4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
27 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
28 RAND_set_rand_method(). Provider based random number generators should
29 be used instead via EVP_RAND(3).
33 * The SRP APIs have been deprecated. The old APIs do not work via providers,
34 and there is no EVP interface to them. Unfortunately there is no replacement
35 for these APIs at this time.
39 * Add a compile time option to prevent the caching of provider fetched
40 algorithms. This is enabled by including the no-cached-fetch option
41 at configuration time.
45 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
46 Typically if OpenSSL has no EC or DH algorithms then it cannot support
47 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
48 through providers. Therefore third party providers may supply group
49 implementations even where there are no built-in ones. Attempting to create
50 TLS connections in such a build without also disabling TLSv1.3 at run time or
51 using third party provider groups may result in handshake failures. TLSv1.3
52 can be disabled at compile time using the "no-tls1_3" Configure option.
56 * The undocumented function X509_certificate_type() has been deprecated;
57 applications can use X509_get0_pubkey() and X509_get0_signature() to
58 get the same information.
62 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
63 functions. They are identical to BN_rand() and BN_rand_range()
68 * Deprecated the obsolete X9.31 RSA key generation related functions
69 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
70 BN_X931_generate_prime_ex().
74 * The default key generation method for the regular 2-prime RSA keys was
75 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
76 Conditions Based on Auxiliary Probable Primes). This method is slower
77 than the original method.
81 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
82 They are replaced with the BN_check_prime() function that avoids possible
83 misuse and always uses at least 64 rounds of the Miller-Rabin
84 primality test. At least 64 rounds of the Miller-Rabin test are now also
85 used for all prime generation, including RSA key generation.
86 This increases key generation time, especially for larger keys.
90 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
91 as they are not useful with non-deprecated functions.
95 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
96 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
97 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
98 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
99 were used to collect all necessary data to form a HTTP request, and to
100 perform the HTTP transfer with that request. With OpenSSL 3.0, the
101 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
102 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
103 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
104 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
105 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
106 OSSL_HTTP_REQ_CTX_set_max_response_length().
108 *Rich Salz and Richard Levitte*
110 * Validation of SM2 keys has been separated from the validation of regular EC
111 keys, allowing to improve the SM2 validation process to reject loaded private
112 keys that are not conforming to the SM2 ISO standard.
113 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
118 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
119 switches: a validation failure triggers an early exit, returning a failure
120 exit status to the parent process.
124 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
125 to ignore unknown ciphers.
129 * The `-cipher-commands` and `-digest-commands` options
130 of the command line utility `list` have been deprecated.
131 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
135 * All of the low level EC_KEY functions have been deprecated including:
137 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
138 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
139 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
140 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
141 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
142 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
143 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
144 EC_KEY_METHOD_get_verify,
145 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
146 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
147 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
148 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
149 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
150 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
151 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
152 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
153 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
154 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
155 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
156 Applications that need to implement an EC_KEY_METHOD need to consider
157 implementation of the functionality in a special provider.
158 For replacement of the functions manipulating the EC_KEY objects
159 see the EVP_PKEY-EC(7) manual page.
161 Additionally functions that read and write EC_KEY objects such as
162 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
163 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
164 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
165 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
166 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
167 have also been deprecated. Applications should instead use the
168 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
170 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
171 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
172 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
173 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
174 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
176 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
178 * Deprecated all the libcrypto and libssl error string loading
179 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
180 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
181 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
182 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
183 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
184 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
185 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
186 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
187 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
188 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
190 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
191 now loads error strings automatically.
195 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
196 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
197 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
198 are to be used by servers requiring ephemeral DH keys. Instead applications
199 should consider using the built-in DH parameters that are available by
200 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
201 necessary then applications can use the alternative functions
202 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
203 replacement for the "callback" functions. The callback was originally useful
204 in order to have different parameters for export and non-export ciphersuites.
205 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
206 functions should be replaced by one of the other methods described above.
210 * The `-crypt` option to the `passwd` command line tool has been removed.
214 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
219 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
221 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
222 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
223 The inverse ciphers use AES decryption for wrapping, and
224 AES encryption for unwrapping.
228 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
229 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
230 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
231 OpenSSL 3.0 these are replaced by the more generic functions
232 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
233 The old versions have been converted to deprecated macros that just call the
238 * The security callback, which can be customised by application code, supports
239 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
240 in the "other" parameter. In most places this is what is passed. All these
241 places occur server side. However there was one client side call of this
242 security operation and it passed a DH object instead. This is incorrect
243 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
244 of the other locations. Therefore this client side call has been changed to
245 pass an EVP_PKEY instead.
249 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
250 interface. Their functionality remains unchanged.
254 * Added new option for 'openssl list', '-providers', which will display the
255 list of loaded providers, their names, version and status. It optionally
256 displays their gettable parameters.
260 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
261 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
262 type is internally recognised so the workaround is no longer needed.
264 Functionality is still retained as it is, but will only work with
265 EVP_PKEYs with a legacy internal key.
269 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
270 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
274 * Changed all "STACK" functions to be macros instead of inline functions. Macro
275 parameters are still checked for type safety at compile time via helper
280 * Remove the RAND_DRBG API
282 The RAND_DRBG API did not fit well into the new provider concept as
283 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
284 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
285 and some of its API calls are rather low-level. This holds in particular
286 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
288 Adding a compatibility layer to continue supporting the RAND_DRBG API as
289 a legacy API for a regular deprecation period turned out to come at the
290 price of complicating the new provider API unnecessarily. Since the
291 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
294 *Paul Dale and Matthias St. Pierre*
296 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
297 as well as actual hostnames.
301 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
302 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
303 conversely, silently ignore DTLS protocol version bounds when configuring
304 TLS-based contexts. The commands can be repeated to set bounds of both
305 types. The same applies with the corresponding "min_protocol" and
306 "max_protocol" command-line switches, in case some application uses both TLS
309 SSL_CTX instances that are created for a fixed protocol version (e.g.
310 `TLSv1_server_method()`) also silently ignore version bounds. Previously
311 attempts to apply bounds to these protocol versions would result in an
312 error. Now only the "version-flexible" SSL_CTX instances are subject to
313 limits in configuration files in command-line options.
317 * Deprecated the `ENGINE` API. Engines should be replaced with providers
322 * Reworked the recorded ERR codes to make better space for system errors.
323 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
324 given code is a system error (true) or an OpenSSL error (false).
328 * Reworked the test perl framework to better allow parallel testing.
330 *Nicola Tuveri and David von Oheimb*
332 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
333 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
337 * 'Configure' has been changed to figure out the configuration target if
338 none is given on the command line. Consequently, the 'config' script is
339 now only a mere wrapper. All documentation is changed to only mention
342 *Rich Salz and Richard Levitte*
344 * Added a library context `OSSL_LIB_CTX` that applications as well as
345 other libraries can use to form a separate context within which
346 libcrypto operations are performed.
348 There are two ways this can be used:
350 - Directly, by passing a library context to functions that take
351 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
353 - Indirectly, by creating a new library context and then assigning
354 it as the new default, with `OSSL_LIB_CTX_set0_default`.
356 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
357 apart from the functions directly related to `OSSL_LIB_CTX`, accept
358 NULL to indicate that the default library context should be used.
360 Library code that changes the default library context using
361 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
362 second call before returning to the caller.
364 _(Note: the library context was initially called `OPENSSL_CTX` and
365 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
369 * Handshake now fails if Extended Master Secret extension is dropped
374 * Dropped interactive mode from the `openssl` program. From now on,
375 running it without arguments is equivalent to `openssl help`.
379 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
380 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
381 While the old function names have been retained for backward compatibility
382 they should not be used in new developments
383 because their return values are confusing: Unlike other `_cmp()` functions
384 they do not return 0 in case their arguments are equal.
388 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
389 `EC_GROUP_get_field_type()`.
393 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
394 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
395 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
396 Applications should rely on the library automatically assigning a suitable
397 EC_METHOD internally upon EC_GROUP construction.
401 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
402 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
403 assigned internally without application intervention.
404 Users of EC_GROUP_new() should switch to a different suitable constructor.
408 * Add CAdES-BES signature verification support, mostly derived
409 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
411 *Filipe Raimundo da Silva*
413 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
417 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
418 functions are not widely used and now OpenSSL automatically perform this
419 conversion when needed.
423 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
424 EC_KEY_precompute_mult(). These functions are not widely used and
425 applications should instead switch to named curves which OpenSSL has
426 hardcoded lookup tables for.
430 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
431 should instead use the L<EC_POINT_mul(3)> function.
435 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
436 that are not applicable to the new provider model. Applications should
437 instead use EVP_default_properties_is_fips_enabled() and
438 EVP_default_properties_enable_fips().
442 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
443 is set, an unexpected EOF is ignored, it pretends a close notify was received
444 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
448 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
449 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
450 used and applications should instead use the
451 L<EC_POINT_set_affine_coordinates(3)> and
452 L<EC_POINT_get_affine_coordinates(3)> functions.
456 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
457 arrays to be more easily constructed via a series of utility functions.
458 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
459 the various push functions and finally convert to a passable OSSL_PARAM
460 array using OSSL_PARAM_BLD_to_param().
464 * The security strength of SHA1 and MD5 based signatures in TLS has been
465 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
466 working at the default security level of 1 and instead requires security
467 level 0. The security level can be changed either using the cipher string
468 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
472 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
473 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
474 internal keys, if they correspond to one of those built in types.
478 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
479 contain a provider side internal key.
483 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
484 They are old functions that we don't use, and that you could disable with
485 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
489 * Project text documents not yet having a proper file name extension
490 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
491 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
495 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
496 have been converted to Markdown with the goal to produce documents
497 which not only look pretty when viewed online in the browser, but
498 remain well readable inside a plain text editor.
500 To achieve this goal, a 'minimalistic' Markdown style has been applied
501 which avoids formatting elements that interfere too much with the
502 reading flow in the text file. For example, it
504 * avoids [ATX headings][] and uses [setext headings][] instead
505 (which works for `<h1>` and `<h2>` headings only).
506 * avoids [inline links][] and uses [reference links][] instead.
507 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
509 [ATX headings]: https://github.github.com/gfm/#atx-headings
510 [setext headings]: https://github.github.com/gfm/#setext-headings
511 [inline links]: https://github.github.com/gfm/#inline-link
512 [reference links]: https://github.github.com/gfm/#reference-link
513 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
514 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
516 *Matthias St. Pierre*
518 * The test suite is changed to preserve results of each test recipe.
519 A new directory test-runs/ with subdirectories named like the
520 test recipes are created in the build tree for this purpose.
524 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
525 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
526 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
528 *David von Oheimb, Martin Peylo*
530 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
531 The legacy OCSP-focused and only partly documented API is retained for
532 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
536 * Added `util/check-format.pl`, a tool for checking adherence to the
537 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
538 The checks performed are incomplete and yield some false positives.
539 Still the tool should be useful for detecting most typical glitches.
543 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
544 If domain name resolution yields multiple IP addresses all of them are tried
545 after `connect()` failures.
549 * All of the low level RSA functions have been deprecated including:
551 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
552 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
553 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
554 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
555 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
556 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
557 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
558 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
559 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
560 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
561 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
562 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
563 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
564 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
565 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
566 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
567 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
568 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
569 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
570 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
571 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
572 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
573 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
574 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
575 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
576 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
577 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
578 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
580 Use of these low level functions has been informally discouraged for a long
581 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
582 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
583 L<EVP_PKEY_decrypt(3)>.
587 * X509 certificates signed using SHA1 are no longer allowed at security
589 In TLS/SSL the default security level is 1. It can be set either
590 using the cipher string with `@SECLEVEL`, or calling
591 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
592 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
594 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
595 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
596 options of the commands.
600 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
601 modified to use PKEY APIs. These commands are now in maintenance mode
602 and no new features will be added to them.
606 * The command line utility rsautl has been deprecated.
