18980 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18981 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18982 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18983 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18984 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18985 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18986 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18987 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18988 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18989 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18990 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18991 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18992 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18993 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18994 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18995 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18996 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18997 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18998 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18999 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19000 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19001 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19002 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19003 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19004 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19005 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19006 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19007 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19008 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19009 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19010 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19011 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19012 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19013 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19014 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19015 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19016 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19017 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19018 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19019 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19020 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19021 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19022 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19023 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19024 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19025 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19026 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19027 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19028 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19029 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19030 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19031 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19032 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19033 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19034 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19035 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19036 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19037 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19038 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19039 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19040 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19041 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19042 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19043 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19044 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19045 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19046 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19047 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19048 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19049 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19050 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19051 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19052 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19053 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19054 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19055 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19056 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19057 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655