607 Instead use the pkeyutl program.
611 * The command line utilities genrsa and rsa have been modified to use PKEY
612 APIs. They now write PKCS#8 keys by default. These commands are now in
613 maintenance mode and no new features will be added to them.
617 * All of the low level DH functions have been deprecated including:
619 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
620 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
621 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
622 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
623 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
624 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
625 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
626 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
627 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
628 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
629 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
630 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
631 DH_meth_get_generate_params and DH_meth_set_generate_params.
633 Use of these low level functions has been informally discouraged for a long
634 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
635 and L<EVP_PKEY_derive(3)>.
637 Additionally functions that read and write DH objects such as d2i_DHparams,
638 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
639 functions have also been deprecated. Applications should instead use the
640 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
642 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
643 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
644 `EVP_PKEY_set1_DH()` are also deprecated.
645 Applications should instead either read or write an
646 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
647 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
649 *Paul Dale and Matt Caswell*
651 * All of the low level DSA functions have been deprecated including:
653 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
654 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
655 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
656 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
657 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
658 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
659 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
660 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
661 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
662 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
663 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
664 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
665 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
666 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
667 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
668 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
669 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
671 Use of these low level functions has been informally discouraged for a long
672 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
673 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
677 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
678 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
679 This means that applications don't have to look at the curve NID and
680 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
681 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
682 a no-op when the EVP_PKEY is already of the given type.
684 Parameter and key generation is also reworked to make it possible
685 to generate EVP_PKEY_SM2 parameters and keys without having to go
686 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
687 However, code that does the latter will still work as before.
691 * Deprecated low level ECDH and ECDSA functions. These include:
693 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
694 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
697 Use of these low level functions has been informally discouraged for a long
698 time. Instead applications should use the EVP_PKEY_derive(3),
699 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
703 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
704 and EVP_PKEY_decrypt() instead.
705 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
706 and EVP_PKEY_encrypt() instead.
710 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
711 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
712 a new formulation to include all the things it can be used for,
713 as well as words of caution.
717 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
718 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
722 * All of the low level HMAC functions have been deprecated including:
724 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
725 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
728 Use of these low level functions has been informally discouraged for a long
729 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
730 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
731 and L<EVP_MAC_final(3)>.
735 * Over two thousand fixes were made to the documentation, including:
736 - Common options (such as -rand/-writerand, TLS version control, etc)
737 were refactored and point to newly-enhanced descriptions in openssl.pod.
738 - Added style conformance for all options (with help from Richard Levitte),
739 documented all reported missing options, added a CI build to check
740 that all options are documented and that no unimplemented options
742 - Documented some internals, such as all use of environment variables.
743 - Addressed all internal broken L<> references.
747 * All of the low level CMAC functions have been deprecated including:
749 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
750 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
752 Use of these low level functions has been informally discouraged for a long
753 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
754 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
755 and L<EVP_MAC_final(3)>.
759 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
760 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
763 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
764 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
765 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
766 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
767 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
768 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
769 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
770 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
771 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
772 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
774 Use of these low level functions has been informally discouraged
775 for a long time. Applications should use the EVP_DigestInit_ex(3),
776 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
780 * Corrected the documentation of the return values from the `EVP_DigestSign*`
781 set of functions. The documentation mentioned negative values for some
782 errors, but this was never the case, so the mention of negative values
785 Code that followed the documentation and thereby check with something
786 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
790 * All of the low level cipher functions have been deprecated including:
792 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
793 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
794 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
795 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
796 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
797 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
798 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
799 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
800 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
801 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
802 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
803 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
804 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
805 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
806 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
807 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
808 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
809 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
810 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
811 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
812 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
813 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
814 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
815 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
816 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
817 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
818 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
819 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
820 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
822 Use of these low level functions has been informally discouraged for
823 a long time. Applications should use the high level EVP APIs, e.g.
824 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
825 equivalently named decrypt functions instead.
827 *Matt Caswell and Paul Dale*
829 * Removed include/openssl/opensslconf.h.in and replaced it with
830 include/openssl/configuration.h.in, which differs in not including
831 <openssl/macros.h>. A short header include/openssl/opensslconf.h
832 was added to include both.
834 This allows internal hacks where one might need to modify the set
835 of configured macros, for example this if deprecated symbols are
836 still supposed to be available internally:
838 #include <openssl/configuration.h>
840 #undef OPENSSL_NO_DEPRECATED
841 #define OPENSSL_SUPPRESS_DEPRECATED
843 #include <openssl/macros.h>
845 This should not be used by applications that use the exported
846 symbols, as that will lead to linking errors.
850 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
851 used in exponentiation with 512-bit moduli. No EC algorithms are
852 affected. Analysis suggests that attacks against 2-prime RSA1024,
853 3-prime RSA1536, and DSA1024 as a result of this defect would be very
854 difficult to perform and are not believed likely. Attacks against DH512
855 are considered just feasible. However, for an attack the target would
856 have to re-use the DH512 private key, which is not recommended anyway.
857 Also applications directly using the low level API BN_mod_exp may be
858 affected if they use BN_FLG_CONSTTIME.
863 * Most memory-debug features have been deprecated, and the functionality
864 replaced with no-ops.
868 * Added documentation for the STACK API.
872 * Introduced a new method type and API, OSSL_ENCODER, to
873 represent generic encoders. An implementation is expected to
874 be able to encode an object associated with a given name (such
875 as an algorithm name for an asymmetric key) into forms given by
876 implementation properties.
878 Encoders are primarily used from inside libcrypto, through
879 calls to functions like EVP_PKEY_print_private(),
880 PEM_write_bio_PrivateKey() and similar.
882 Encoders are specified in such a way that they can be made to
883 directly handle the provider side portion of an object, if this
884 provider side part comes from the same provider as the encoder
885 itself, but can also be made to handle objects in parametrized
886 form (as an OSSL_PARAM array of data). This allows a provider to
887 offer generic encoders as a service for any other provider.
891 * Added a .pragma directive to the syntax of configuration files, to
892 allow varying behavior in a supported and predictable manner.
893 Currently added pragma:
897 This allows dollar signs to be a keyword character unless it's
898 followed by a opening brace or parenthesis. This is useful for
899 platforms where dollar signs are commonly used in names, such as
900 volume names and system directory names on VMS.
904 * Added functionality to create an EVP_PKEY from user data. This
905 is effectively the same as creating a RSA, DH or DSA object and
906 then assigning them to an EVP_PKEY, but directly using algorithm
907 agnostic EVP functions. A benefit is that this should be future
908 proof for public key algorithms to come.
912 * Change the interpretation of the '--api' configuration option to
913 mean that this is a desired API compatibility level with no
914 further meaning. The previous interpretation, that this would
915 also mean to remove all deprecated symbols up to and including
916 the given version, no requires that 'no-deprecated' is also used
917 in the configuration.
919 When building applications, the desired API compatibility level
920 can be set with the OPENSSL_API_COMPAT macro like before. For
921 API compatibility version below 3.0, the old style numerical
922 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
923 For version 3.0 and on, the value is expected to be the decimal
924 value calculated from the major and minor version like this:
926 MAJOR * 10000 + MINOR * 100
930 -DOPENSSL_API_COMPAT=30000 For 3.0
931 -DOPENSSL_API_COMPAT=30200 For 3.2
933 To hide declarations that are deprecated up to and including the
934 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
935 given when building the application as well.
939 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
940 access to certificate and CRL stores via URIs and OSSL_STORE
943 This adds the following functions:
945 - X509_LOOKUP_store()
946 - X509_STORE_load_file()
947 - X509_STORE_load_path()
948 - X509_STORE_load_store()
949 - SSL_add_store_cert_subjects_to_stack()
950 - SSL_CTX_set_default_verify_store()
951 - SSL_CTX_load_verify_file()
952 - SSL_CTX_load_verify_dir()
953 - SSL_CTX_load_verify_store()
957 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
958 The presence of this system service is determined at run-time.
962 * Added functionality to create an EVP_PKEY context based on data
963 for methods from providers. This takes an algorithm name and a
964 property query string and simply stores them, with the intent
965 that any operation that uses this context will use those strings
966 to fetch the needed methods implicitly, thereby making the port
967 of application written for pre-3.0 OpenSSL easier.
971 * The undocumented function NCONF_WIN32() has been deprecated; for
972 conversion details see the HISTORY section of doc/man5/config.pod
976 * Introduced the new functions EVP_DigestSignInit_ex() and
977 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
978 EVP_DigestVerifyUpdate() have been converted to functions. See the man
979 pages for further details.
983 * Over two thousand fixes were made to the documentation, including:
984 adding missing command flags, better style conformance, documentation
987 *Rich Salz, Richard Levitte*
989 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
990 X25519, X448, Ed25519 and Ed448.
994 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
999 * Deprecated the public definition of `ERR_STATE` as well as the function
1000 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1005 * Added ERR functionality to give callers access to the stored function
1006 names that have replaced the older function code based functions.
1008 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1009 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1010 ERR_peek_error_all() and ERR_peek_last_error_all().
1012 These functions have become deprecated: ERR_get_error_line(),
1013 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1014 ERR_peek_last_error_line_data() and ERR_func_error_string().
1016 Users are recommended to use ERR_get_error_all(), or to pick information
1017 with ERR_peek functions and finish off with getting the error code by using
1022 * Extended testing to be verbose for failing tests only. The make variables
1023 VERBOSE_FAILURE or VF can be used to enable this:
1025 $ make VF=1 test # Unix
1026 $ mms /macro=(VF=1) test ! OpenVMS
1027 $ nmake VF=1 test # Windows
1031 * Added the `-copy_extensions` option to the `x509` command for use with
1032 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1033 all extensions in the request are copied to the certificate or vice versa.
1035 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1037 * Added the `-copy_extensions` option to the `req` command for use with
1038 `-x509`. When given with the `copy` or `copyall` argument,
1039 all extensions in the certification request are copied to the certificate.
1043 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1044 they generate are by default RFC 5280 compliant in the following sense:
1045 There is a subjectKeyIdentifier extension with a hash value of the public key
1046 and for not self-signed certs there is an authorityKeyIdentifier extension
1047 with a keyIdentifier field or issuer information identifying the signing key.
1048 This is done unless some configuration overrides the new default behavior,
1049 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1053 * Added several checks to `X509_verify_cert()` according to requirements in
1054 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1055 (which may be done by using the CLI option `-x509_strict`):
1056 * The basicConstraints of CA certificates must be marked critical.
1057 * CA certificates must explicitly include the keyUsage extension.
1058 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1059 * The issuer name of any certificate must not be empty.
1060 * The subject name of CA certs, certs with keyUsage crlSign,
1061 and certs without subjectAlternativeName must not be empty.
1062 * If a subjectAlternativeName extension is given it must not be empty.
1063 * The signatureAlgorithm field and the cert signature must be consistent.
1064 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1065 must not be marked critical.
1066 * The authorityKeyIdentifier must be given for X.509v3 certs
1067 unless they are self-signed.
1068 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1072 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1073 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1077 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1078 used even when parsing explicit parameters, when loading a encoded key
1079 or calling `EC_GROUP_new_from_ecpkparameters()`/
1080 `EC_GROUP_new_from_ecparameters()`.
1081 This prevents bypass of security hardening and performance gains,
1082 especially for curves with specialized EC_METHODs.
1083 By default, if a key encoded with explicit parameters is loaded and later
1084 encoded, the output is still encoded with explicit parameters, even if
1085 internally a "named" EC_GROUP is used for computation.
1089 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1090 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1091 NULL. After this change, only the cofactor parameter can be NULL. It also
1092 does some minimal sanity checks on the passed order.
1097 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1098 An attack is simple, if the first CMS_recipientInfo is valid but the
1099 second CMS_recipientInfo is chosen ciphertext. If the second
1100 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1101 encryption key will be replaced by garbage, and the message cannot be
1102 decoded, but if the RSA decryption fails, the correct encryption key is
1103 used and the recipient will not notice the attack.
1104 As a work around for this potential attack the length of the decrypted
1105 key must be equal to the cipher default key length, in case the
1106 certifiate is not given and all recipientInfo are tried out.
1107 The old behaviour can be re-enabled in the CMS code by setting the
1108 CMS_DEBUG_DECRYPT flag.
1112 * Early start up entropy quality from the DEVRANDOM seed source has been
1113 improved for older Linux systems. The RAND subsystem will wait for
1114 /dev/random to be producing output before seeding from /dev/urandom.
1115 The seeded state is stored for future library initialisations using
1116 a system global shared memory segment. The shared memory identifier
1117 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1118 the desired value. The default identifier is 114.
1122 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1123 when primes for RSA keys are computed.
1124 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1125 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1126 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1127 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1128 This avoids possible fingerprinting of newly generated RSA modules.
1132 * Correct the extended master secret constant on EBCDIC systems. Without this
1133 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1134 negotiate EMS will fail. Unfortunately this also means that TLS connections
1135 between EBCDIC systems with this fix, and EBCDIC systems without this
1136 fix will fail if they negotiate EMS.
1140 * Changed the library initialisation so that the config file is now loaded
1141 by default. This was already the case for libssl. It now occurs for both
1142 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1143 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1147 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1148 where the former acts as a replacement for `ERR_put_error()`, and the
1149 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1150 `ERR_raise_data()` adds more flexibility by taking a format string and
1151 an arbitrary number of arguments following it, to be processed with
1156 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1157 to check if a named provider is loaded and available. When called, it
1158 will also activate all fallback providers if such are still present.
1162 * Enforce a minimum DH modulus size of 512 bits.
1166 * Changed DH parameters to generate the order q subgroup instead of 2q.
1167 Previously generated DH parameters are still accepted by DH_check
1168 but DH_generate_key works around that by clearing bit 0 of the
1169 private key for those. This avoids leaking bit 0 of the private key.
1173 * Significantly reduce secure memory usage by the randomness pools.
1177 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1182 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1183 algorithms. An implementation of a key exchange algorithm can be obtained
1184 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1185 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1186 the older EVP_PKEY_derive_init() function. See the man pages for the new
1187 functions for further details.
1191 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1195 * Removed the function names from error messages and deprecated the
1198 * Removed NextStep support and the macro OPENSSL_UNISTD
1202 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1203 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1204 Also removed "export var as function" capability; we do not export
1205 variables, only functions.
1209 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1210 an error and 1 indicating success. In previous versions of OpenSSL this
1211 was a void type. If a key was set longer than the maximum possible this
1216 * Support SM2 signing and verification schemes with X509 certificate.
1220 * Use SHA256 as the default digest for TS query in the `ts` app.
1224 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1225 This checks that the salt length is at least 128 bits, the derived key
1226 length is at least 112 bits, and that the iteration count is at least 1000.
1227 For backwards compatibility these checks are disabled by default in the
1228 default provider, but are enabled by default in the fips provider.
1229 To enable or disable these checks use the control
1230 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1234 * Default cipher lists/suites are now available via a function, the
1235 #defines are deprecated.
1239 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1240 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1241 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1245 * Join the directories crypto/x509 and crypto/x509v3
1249 * Added command 'openssl kdf' that uses the EVP_KDF API.
1253 * Added command 'openssl mac' that uses the EVP_MAC API.
1257 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1258 as default directories. Also added the command 'openssl info'
1259 for scripting purposes.
1263 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1264 deprecated. These undocumented functions were never integrated into the EVP
1265 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1266 Bi-directional IGE mode. These modes were never formally standardised and
1267 usage of these functions is believed to be very small. In particular
1268 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1269 is ever used. The security implications are believed to be minimal, but
1270 this issue was never fixed for backwards compatibility reasons. New code
1271 should not use these modes.
1275 * Add prediction resistance to the DRBG reseeding process.
1279 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1280 mandated by IEEE Std 1619-2018.
1284 * Added newline escaping functionality to a filename when using openssl dgst.
1285 This output format is to replicate the output format found in the `*sum`
1286 checksum programs. This aims to preserve backward compatibility.
1288 *Matt Eaton, Richard Levitte, and Paul Dale*
1290 * Removed the heartbeat message in DTLS feature, as it has very
1291 little usage and doesn't seem to fulfill a valuable purpose.
1292 The configuration option is now deprecated.
1296 * Changed the output of 'openssl {digestname} < file' to display the
1297 digest name in its output.
1301 * Added a new generic trace API which provides support for enabling
1302 instrumentation through trace output. This feature is mainly intended
1303 as an aid for developers and is disabled by default. To utilize it,
1304 OpenSSL needs to be configured with the `enable-trace` option.
1306 If the tracing API is enabled, the application can activate trace output
1307 by registering BIOs as trace channels for a number of tracing and debugging
1310 The `openssl` program has been expanded to enable any of the types
1311 available via environment variables defined by the user, and serves as
1312 one possible example on how to use this functionality.
1314 *Richard Levitte & Matthias St. Pierre*
1316 * Added build tests for C++. These are generated files that only do one
1317 thing, to include one public OpenSSL head file each. This tests that
1318 the public header files can be usefully included in a C++ application.
1320 This test isn't enabled by default. It can be enabled with the option
1321 'enable-buildtest-c++'.
1325 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1329 * Add KMAC to EVP_MAC.
1333 * Added property based algorithm implementation selection framework to
1338 * Added SCA hardening for modular field inversion in EC_GROUP through
1339 a new dedicated field_inv() pointer in EC_METHOD.
1340 This also addresses a leakage affecting conversions from projective
1341 to affine coordinates.
1343 *Billy Bob Brumley, Nicola Tuveri*
1345 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1346 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1347 those algorithms that were already supported through the EVP_PKEY API
1348 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1349 and scrypt are now wrappers that call EVP_KDF.
1353 * Build devcrypto engine as a dynamic engine.
1355 *Eneas U de Queiroz*
1357 * Add keyed BLAKE2 to EVP_MAC.
1361 * Fix a bug in the computation of the endpoint-pair shared secret used
1362 by DTLS over SCTP. This breaks interoperability with older versions
1363 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1364 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1365 interoperability with such broken implementations. However, enabling
1366 this switch breaks interoperability with correct implementations.
1368 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1369 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1373 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1377 * Change the license to the Apache License v2.0.
1381 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1383 - Major releases (indicated by incrementing the MAJOR release number)
1384 may introduce incompatible API/ABI changes.
1385 - Minor releases (indicated by incrementing the MINOR release number)
1386 may introduce new features but retain API/ABI compatibility.
1387 - Patch releases (indicated by incrementing the PATCH number)
1388 are intended for bug fixes and other improvements of existing
1389 features only (like improving performance or adding documentation)
1390 and retain API/ABI compatibility.
1394 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1398 * Remove the 'dist' target and add a tarball building script. The
1399 'dist' target has fallen out of use, and it shouldn't be
1400 necessary to configure just to create a source distribution.
1404 * Recreate the OS390-Unix config target. It no longer relies on a
1405 special script like it did for OpenSSL pre-1.1.0.
1409 * Instead of having the source directories listed in Configure, add
1410 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1415 * Add GMAC to EVP_MAC.
1419 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1423 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1424 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1425 to facilitate the continued use of MACs through raw private keys in
1426 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1430 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1431 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1435 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1436 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1437 are retained for backwards compatibility.
1441 * AES-XTS mode now enforces that its two keys are different to mitigate
1442 the attacked described in "Efficient Instantiations of Tweakable
1443 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1444 Details of this attack can be obtained from:
1445 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1449 * Rename the object files, i.e. give them other names than in previous
1450 versions. Their names now include the name of the final product, as
1451 well as its type mnemonic (bin, lib, shlib).
1455 * Added new option for 'openssl list', '-objects', which will display the
1456 list of built in objects, i.e. OIDs with names.
1460 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1461 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1466 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1467 improves application performance by removing data copies and providing
1468 applications with zero-copy system calls such as sendfile and splice.
1472 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1473 option is set, openssl cleanses (zeroize) plaintext bytes from
1474 internal buffers after delivering them to the application. Note,
1475 the application is still responsible for cleansing other copies
1476 (e.g.: data received by SSL_read(3)).
1480 * `PKCS12_parse` now maintains the order of the parsed certificates
1481 when outputting them via `*ca` (rather than reversing it).
1485 * Deprecated pthread fork support methods. These were unused so no
1486 replacement is required.
1488 - OPENSSL_fork_prepare()
1489 - OPENSSL_fork_parent()
1490 - OPENSSL_fork_child()
1497 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1499 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1501 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1502 create a unique hash value based on the issuer and serial number data
1503 contained within an X509 certificate. However it was failing to correctly
1504 handle any errors that may occur while parsing the issuer field (which might
1505 occur if the issuer field is maliciously constructed). This may subsequently
1506 result in a NULL pointer deref and a crash leading to a potential denial of
1512 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1513 padding mode to correctly check for rollback attacks. This is considered a
1514 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1519 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1520 functions. Previously they could overflow the output length argument in some
1521 cases where the input length is close to the maximum permissable length for
1522 an integer on the platform. In such cases the return value from the function
1523 call would be 1 (indicating success), but the output length value would be
1524 negative. This could cause applications to behave incorrectly or crash.
1529 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1530 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1531 could be exploited in a side channel attack to recover the password. Since
1532 the attack is local host only this is outside of the current OpenSSL
1533 threat model and therefore no CVE is assigned.
1535 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1540 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1542 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1543 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1544 If an attacker can control both items being compared then this could lead
1545 to a possible denial of service attack. OpenSSL itself uses the
1546 GENERAL_NAME_cmp function for two purposes:
1547 1) Comparing CRL distribution point names between an available CRL and a
1548 CRL distribution point embedded in an X509 certificate
1549 2) When verifying that a timestamp response token signer matches the
1550 timestamp authority name (exposed via the API functions
1551 TS_RESP_verify_response and TS_RESP_verify_token)
1556 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1558 * Certificates with explicit curve parameters are now disallowed in
1559 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1563 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1564 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1565 conversely, silently ignore DTLS protocol version bounds when configuring
1566 TLS-based contexts. The commands can be repeated to set bounds of both
1567 types. The same applies with the corresponding "min_protocol" and
1568 "max_protocol" command-line switches, in case some application uses both TLS
1571 SSL_CTX instances that are created for a fixed protocol version (e.g.
1572 TLSv1_server_method()) also silently ignore version bounds. Previously
1573 attempts to apply bounds to these protocol versions would result in an
1574 error. Now only the "version-flexible" SSL_CTX instances are subject to
1575 limits in configuration files in command-line options.
1579 * Handshake now fails if Extended Master Secret extension is dropped
1584 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1586 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1588 * Fixed segmentation fault in SSL_check_chain()
1589 Server or client applications that call the SSL_check_chain() function
1590 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1591 dereference as a result of incorrect handling of the
1592 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1593 or unrecognised signature algorithm is received from the peer. This could
1594 be exploited by a malicious peer in a Denial of Service attack.
1599 * Added AES consttime code for no-asm configurations
1600 an optional constant time support for AES was added
1601 when building openssl for no-asm.
1602 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1603 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1604 At this time this feature is by default disabled.
1605 It will be enabled by default in 3.0.
1609 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1611 * Revert the change of EOF detection while reading in libssl to avoid
1612 regressions in applications depending on the current way of reporting
1613 the EOF. As the existing method is not fully accurate the change to
1614 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1615 branch and will be present in the 3.0 release.
1619 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1620 when primes for RSA keys are computed.
1621 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1622 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1623 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1624 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1625 This avoids possible fingerprinting of newly generated RSA modules.
1629 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1631 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1632 while reading in libssl then we would report an error back to the
1633 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1634 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1635 therefore give a hint as to what went wrong.
1639 * Check that ed25519 and ed448 are allowed by the security level. Previously
1640 signature algorithms not using an MD were not being checked that they were
1641 allowed by the security level.
1645 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1646 was not quite right. The behaviour was not consistent between resumption
1647 and normal handshakes, and also not quite consistent with historical
1648 behaviour. The behaviour in various scenarios has been clarified and
1649 it has been updated to make it match historical behaviour as closely as
1654 * *[VMS only]* The header files that the VMS compilers include automatically,
1655 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1656 that the C++ compiler doesn't understand. This is a shortcoming in the
1657 compiler, but can be worked around with `__cplusplus` guards.
1659 C++ applications that use OpenSSL libraries must be compiled using the
1660 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1661 functions. Otherwise, only functions with symbols of less than 31
1662 characters can be used, as the linker will not be able to successfully
1663 resolve symbols with longer names.
1667 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1668 The presence of this system service is determined at run-time.
1672 * Added newline escaping functionality to a filename when using openssl dgst.
1673 This output format is to replicate the output format found in the `*sum`
1674 checksum programs. This aims to preserve backward compatibility.
1676 *Matt Eaton, Richard Levitte, and Paul Dale*
1678 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1683 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1685 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1686 number generator (RNG). This was intended to include protection in the
1687 event of a fork() system call in order to ensure that the parent and child
1688 processes did not share the same RNG state. However this protection was not
1689 being used in the default case.
1691 A partial mitigation for this issue is that the output from a high
1692 precision timer is mixed into the RNG state so the likelihood of a parent
1693 and child process sharing state is significantly reduced.
1695 If an application already calls OPENSSL_init_crypto() explicitly using
1696 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1699 *Matthias St. Pierre*
1701 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1702 used even when parsing explicit parameters, when loading a encoded key
1703 or calling `EC_GROUP_new_from_ecpkparameters()`/
1704 `EC_GROUP_new_from_ecparameters()`.
1705 This prevents bypass of security hardening and performance gains,
1706 especially for curves with specialized EC_METHODs.
1707 By default, if a key encoded with explicit parameters is loaded and later
1708 encoded, the output is still encoded with explicit parameters, even if
1709 internally a "named" EC_GROUP is used for computation.
1713 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1714 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1715 NULL. After this change, only the cofactor parameter can be NULL. It also
1716 does some minimal sanity checks on the passed order.
1721 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1722 An attack is simple, if the first CMS_recipientInfo is valid but the
1723 second CMS_recipientInfo is chosen ciphertext. If the second
1724 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1725 encryption key will be replaced by garbage, and the message cannot be
1726 decoded, but if the RSA decryption fails, the correct encryption key is
1727 used and the recipient will not notice the attack.
1728 As a work around for this potential attack the length of the decrypted
1729 key must be equal to the cipher default key length, in case the
1730 certifiate is not given and all recipientInfo are tried out.
1731 The old behaviour can be re-enabled in the CMS code by setting the
1732 CMS_DEBUG_DECRYPT flag.
1737 * Early start up entropy quality from the DEVRANDOM seed source has been
1738 improved for older Linux systems. The RAND subsystem will wait for
1739 /dev/random to be producing output before seeding from /dev/urandom.
1740 The seeded state is stored for future library initialisations using
1741 a system global shared memory segment. The shared memory identifier
1742 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1743 the desired value. The default identifier is 114.
1747 * Correct the extended master secret constant on EBCDIC systems. Without this
1748 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1749 negotiate EMS will fail. Unfortunately this also means that TLS connections
1750 between EBCDIC systems with this fix, and EBCDIC systems without this
1751 fix will fail if they negotiate EMS.
1755 * Use Windows installation paths in the mingw builds
1757 Mingw isn't a POSIX environment per se, which means that Windows
1758 paths should be used for installation.
1763 * Changed DH_check to accept parameters with order q and 2q subgroups.
1764 With order 2q subgroups the bit 0 of the private key is not secret
1765 but DH_generate_key works around that by clearing bit 0 of the
1766 private key for those. This avoids leaking bit 0 of the private key.
1770 * Significantly reduce secure memory usage by the randomness pools.
1774 * Revert the DEVRANDOM_WAIT feature for Linux systems
1776 The DEVRANDOM_WAIT feature added a select() call to wait for the
1777 /dev/random device to become readable before reading from the
1778 /dev/urandom device.
1780 It turned out that this change had negative side effects on
1781 performance which were not acceptable. After some discussion it
1782 was decided to revert this feature and leave it up to the OS
1783 resp. the platform maintainer to ensure a proper initialization
1784 during early boot time.
1786 *Matthias St. Pierre*
1788 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1790 * Add build tests for C++. These are generated files that only do one
1791 thing, to include one public OpenSSL head file each. This tests that
1792 the public header files can be usefully included in a C++ application.
1794 This test isn't enabled by default. It can be enabled with the option
1795 'enable-buildtest-c++'.
1799 * Enable SHA3 pre-hashing for ECDSA and DSA.
1803 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1804 This changes the size when using the `genpkey` command when no size is given.
1805 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1806 generation commands to use 2048 bits by default.
1810 * Reorganize the manual pages to consistently have RETURN VALUES,
1811 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1812 util/fix-doc-nits accordingly.
1814 *Paul Yang, Joshua Lock*
1816 * Add the missing accessor EVP_PKEY_get0_engine()
1820 * Have commands like `s_client` and `s_server` output the signature scheme
1821 along with other cipher suite parameters when debugging.
1825 * Make OPENSSL_config() error agnostic again.
1829 * Do the error handling in RSA decryption constant time.
1833 * Prevent over long nonces in ChaCha20-Poly1305.
1835 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1836 for every encryption operation. RFC 7539 specifies that the nonce value
1837 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1838 and front pads the nonce with 0 bytes if it is less than 12
1839 bytes. However it also incorrectly allows a nonce to be set of up to 16
1840 bytes. In this case only the last 12 bytes are significant and any
1841 additional leading bytes are ignored.
1843 It is a requirement of using this cipher that nonce values are
1844 unique. Messages encrypted using a reused nonce value are susceptible to
1845 serious confidentiality and integrity attacks. If an application changes
1846 the default nonce length to be longer than 12 bytes and then makes a
1847 change to the leading bytes of the nonce expecting the new value to be a
1848 new unique nonce then such an application could inadvertently encrypt
1849 messages with a reused nonce.
1851 Additionally the ignored bytes in a long nonce are not covered by the
1852 integrity guarantee of this cipher. Any application that relies on the
1853 integrity of these ignored leading bytes of a long nonce may be further
1854 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1855 is safe because no such use sets such a long nonce value. However user
1856 applications that use this cipher directly and set a non-default nonce
1857 length to be longer than 12 bytes may be vulnerable.
1859 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1865 * Add DEVRANDOM_WAIT feature for Linux systems
1867 On older Linux systems where the getrandom() system call is not available,
1868 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1869 Contrary to getrandom(), the /dev/urandom device will not block during
1870 early boot when the kernel CSPRNG has not been seeded yet.
1872 To mitigate this known weakness, use select() to wait for /dev/random to
1873 become readable before reading from /dev/urandom.
1875 * Ensure that SM2 only uses SM3 as digest algorithm
1879 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1881 * Change the info callback signals for the start and end of a post-handshake
1882 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1883 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1884 confused by this and assume that a TLSv1.2 renegotiation has started. This
1885 can break KeyUpdate handling. Instead we no longer signal the start and end
1886 of a post handshake message exchange (although the messages themselves are
1887 still signalled). This could break some applications that were expecting
1888 the old signals. However without this KeyUpdate is not usable for many
1893 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1895 * Timing vulnerability in DSA signature generation
1897 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1898 timing side channel attack. An attacker could use variations in the signing
1899 algorithm to recover the private key.
1901 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1906 * Timing vulnerability in ECDSA signature generation
1908 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1909 timing side channel attack. An attacker could use variations in the signing
1910 algorithm to recover the private key.
1912 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1917 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1918 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1919 of two gigabytes and the error handling improved.
1921 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1922 categorized as a normal bug, not a security issue, because the DRBG reseeds
1923 automatically and is fully functional even without additional randomness
1924 provided by the application.
1926 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1928 * Add a new ClientHello callback. Provides a callback interface that gives
1929 the application the ability to adjust the nascent SSL object at the
1930 earliest stage of ClientHello processing, immediately after extensions have
1931 been collected but before they have been processed. In particular, this
1932 callback can adjust the supported TLS versions in response to the contents
1937 * Add SM2 base algorithm support.
1941 * s390x assembly pack: add (improved) hardware-support for the following
1942 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1943 aes-cfb/cfb8, aes-ecb.
1947 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1948 parameter is no longer accepted, as it leads to a corrupt table. NULL
1949 pem_str is reserved for alias entries only.
1953 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1954 step for prime curves. The new implementation is based on formulae from
1955 differential addition-and-doubling in homogeneous projective coordinates
1956 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1957 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1958 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1959 to work in projective coordinates.
1961 *Billy Bob Brumley, Nicola Tuveri*
1963 * Change generating and checking of primes so that the error rate of not
1964 being prime depends on the intended use based on the size of the input.
1965 For larger primes this will result in more rounds of Miller-Rabin.
1966 The maximal error rate for primes with more than 1080 bits is lowered
1969 *Kurt Roeckx, Annie Yousar*
1971 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1975 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1976 moving between systems, and to avoid confusion when a Windows build is
1977 done with mingw vs with MSVC. For POSIX installs, there's still a
1978 symlink or copy named 'tsget' to avoid that confusion as well.
1982 * Revert blinding in ECDSA sign and instead make problematic addition
1983 length-invariant. Switch even to fixed-length Montgomery multiplication.
1987 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1988 step for binary curves. The new implementation is based on formulae from
1989 differential addition-and-doubling in mixed Lopez-Dahab projective
1990 coordinates, modified to independently blind the operands.
1992 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1994 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1995 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1996 EC_METHODs to implement their own specialized "ladder step", to take
1997 advantage of more favorable coordinate systems or more efficient
1998 differential addition-and-doubling algorithms.
2000 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2002 * Modified the random device based seed sources to keep the relevant
2003 file descriptors open rather than reopening them on each access.
2004 This allows such sources to operate in a chroot() jail without
2005 the associated device nodes being available. This behaviour can be
2006 controlled using RAND_keep_random_devices_open().
2010 * Numerous side-channel attack mitigations have been applied. This may have
2011 performance impacts for some algorithms for the benefit of improved
2012 security. Specific changes are noted in this change log by their respective
2017 * AIX shared library support overhaul. Switch to AIX "natural" way of
2018 handling shared libraries, which means collecting shared objects of
2019 different versions and bitnesses in one common archive. This allows to
2020 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2021 doesn't affect the way 3rd party applications are linked, only how
2022 multi-version installation is managed.
2026 * Make ec_group_do_inverse_ord() more robust and available to other
2027 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2028 mitigations are applied to the fallback BN_mod_inverse().
2029 When using this function rather than BN_mod_inverse() directly, new
2030 EC cryptosystem implementations are then safer-by-default.
2034 * Add coordinate blinding for EC_POINT and implement projective
2035 coordinate blinding for generic prime curves as a countermeasure to
2036 chosen point SCA attacks.
2038 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2040 * Add blinding to ECDSA and DSA signatures to protect against side channel
2041 attacks discovered by Keegan Ryan (NCC Group).
2045 * Enforce checking in the `pkeyutl` command to ensure that the input
2046 length does not exceed the maximum supported digest length when performing
2047 a sign, verify or verifyrecover operation.
2051 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2052 I/O in combination with something like select() or poll() will hang. This
2053 can be turned off again using SSL_CTX_clear_mode().
2054 Many applications do not properly handle non-application data records, and
2055 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2056 around the problems in those applications, but can also break some.
2057 It's recommended to read the manpages about SSL_read(), SSL_write(),
2058 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2059 SSL_CTX_set_read_ahead() again.
2063 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2064 now allow empty (zero character) pass phrases.
2068 * Apply blinding to binary field modular inversion and remove patent
2069 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2073 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2074 binary and prime elliptic curves.
2078 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2079 constant time fixed point multiplication.
2083 * Revise elliptic curve scalar multiplication with timing attack
2084 defenses: ec_wNAF_mul redirects to a constant time implementation
2085 when computing fixed point and variable point multiplication (which
2086 in OpenSSL are mostly used with secret scalars in keygen, sign,
2087 ECDH derive operations).
2088 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2091 * Updated CONTRIBUTING
2095 * Updated DRBG / RAND to request nonce and additional low entropy
2096 randomness from the system.
2098 *Matthias St. Pierre*
2100 * Updated 'openssl rehash' to use OpenSSL consistent default.
2104 * Moved the load of the ssl_conf module to libcrypto, which helps
2105 loading engines that libssl uses before libssl is initialised.
2109 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2113 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2115 *Ingo Schwarze, Rich Salz*
2117 * Added output of accepting IP address and port for 'openssl s_server'
2121 * Added a new API for TLSv1.3 ciphersuites:
2122 SSL_CTX_set_ciphersuites()
2123 SSL_set_ciphersuites()
2127 * Memory allocation failures consistently add an error to the error
2132 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2133 in libcrypto when run as setuid/setgid.
2137 * Load any config file by default when libssl is used.
2141 * Added new public header file <openssl/rand_drbg.h> and documentation
2142 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2144 *Matthias St. Pierre*
2146 * QNX support removed (cannot find contributors to get their approval
2147 for the license change).
2151 * TLSv1.3 replay protection for early data has been implemented. See the
2152 SSL_read_early_data() man page for further details.
2156 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2157 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2158 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2159 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2160 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2161 configuration has been separated out. See the ciphers man page or the
2162 SSL_CTX_set_ciphersuites() man page for more information.
2166 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2167 in responder mode now supports the new "-multi" option, which
2168 spawns the specified number of child processes to handle OCSP
2169 requests. The "-timeout" option now also limits the OCSP
2170 responder's patience to wait to receive the full client request
2171 on a newly accepted connection. Child processes are respawned
2172 as needed, and the CA index file is automatically reloaded
2173 when changed. This makes it possible to run the "ocsp" responder
2174 as a long-running service, making the OpenSSL CA somewhat more
2175 feature-complete. In this mode, most diagnostic messages logged
2176 after entering the event loop are logged via syslog(3) rather than
2181 * Added support for X448 and Ed448. Heavily based on original work by
2186 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2187 objects loaded. This adds the functions OSSL_STORE_expect() and
2188 OSSL_STORE_find() as well as needed tools to construct searches and
2189 get the search data out of them.
2193 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2194 version of OpenSSL should review their configuration settings to ensure
2195 that they are still appropriate for TLSv1.3. For further information see:
2196 <https://wiki.openssl.org/index.php/TLS1.3>
2200 * Grand redesign of the OpenSSL random generator
2202 The default RAND method now utilizes an AES-CTR DRBG according to
2203 NIST standard SP 800-90Ar1. The new random generator is essentially
2204 a port of the default random generator from the OpenSSL FIPS 2.0
2205 object module. It is a hybrid deterministic random bit generator
2206 using an AES-CTR bit stream and which seeds and reseeds itself
2207 automatically using trusted system entropy sources.
2209 Some of its new features are:
2210 - Support for multiple DRBG instances with seed chaining.
2211 - The default RAND method makes use of a DRBG.
2212 - There is a public and private DRBG instance.
2213 - The DRBG instances are fork-safe.
2214 - Keep all global DRBG instances on the secure heap if it is enabled.
2215 - The public and private DRBG instance are per thread for lock free
2218 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2220 * Changed Configure so it only says what it does and doesn't dump
2221 so much data. Instead, ./configdata.pm should be used as a script
2222 to display all sorts of configuration data.
2226 * Added processing of "make variables" to Configure.
2230 * Added SHA512/224 and SHA512/256 algorithm support.
2234 * The last traces of Netware support, first removed in 1.1.0, have
2239 * Get rid of Makefile.shared, and in the process, make the processing
2240 of certain files (rc.obj, or the .def/.map/.opt files produced from
2241 the ordinal files) more visible and hopefully easier to trace and
2242 debug (or make silent).
2246 * Make it possible to have environment variable assignments as
2247 arguments to config / Configure.
2251 * Add multi-prime RSA (RFC 8017) support.
2255 * Add SM3 implemented according to GB/T 32905-2016
2256 *Jack Lloyd <jack.lloyd@ribose.com>,*
2257 *Ronald Tse <ronald.tse@ribose.com>,*
2258 *Erick Borsboom <erick.borsboom@ribose.com>*
2260 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2261 as documented in RFC6066.
2262 Based on a patch from Tomasz Moń
2264 *Filipe Raimundo da Silva*
2266 * Add SM4 implemented according to GB/T 32907-2016.
2267 *Jack Lloyd <jack.lloyd@ribose.com>,*
2268 *Ronald Tse <ronald.tse@ribose.com>,*
2269 *Erick Borsboom <erick.borsboom@ribose.com>*
2271 * Reimplement -newreq-nodes and ERR_error_string_n; the
2272 original author does not agree with the license change.
2276 * Add ARIA AEAD TLS support.
2280 * Some macro definitions to support VS6 have been removed. Visual
2281 Studio 6 has not worked since 1.1.0
2285 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2286 without clearing the errors.
2290 * Add "atfork" functions. If building on a system that without
2291 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2292 requirements. The RAND facility now uses/requires this.
2300 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2301 not possible to disable entirely. However, it's still possible to
2302 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2305 To disable, configure with 'no-ui-console'. 'no-ui' is still
2306 possible to use as an alias. Check at compile time with the
2307 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2308 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2312 * Add a STORE module, which implements a uniform and URI based reader of
2313 stores that can contain keys, certificates, CRLs and numerous other
2314 objects. The main API is loosely based on a few stdio functions,
2315 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2316 OSSL_STORE_error and OSSL_STORE_close.
2317 The implementation uses backends called "loaders" to implement arbitrary
2318 URI schemes. There is one built in "loader" for the 'file' scheme.
2322 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2323 then adjusted to work on FreeBSD 8.4 as well.
2324 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2325 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2329 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2330 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2331 error code calls like this:
2333 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2335 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2336 that can be encoded in C. For the foreseeable future, this will only
2339 *Richard Levitte and Tim Hudson*
2341 * Removed BSD cryptodev engine.
2345 * Add a build target 'build_all_generated', to build all generated files
2346 and only that. This can be used to prepare everything that requires
2347 things like perl for a system that lacks perl and then move everything
2348 to that system and do the rest of the build there.
2352 * In the UI interface, make it possible to duplicate the user data. This
2353 can be used by engines that need to retain the data for a longer time
2354 than just the call where this user data is passed.
2358 * Ignore the '-named_curve auto' value for compatibility of applications
2361 *Tomáš Mráz <tmraz@fedoraproject.org>*
2363 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2364 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2365 alerts across multiple records (some of which could be empty). In practice
2366 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2367 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2368 support this at all. Supporting it adds significant complexity to the
2369 record layer, and its removal is unlikely to cause interoperability
2374 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2375 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2376 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2381 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2382 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2384 *Richard Levitte, Andy Polyakov*
2386 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2391 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2392 platform rather than 'mingw'.
2396 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2397 success if they are asked to add an object which already exists
2398 in the store. This change cascades to other functions which load
2399 certificates and CRLs.
2403 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2404 facilitate stack unwinding even from assembly subroutines.
2408 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2409 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2413 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2414 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2415 which is the minimum version we support.
2419 * Certificate time validation (X509_cmp_time) enforces stricter
2420 compliance with RFC 5280. Fractional seconds and timezone offsets
2421 are no longer allowed.
2425 * Add support for ARIA
2429 * s_client will now send the Server Name Indication (SNI) extension by
2430 default unless the new "-noservername" option is used. The server name is
2431 based on the host provided to the "-connect" option unless overridden by
2432 using "-servername".
2436 * Add support for SipHash
2440 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2441 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2442 prevent issues where no progress is being made and the peer continually
2443 sends unrecognised record types, using up resources processing them.
2447 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2448 using the algorithm defined in
2449 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2453 * Heartbeat support has been removed; the ABI is changed for now.
2455 *Richard Levitte, Rich Salz*
2457 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2461 * The RSA "null" method, which was partially supported to avoid patent
2462 issues, has been replaced to always returns NULL.
2469 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2471 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2472 used even when parsing explicit parameters, when loading a encoded key
2473 or calling `EC_GROUP_new_from_ecpkparameters()`/
2474 `EC_GROUP_new_from_ecparameters()`.
2475 This prevents bypass of security hardening and performance gains,
2476 especially for curves with specialized EC_METHODs.
2477 By default, if a key encoded with explicit parameters is loaded and later
2478 encoded, the output is still encoded with explicit parameters, even if
2479 internally a "named" EC_GROUP is used for computation.
2483 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2484 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2485 NULL. After this change, only the cofactor parameter can be NULL. It also
2486 does some minimal sanity checks on the passed order.
2491 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2492 An attack is simple, if the first CMS_recipientInfo is valid but the
2493 second CMS_recipientInfo is chosen ciphertext. If the second
2494 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2495 encryption key will be replaced by garbage, and the message cannot be
2496 decoded, but if the RSA decryption fails, the correct encryption key is
2497 used and the recipient will not notice the attack.
2498 As a work around for this potential attack the length of the decrypted
2499 key must be equal to the cipher default key length, in case the
2500 certifiate is not given and all recipientInfo are tried out.
2501 The old behaviour can be re-enabled in the CMS code by setting the
2502 CMS_DEBUG_DECRYPT flag.
2507 * Use Windows installation paths in the mingw builds
2509 Mingw isn't a POSIX environment per se, which means that Windows
2510 paths should be used for installation.
2515 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2517 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2518 This changes the size when using the `genpkey` command when no size is given.
2519 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2520 generation commands to use 2048 bits by default.
2524 * Prevent over long nonces in ChaCha20-Poly1305.
2526 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2527 for every encryption operation. RFC 7539 specifies that the nonce value
2528 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2529 and front pads the nonce with 0 bytes if it is less than 12
2530 bytes. However it also incorrectly allows a nonce to be set of up to 16
2531 bytes. In this case only the last 12 bytes are significant and any
2532 additional leading bytes are ignored.
2534 It is a requirement of using this cipher that nonce values are
2535 unique. Messages encrypted using a reused nonce value are susceptible to
2536 serious confidentiality and integrity attacks. If an application changes
2537 the default nonce length to be longer than 12 bytes and then makes a
2538 change to the leading bytes of the nonce expecting the new value to be a
2539 new unique nonce then such an application could inadvertently encrypt
2540 messages with a reused nonce.
2542 Additionally the ignored bytes in a long nonce are not covered by the
2543 integrity guarantee of this cipher. Any application that relies on the
2544 integrity of these ignored leading bytes of a long nonce may be further
2545 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2546 is safe because no such use sets such a long nonce value. However user
2547 applications that use this cipher directly and set a non-default nonce
2548 length to be longer than 12 bytes may be vulnerable.
2550 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2556 * Added SCA hardening for modular field inversion in EC_GROUP through
2557 a new dedicated field_inv() pointer in EC_METHOD.
2558 This also addresses a leakage affecting conversions from projective
2559 to affine coordinates.
2561 *Billy Bob Brumley, Nicola Tuveri*
2563 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2564 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2568 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2572 * Remove the 'dist' target and add a tarball building script. The
2573 'dist' target has fallen out of use, and it shouldn't be
2574 necessary to configure just to create a source distribution.
2578 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2580 * Timing vulnerability in DSA signature generation
2582 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2583 timing side channel attack. An attacker could use variations in the signing
2584 algorithm to recover the private key.
2586 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2591 * Timing vulnerability in ECDSA signature generation
2593 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2594 timing side channel attack. An attacker could use variations in the signing
2595 algorithm to recover the private key.
2597 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2602 * Add coordinate blinding for EC_POINT and implement projective
2603 coordinate blinding for generic prime curves as a countermeasure to
2604 chosen point SCA attacks.
2606 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2608 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2610 * Client DoS due to large DH parameter
2612 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2613 malicious server can send a very large prime value to the client. This will
2614 cause the client to spend an unreasonably long period of time generating a
2615 key for this prime resulting in a hang until the client has finished. This
2616 could be exploited in a Denial Of Service attack.
2618 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2623 * Cache timing vulnerability in RSA Key Generation
2625 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2626 a cache timing side channel attack. An attacker with sufficient access to
2627 mount cache timing attacks during the RSA key generation process could
2628 recover the private key.
2630 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2631 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2636 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2637 parameter is no longer accepted, as it leads to a corrupt table. NULL
2638 pem_str is reserved for alias entries only.
2642 * Revert blinding in ECDSA sign and instead make problematic addition
2643 length-invariant. Switch even to fixed-length Montgomery multiplication.
2647 * Change generating and checking of primes so that the error rate of not
2648 being prime depends on the intended use based on the size of the input.
2649 For larger primes this will result in more rounds of Miller-Rabin.
2650 The maximal error rate for primes with more than 1080 bits is lowered
2653 *Kurt Roeckx, Annie Yousar*
2655 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2659 * Add blinding to ECDSA and DSA signatures to protect against side channel
2660 attacks discovered by Keegan Ryan (NCC Group).
2664 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2665 now allow empty (zero character) pass phrases.
2669 * Certificate time validation (X509_cmp_time) enforces stricter
2670 compliance with RFC 5280. Fractional seconds and timezone offsets
2671 are no longer allowed.
2675 * Fixed a text canonicalisation bug in CMS
2677 Where a CMS detached signature is used with text content the text goes
2678 through a canonicalisation process first prior to signing or verifying a
2679 signature. This process strips trailing space at the end of lines, converts
2680 line terminators to CRLF and removes additional trailing line terminators
2681 at the end of a file. A bug in the canonicalisation process meant that
2682 some characters, such as form-feed, were incorrectly treated as whitespace
2683 and removed. This is contrary to the specification (RFC5485). This fix
2684 could mean that detached text data signed with an earlier version of
2685 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2686 signed with a fixed OpenSSL may fail to verify with an earlier version of
2687 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2688 and use the "-binary" flag (for the "cms" command line application) or set
2689 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2693 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2695 * Constructed ASN.1 types with a recursive definition could exceed the stack
2697 Constructed ASN.1 types with a recursive definition (such as can be found
2698 in PKCS7) could eventually exceed the stack given malicious input with
2699 excessive recursion. This could result in a Denial Of Service attack. There
2700 are no such structures used within SSL/TLS that come from untrusted sources
2701 so this is considered safe.
2703 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2709 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2711 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2712 effectively reduced to only comparing the least significant bit of each
2713 byte. This allows an attacker to forge messages that would be considered as
2714 authenticated in an amount of tries lower than that guaranteed by the
2715 security claims of the scheme. The module can only be compiled by the
2716 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2718 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2724 * Add a build target 'build_all_generated', to build all generated files
2725 and only that. This can be used to prepare everything that requires
2726 things like perl for a system that lacks perl and then move everything
2727 to that system and do the rest of the build there.
2731 * Backport SSL_OP_NO_RENGOTIATION
2733 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2734 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2735 changes this is no longer possible in 1.1.0. Therefore the new
2736 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2737 1.1.0 to provide equivalent functionality.
2739 Note that if an application built against 1.1.0h headers (or above) is run
2740 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2741 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2745 * Removed the OS390-Unix config target. It relied on a script that doesn't
2750 * rsaz_1024_mul_avx2 overflow bug on x86_64
2752 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2753 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2754 Analysis suggests that attacks against RSA and DSA as a result of this
2755 defect would be very difficult to perform and are not believed likely.
2756 Attacks against DH1024 are considered just feasible, because most of the
2757 work necessary to deduce information about a private key may be performed
2758 offline. The amount of resources required for such an attack would be
2759 significant. However, for an attack on TLS to be meaningful, the server
2760 would have to share the DH1024 private key among multiple clients, which is
2761 no longer an option since CVE-2016-0701.
2763 This only affects processors that support the AVX2 but not ADX extensions
2764 like Intel Haswell (4th generation).
2766 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2767 was originally found via the OSS-Fuzz project.
2772 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2774 * bn_sqrx8x_internal carry bug on x86_64
2776 There is a carry propagating bug in the x86_64 Montgomery squaring
2777 procedure. No EC algorithms are affected. Analysis suggests that attacks
2778 against RSA and DSA as a result of this defect would be very difficult to
2779 perform and are not believed likely. Attacks against DH are considered just
2780 feasible (although very difficult) because most of the work necessary to
2781 deduce information about a private key may be performed offline. The amount
2782 of resources required for such an attack would be very significant and
2783 likely only accessible to a limited number of attackers. An attacker would
2784 additionally need online access to an unpatched system using the target
2785 private key in a scenario with persistent DH parameters and a private
2786 key that is shared between multiple clients.
2788 This only affects processors that support the BMI1, BMI2 and ADX extensions
2789 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2791 This issue was reported to OpenSSL by the OSS-Fuzz project.
2796 * Malformed X.509 IPAddressFamily could cause OOB read
2798 If an X.509 certificate has a malformed IPAddressFamily extension,
2799 OpenSSL could do a one-byte buffer overread. The most likely result
2800 would be an erroneous display of the certificate in text format.
2802 This issue was reported to OpenSSL by the OSS-Fuzz project.
2807 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2809 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2810 platform rather than 'mingw'.
2814 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2815 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2816 which is the minimum version we support.
2820 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2822 * Encrypt-Then-Mac renegotiation crash
2824 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2825 negotiated where it was not in the original handshake (or vice-versa) then
2826 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2827 and servers are affected.
2829 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2834 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2836 * Truncated packet could crash via OOB read
2838 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2839 cipher is being used, then a truncated packet can cause that host to
2840 perform an out-of-bounds read, usually resulting in a crash.
2842 This issue was reported to OpenSSL by Robert Święcki of Google.
2847 * Bad (EC)DHE parameters cause a client crash
2849 If a malicious server supplies bad parameters for a DHE or ECDHE key
2850 exchange then this can result in the client attempting to dereference a
2851 NULL pointer leading to a client crash. This could be exploited in a Denial
2854 This issue was reported to OpenSSL by Guido Vranken.
2859 * BN_mod_exp may produce incorrect results on x86_64
2861 There is a carry propagating bug in the x86_64 Montgomery squaring
2862 procedure. No EC algorithms are affected. Analysis suggests that attacks
2863 against RSA and DSA as a result of this defect would be very difficult to
2864 perform and are not believed likely. Attacks against DH are considered just
2865 feasible (although very difficult) because most of the work necessary to
2866 deduce information about a private key may be performed offline. The amount
2867 of resources required for such an attack would be very significant and
2868 likely only accessible to a limited number of attackers. An attacker would
2869 additionally need online access to an unpatched system using the target
2870 private key in a scenario with persistent DH parameters and a private
2871 key that is shared between multiple clients. For example this can occur by
2872 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2873 similar to CVE-2015-3193 but must be treated as a separate problem.
2875 This issue was reported to OpenSSL by the OSS-Fuzz project.
2880 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2882 * ChaCha20/Poly1305 heap-buffer-overflow
2884 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2885 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2886 crash. This issue is not considered to be exploitable beyond a DoS.
2888 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2893 * CMS Null dereference
2895 Applications parsing invalid CMS structures can crash with a NULL pointer
2896 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2897 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2898 structure callback if an attempt is made to free certain invalid encodings.
2899 Only CHOICE structures using a callback which do not handle NULL value are
2902 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2907 * Montgomery multiplication may produce incorrect results
2909 There is a carry propagating bug in the Broadwell-specific Montgomery
2910 multiplication procedure that handles input lengths divisible by, but
2911 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2912 and DH private keys are impossible. This is because the subroutine in
2913 question is not used in operations with the private key itself and an input
2914 of the attacker's direct choice. Otherwise the bug can manifest itself as
2915 transient authentication and key negotiation failures or reproducible
2916 erroneous outcome of public-key operations with specially crafted input.
2917 Among EC algorithms only Brainpool P-512 curves are affected and one
2918 presumably can attack ECDH key negotiation. Impact was not analyzed in
2919 detail, because pre-requisites for attack are considered unlikely. Namely
2920 multiple clients have to choose the curve in question and the server has to
2921 share the private key among them, neither of which is default behaviour.
2922 Even then only clients that chose the curve will be affected.
2924 This issue was publicly reported as transient failures and was not
2925 initially recognized as a security issue. Thanks to Richard Morgan for
2926 providing reproducible case.
2931 * Removed automatic addition of RPATH in shared libraries and executables,
2932 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2936 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2938 * Fix Use After Free for large message sizes
2940 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2941 message larger than approx 16k is received then the underlying buffer to
2942 store the incoming message is reallocated and moved. Unfortunately a
2943 dangling pointer to the old location is left which results in an attempt to
2944 write to the previously freed location. This is likely to result in a
2945 crash, however it could potentially lead to execution of arbitrary code.
2947 This issue only affects OpenSSL 1.1.0a.
2949 This issue was reported to OpenSSL by Robert Święcki.
2954 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2956 * OCSP Status Request extension unbounded memory growth
2958 A malicious client can send an excessively large OCSP Status Request
2959 extension. If that client continually requests renegotiation, sending a
2960 large OCSP Status Request extension each time, then there will be unbounded
2961 memory growth on the server. This will eventually lead to a Denial Of
2962 Service attack through memory exhaustion. Servers with a default
2963 configuration are vulnerable even if they do not support OCSP. Builds using
2964 the "no-ocsp" build time option are not affected.
2966 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2971 * SSL_peek() hang on empty record
2973 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2974 sends an empty record. This could be exploited by a malicious peer in a
2975 Denial Of Service attack.
2977 This issue was reported to OpenSSL by Alex Gaynor.
2982 * Excessive allocation of memory in tls_get_message_header() and
2983 dtls1_preprocess_fragment()
2985 A (D)TLS message includes 3 bytes for its length in the header for the
2986 message. This would allow for messages up to 16Mb in length. Messages of
2987 this length are excessive and OpenSSL includes a check to ensure that a
2988 peer is sending reasonably sized messages in order to avoid too much memory
2989 being consumed to service a connection. A flaw in the logic of version
2990 1.1.0 means that memory for the message is allocated too early, prior to
2991 the excessive message length check. Due to way memory is allocated in
2992 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2993 to service a connection. This could lead to a Denial of Service through
2994 memory exhaustion. However, the excessive message length check still takes
2995 place, and this would cause the connection to immediately fail. Assuming
2996 that the application calls SSL_free() on the failed connection in a timely
2997 manner then the 21Mb of allocated memory will then be immediately freed
2998 again. Therefore the excessive memory allocation will be transitory in
2999 nature. This then means that there is only a security impact if:
3001 1) The application does not call SSL_free() in a timely manner in the event
3002 that the connection fails
3004 2) The application is working in a constrained environment where there is
3005 very little free memory
3007 3) The attacker initiates multiple connection attempts such that there are
3008 multiple connections in a state where memory has been allocated for the
3009 connection; SSL_free() has not yet been called; and there is insufficient
3010 memory to service the multiple requests.
3012 Except in the instance of (1) above any Denial Of Service is likely to be
3013 transitory because as soon as the connection fails the memory is
3014 subsequently freed again in the SSL_free() call. However there is an
3015 increased risk during this period of application crashes due to the lack of
3016 memory - which would then mean a more serious Denial of Service.
3018 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3019 (CVE-2016-6307 and CVE-2016-6308)
3023 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3024 had to be removed. Primary reason is that vendor assembler can't
3025 assemble our modules with -KPIC flag. As result it, assembly
3026 support, was not even available as option. But its lack means
3027 lack of side-channel resistant code, which is incompatible with
3028 security by todays standards. Fortunately gcc is readily available
3029 prepackaged option, which we firmly point at...
3033 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3035 * Windows command-line tool supports UTF-8 opt-in option for arguments
3036 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3037 (to any value) allows Windows user to access PKCS#12 file generated
3038 with Windows CryptoAPI and protected with non-ASCII password, as well
3039 as files generated under UTF-8 locale on Linux also protected with
3044 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3045 have been disabled by default and removed from DEFAULT, just like RC4.
3046 See the RC4 item below to re-enable both.
3050 * The method for finding the storage location for the Windows RAND seed file
3051 has changed. First we check %RANDFILE%. If that is not set then we check
3052 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3053 all else fails we fall back to C:\.
3057 * The EVP_EncryptUpdate() function has had its return type changed from void
3058 to int. A return of 0 indicates and error while a return of 1 indicates
3063 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3064 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3065 off the constant time implementation for RSA, DSA and DH have been made
3066 no-ops and deprecated.
3070 * Windows RAND implementation was simplified to only get entropy by
3071 calling CryptGenRandom(). Various other RAND-related tickets
3074 *Joseph Wylie Yandle, Rich Salz*
3076 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3077 and `OPENSSL_LH_`, respectively. The old names are available
3078 with API compatibility. They new names are now completely documented.
3082 * Unify TYPE_up_ref(obj) methods signature.
3083 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3084 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3085 int (instead of void) like all others TYPE_up_ref() methods.
3086 So now these methods also check the return value of CRYPTO_atomic_add(),
3087 and the validity of object reference counter.
3089 *fdasilvayy@gmail.com*
3091 * With Windows Visual Studio builds, the .pdb files are installed
3092 alongside the installed libraries and executables. For a static
3093 library installation, ossl_static.pdb is the associate compiler
3094 generated .pdb file to be used when linking programs.
3098 * Remove openssl.spec. Packaging files belong with the packagers.
3102 * Automatic Darwin/OSX configuration has had a refresh, it will now
3103 recognise x86_64 architectures automatically. You can still decide
3104 to build for a different bitness with the environment variable
3105 KERNEL_BITS (can be 32 or 64), for example:
3107 KERNEL_BITS=32 ./config
3111 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3112 256 bit AES and HMAC with SHA256.
3116 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3120 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3124 * To enable users to have their own config files and build file templates,
3125 Configure looks in the directory indicated by the environment variable
3126 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3127 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3128 name and is used as is.
3132 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3133 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3134 X509_CERT_FILE_CTX was removed.
3138 * "shared" builds are now the default. To create only static libraries use
3139 the "no-shared" Configure option.
3143 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3144 All of these option have not worked for some while and are fundamental
3149 * Make various cleanup routines no-ops and mark them as deprecated. Most
3150 global cleanup functions are no longer required because they are handled
3151 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3152 Explicitly de-initing can cause problems (e.g. where a library that uses
3153 OpenSSL de-inits, but an application is still using it). The affected
3154 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3155 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3156 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3157 COMP_zlib_cleanup().
3161 * --strict-warnings no longer enables runtime debugging options
3162 such as REF_DEBUG. Instead, debug options are automatically
3163 enabled with '--debug' builds.
3165 *Andy Polyakov, Emilia Käsper*
3167 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3168 have been moved out of the public header files. New functions for managing
3169 these have been added.
3173 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3174 objects have been moved out of the public header files. New
3175 functions for managing these have been added.
3179 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3180 have been moved out of the public header files. New functions for managing
3181 these have been added.
3185 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3186 moved out of the public header files. New functions for managing these
3191 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3195 * Removed the mk1mf build scripts.
3199 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3200 it is always safe to #include a header now.
3204 * Removed the aged BC-32 config and all its supporting scripts
3208 * Removed support for Ultrix, Netware, and OS/2.
3212 * Add support for HKDF.
3214 *Alessandro Ghedini*
3216 * Add support for blake2b and blake2s
3220 * Added support for "pipelining". Ciphers that have the
3221 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3222 encryptions/decryptions simultaneously. There are currently no built-in
3223 ciphers with this property but the expectation is that engines will be able
3224 to offer it to significantly improve throughput. Support has been extended
3225 into libssl so that multiple records for a single connection can be
3226 processed in one go (for >=TLS 1.1).
3230 * Added the AFALG engine. This is an async capable engine which is able to
3231 offload work to the Linux kernel. In this initial version it only supports
3232 AES128-CBC. The kernel must be version 4.1.0 or greater.
3236 * OpenSSL now uses a new threading API. It is no longer necessary to
3237 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3238 are two supported threading models: pthreads and windows threads. It is
3239 also possible to configure OpenSSL at compile time for "no-threads". The
3240 old threading API should no longer be used. The functions have been
3241 replaced with "no-op" compatibility macros.
3243 *Alessandro Ghedini, Matt Caswell*
3245 * Modify behavior of ALPN to invoke callback after SNI/servername
3246 callback, such that updates to the SSL_CTX affect ALPN.
3250 * Add SSL_CIPHER queries for authentication and key-exchange.
3254 * Changes to the DEFAULT cipherlist:
3255 - Prefer (EC)DHE handshakes over plain RSA.
3256 - Prefer AEAD ciphers over legacy ciphers.
3257 - Prefer ECDSA over RSA when both certificates are available.
3258 - Prefer TLSv1.2 ciphers/PRF.
3259 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3264 * Change the ECC default curve list to be this, in order: x25519,
3265 secp256r1, secp521r1, secp384r1.
3269 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3270 disabled by default. They can be re-enabled using the
3271 enable-weak-ssl-ciphers option to Configure.
3275 * If the server has ALPN configured, but supports no protocols that the
3276 client advertises, send a fatal "no_application_protocol" alert.
3277 This behaviour is SHALL in RFC 7301, though it isn't universally
3278 implemented by other servers.
3282 * Add X25519 support.
3283 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3284 for public and private key encoding using the format documented in
3285 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3286 key generation and key derivation.
3288 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3293 * Deprecate SRP_VBASE_get_by_user.
3294 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3295 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3296 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3297 seed, even if the seed is configured.
3299 Users should use SRP_VBASE_get1_by_user instead. Note that in
3300 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3301 also that even though configuring the SRP seed attempts to hide
3302 invalid usernames by continuing the handshake with fake
3303 credentials, this behaviour is not constant time and no strong
3304 guarantees are made that the handshake is indistinguishable from
3305 that of a valid user.
3309 * Configuration change; it's now possible to build dynamic engines
3310 without having to build shared libraries and vice versa. This
3311 only applies to the engines in `engines/`, those in `crypto/engine/`
3312 will always be built into libcrypto (i.e. "static").
3314 Building dynamic engines is enabled by default; to disable, use
3315 the configuration option "disable-dynamic-engine".
3317 The only requirements for building dynamic engines are the
3318 presence of the DSO module and building with position independent
3319 code, so they will also automatically be disabled if configuring
3320 with "disable-dso" or "disable-pic".
3322 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3323 are also taken away from openssl/opensslconf.h, as they are
3328 * Configuration change; if there is a known flag to compile
3329 position independent code, it will always be applied on the
3330 libcrypto and libssl object files, and never on the application
3331 object files. This means other libraries that use routines from
3332 libcrypto / libssl can be made into shared libraries regardless
3333 of how OpenSSL was configured.
3335 If this isn't desirable, the configuration options "disable-pic"
3336 or "no-pic" can be used to disable the use of PIC. This will
3337 also disable building shared libraries and dynamic engines.
3341 * Removed JPAKE code. It was experimental and has no wide use.
3345 * The INSTALL_PREFIX Makefile variable has been renamed to
3346 DESTDIR. That makes for less confusion on what this variable
3347 is for. Also, the configuration option --install_prefix is
3352 * Heartbeat for TLS has been removed and is disabled by default
3353 for DTLS; configure with enable-heartbeats. Code that uses the
3354 old #define's might need to be updated.
3356 *Emilia Käsper, Rich Salz*
3358 * Rename REF_CHECK to REF_DEBUG.
3362 * New "unified" build system
3364 The "unified" build system is aimed to be a common system for all
3365 platforms we support. With it comes new support for VMS.
3367 This system builds supports building in a different directory tree
3368 than the source tree. It produces one Makefile (for unix family
3369 or lookalikes), or one descrip.mms (for VMS).
3371 The source of information to make the Makefile / descrip.mms is
3372 small files called 'build.info', holding the necessary
3373 information for each directory with source to compile, and a
3374 template in Configurations, like unix-Makefile.tmpl or
3377 With this change, the library names were also renamed on Windows
3378 and on VMS. They now have names that are closer to the standard
3379 on Unix, and include the major version number, and in certain
3380 cases, the architecture they are built for. See "Notes on shared
3381 libraries" in INSTALL.
3383 We rely heavily on the perl module Text::Template.
3387 * Added support for auto-initialisation and de-initialisation of the library.
3388 OpenSSL no longer requires explicit init or deinit routines to be called,
3389 except in certain circumstances. See the OPENSSL_init_crypto() and
3390 OPENSSL_init_ssl() man pages for further information.
3394 * The arguments to the DTLSv1_listen function have changed. Specifically the
3395 "peer" argument is now expected to be a BIO_ADDR object.
3397 * Rewrite of BIO networking library. The BIO library lacked consistent
3398 support of IPv6, and adding it required some more extensive
3399 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3400 which hold all types of addresses and chains of address information.
3401 It also introduces a new API, with functions like BIO_socket,
3402 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3403 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3404 have been adapted accordingly.
3408 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3413 * CRIME protection: disable compression by default, even if OpenSSL is
3414 compiled with zlib enabled. Applications can still enable compression
3415 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3416 using the SSL_CONF library to configure compression.
3420 * The signature of the session callback configured with
3421 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3422 was explicitly marked as `const unsigned char*` instead of
3427 * Always DPURIFY. Remove the use of uninitialized memory in the
3428 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3432 * Removed many obsolete configuration items, including
3433 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3434 MD2_CHAR, MD2_INT, MD2_LONG
3436 IDEA_SHORT, IDEA_LONG
3437 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3439 *Rich Salz, with advice from Andy Polyakov*
3441 * Many BN internals have been moved to an internal header file.
3443 *Rich Salz with help from Andy Polyakov*
3445 * Configuration and writing out the results from it has changed.
3446 Files such as Makefile include/openssl/opensslconf.h and are now
3447 produced through general templates, such as Makefile.in and
3448 crypto/opensslconf.h.in and some help from the perl module
3451 Also, the center of configuration information is no longer
3452 Makefile. Instead, Configure produces a perl module in
3453 configdata.pm which holds most of the config data (in the hash
3454 table %config), the target data that comes from the target
3455 configuration in one of the `Configurations/*.conf` files (in
3460 * To clarify their intended purposes, the Configure options
3461 --prefix and --openssldir change their semantics, and become more
3462 straightforward and less interdependent.
3464 --prefix shall be used exclusively to give the location INSTALLTOP
3465 where programs, scripts, libraries, include files and manuals are
3466 going to be installed. The default is now /usr/local.
3468 --openssldir shall be used exclusively to give the default
3469 location OPENSSLDIR where certificates, private keys, CRLs are
3470 managed. This is also where the default openssl.cnf gets
3472 If the directory given with this option is a relative path, the
3473 values of both the --prefix value and the --openssldir value will
3474 be combined to become OPENSSLDIR.
3475 The default for --openssldir is INSTALLTOP/ssl.
3477 Anyone who uses --openssldir to specify where OpenSSL is to be
3478 installed MUST change to use --prefix instead.
3482 * The GOST engine was out of date and therefore it has been removed. An up
3483 to date GOST engine is now being maintained in an external repository.
3484 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3485 support for GOST ciphersuites (these are only activated if a GOST engine
3490 * EGD is no longer supported by default; use enable-egd when
3493 *Ben Kaduk and Rich Salz*
3495 * The distribution now has Makefile.in files, which are used to
3496 create Makefile's when Configure is run. *Configure must be run
3497 before trying to build now.*
3501 * The return value for SSL_CIPHER_description() for error conditions
3506 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3508 Obtaining and performing DNSSEC validation of TLSA records is
3509 the application's responsibility. The application provides
3510 the TLSA records of its choice to OpenSSL, and these are then
3511 used to authenticate the peer.
3513 The TLSA records need not even come from DNS. They can, for
3514 example, be used to implement local end-entity certificate or
3515 trust-anchor "pinning", where the "pin" data takes the form
3516 of TLSA records, which can augment or replace verification
3517 based on the usual WebPKI public certification authorities.
3521 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3522 continues to support deprecated interfaces in default builds.
3523 However, applications are strongly advised to compile their
3524 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3525 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3526 or the 1.1.0 releases.
3528 In environments in which all applications have been ported to
3529 not use any deprecated interfaces OpenSSL's Configure script
3530 should be used with the --api=1.1.0 option to entirely remove
3531 support for the deprecated features from the library and
3532 unconditionally disable them in the installed headers.
3533 Essentially the same effect can be achieved with the "no-deprecated"
3534 argument to Configure, except that this will always restrict
3535 the build to just the latest API, rather than a fixed API
3538 As applications are ported to future revisions of the API,
3539 they should update their compile-time OPENSSL_API_COMPAT define
3540 accordingly, but in most cases should be able to continue to
3541 compile with later releases.
3543 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3544 0x10000000L and 0x00908000L, respectively. However those
3545 versions did not support the OPENSSL_API_COMPAT feature, and
3546 so applications are not typically tested for explicit support
3547 of just the undeprecated features of either release.
3551 * Add support for setting the minimum and maximum supported protocol.
3552 It can bet set via the SSL_set_min_proto_version() and
3553 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3554 MaxProtocol. It's recommended to use the new APIs to disable
3555 protocols instead of disabling individual protocols using
3556 SSL_set_options() or SSL_CONF's Protocol. This change also
3557 removes support for disabling TLS 1.2 in the OpenSSL TLS
3558 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3562 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3566 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3567 and integrates ECDSA and ECDH functionality into EC. Implementations can
3568 now redirect key generation and no longer need to convert to or from
3571 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3572 include the ec.h header file instead.
3576 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3577 ciphers who are no longer supported and drops support the ephemeral RSA key
3578 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3582 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3583 opaque. For HMAC_CTX, the following constructors and destructors
3586 HMAC_CTX *HMAC_CTX_new(void);
3587 void HMAC_CTX_free(HMAC_CTX *ctx);
3589 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3590 destroy such methods has been added. See EVP_MD_meth_new(3) and
3591 EVP_CIPHER_meth_new(3) for documentation.
3594 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3595 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3596 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3597 an already created structure.
3598 2) For consistency with the majority of our object creators and
3599 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3600 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3601 for deprecated builds.
3605 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3606 cryptographic operations to be performed asynchronously as long as an
3607 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3608 further details. Libssl has also had this capability integrated with the
3609 introduction of the new mode SSL_MODE_ASYNC and associated error
3610 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3611 pages. This work was developed in partnership with Intel Corp.
3615 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3616 always enabled now. If you want to disable the support you should
3617 exclude it using the list of supported ciphers. This also means that the
3618 "-no_ecdhe" option has been removed from s_server.
3622 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3623 SSL_{CTX_}set1_curves() which can set a list.
3627 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3628 curve you want to support using SSL_{CTX_}set1_curves().
3632 * State machine rewrite. The state machine code has been significantly
3633 refactored in order to remove much duplication of code and solve issues
3634 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3635 further details). This change does have some associated API changes.
3636 Notably the SSL_state() function has been removed and replaced by
3637 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3638 SSL_set_state() has been removed altogether. The previous handshake states
3639 defined in ssl.h and ssl3.h have also been removed.
3643 * All instances of the string "ssleay" in the public API were replaced
3644 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3645 Some error codes related to internal RSA_eay API's were renamed.
3649 * The demo files in crypto/threads were moved to demo/threads.
3653 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3656 *Matt Caswell, Rich Salz*
3658 * New ASN.1 embed macro.
3660 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3661 structure is not allocated: it is part of the parent. That is instead of
3669 This reduces memory fragmentation and make it impossible to accidentally
3670 set a mandatory field to NULL.
3672 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3673 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3674 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3679 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3683 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3684 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3685 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3686 DES and RC4 ciphersuites.
3690 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3691 This changes the decoding behaviour for some invalid messages,
3692 though the change is mostly in the more lenient direction, and
3693 legacy behaviour is preserved as much as possible.
3697 * Fix no-stdio build.
3698 *David Woodhouse <David.Woodhouse@intel.com> and also*
3699 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3701 * New testing framework
3702 The testing framework has been largely rewritten and is now using
3703 perl and the perl modules Test::Harness and an extended variant of
3704 Test::More called OpenSSL::Test to do its work. All test scripts in
3705 test/ have been rewritten into test recipes, and all direct calls to
3706 executables in test/Makefile have become individual recipes using the
3707 simplified testing OpenSSL::Test::Simple.
3709 For documentation on our testing modules, do:
3711 perldoc test/testlib/OpenSSL/Test/Simple.pm
3712 perldoc test/testlib/OpenSSL/Test.pm
3716 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3717 are used; the latter aborts on memory leaks (usually checked on exit).
3718 Some undocumented "set malloc, etc., hooks" functions were removed
3719 and others were changed. All are now documented.
3723 * In DSA_generate_parameters_ex, if the provided seed is too short,
3726 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3728 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3729 from RFC4279, RFC4785, RFC5487, RFC5489.
3731 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3732 original RSA_PSK patch.
3736 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3737 era flag was never set throughout the codebase (only read). Also removed
3738 SSL3_FLAGS_POP_BUFFER which was only used if
3739 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3743 * Changed the default name options in the "ca", "crl", "req" and "x509"
3744 to be "oneline" instead of "compat".
3748 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3749 not aware of clients that still exhibit this bug, and the workaround
3750 hasn't been working properly for a while.
3754 * The return type of BIO_number_read() and BIO_number_written() as well as
3755 the corresponding num_read and num_write members in the BIO structure has
3756 changed from unsigned long to uint64_t. On platforms where an unsigned
3757 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3762 * Given the pervasive nature of TLS extensions it is inadvisable to run
3763 OpenSSL without support for them. It also means that maintaining
3764 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3765 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3769 * Removed support for the two export grade static DH ciphersuites
3770 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3771 were newly added (along with a number of other static DH ciphersuites) to
3772 1.0.2. However the two export ones have *never* worked since they were
3773 introduced. It seems strange in any case to be adding new export
3774 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3778 * Version negotiation has been rewritten. In particular SSLv23_method(),
3779 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3780 and turned into macros which simply call the new preferred function names
3781 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3782 should use the new names instead. Also as part of this change the ssl23.h
3783 header file has been removed.
3787 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3788 code and the associated standard is no longer considered fit-for-purpose.
3792 * RT2547 was closed. When generating a private key, try to make the
3793 output file readable only by the owner. This behavior change might
3794 be noticeable when interacting with other software.
3796 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3801 * Added HTTP GET support to the ocsp command.
3805 * Changed default digest for the dgst and enc commands from MD5 to
3810 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3814 * Added support for TLS extended master secret from
3815 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3816 initial patch which was a great help during development.
3820 * All libssl internal structures have been removed from the public header
3821 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3822 now redundant). Users should not attempt to access internal structures
3823 directly. Instead they should use the provided API functions.
3827 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3828 Access to deprecated functions can be re-enabled by running config with
3829 "enable-deprecated". In addition applications wishing to use deprecated
3830 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3831 will, by default, disable some transitive includes that previously existed
3832 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3836 * Added support for OCB mode. OpenSSL has been granted a patent license
3837 compatible with the OpenSSL license for use of OCB. Details are available
3838 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3839 for OCB can be removed by calling config with no-ocb.
3843 * SSLv2 support has been removed. It still supports receiving a SSLv2
3844 compatible client hello.
3848 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3849 done while fixing the error code for the key-too-small case.
3851 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3853 * CA.sh has been removed; use CA.pl instead.
3857 * Removed old DES API.
3861 * Remove various unsupported platforms:
3867 Sinix/ReliantUNIX RM400
3872 16-bit platforms such as WIN16
3876 * Clean up OPENSSL_NO_xxx #define's
3877 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3878 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3879 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3880 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3881 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3882 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3883 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3884 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3885 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3886 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3890 * Cleaned up dead code
3891 Remove all but one '#ifdef undef' which is to be looked at.
3895 * Clean up calling of xxx_free routines.
3896 Just like free(), fix most of the xxx_free routines to accept
3897 NULL. Remove the non-null checks from callers. Save much code.
3901 * Add secure heap for storage of private keys (when possible).
3902 Add BIO_s_secmem(), CBIGNUM, etc.
3903 Contributed by Akamai Technologies under our Corporate CLA.
3907 * Experimental support for a new, fast, unbiased prime candidate generator,
3908 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3910 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3912 * New output format NSS in the sess_id command line tool. This allows
3913 exporting the session id and the master key in NSS keylog format.
3915 *Martin Kaiser <martin@kaiser.cx>*
3917 * Harmonize version and its documentation. -f flag is used to display
3920 *mancha <mancha1@zoho.com>*
3922 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3923 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3925 *mancha <mancha1@zoho.com>*
3927 * Fix some double frees. These are not thought to be exploitable.
3929 *mancha <mancha1@zoho.com>*
3931 * A missing bounds check in the handling of the TLS heartbeat extension
3932 can be used to reveal up to 64k of memory to a connected client or
3935 Thanks for Neel Mehta of Google Security for discovering this bug and to
3936 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3937 preparing the fix ([CVE-2014-0160])
3939 *Adam Langley, Bodo Moeller*
3941 * Fix for the attack described in the paper "Recovering OpenSSL
3942 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3943 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3944 <http://eprint.iacr.org/2014/140>
3946 Thanks to Yuval Yarom and Naomi Benger for discovering this
3947 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3949 *Yuval Yarom and Naomi Benger*
3951 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3952 this fixes a limitation in previous versions of OpenSSL.
3956 * Experimental encrypt-then-mac support.
3958 Experimental support for encrypt then mac from
3959 draft-gutmann-tls-encrypt-then-mac-02.txt
3961 To enable it set the appropriate extension number (0x42 for the test
3962 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3964 For non-compliant peers (i.e. just about everything) this should have no
3967 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3971 * Add EVP support for key wrapping algorithms, to avoid problems with
3972 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3973 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3974 algorithms and include tests cases.
3978 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3983 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3984 MGF1 digest and OAEP label.
3988 * Make openssl verify return errors.
3990 *Chris Palmer <palmer@google.com> and Ben Laurie*
3992 * New function ASN1_TIME_diff to calculate the difference between two
3993 ASN1_TIME structures or one structure and the current time.
3997 * Update fips_test_suite to support multiple command line options. New
3998 test to induce all self test errors in sequence and check expected
4003 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4004 sign or verify all in one operation.
4008 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4009 test programs and fips_test_suite. Includes functionality to parse
4010 the minimal script output of fipsalgest.pl directly.
4014 * Add authorisation parameter to FIPS_module_mode_set().
4018 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4022 * Use separate DRBG fields for internal and external flags. New function
4023 FIPS_drbg_health_check() to perform on demand health checking. Add
4024 generation tests to fips_test_suite with reduced health check interval to
4025 demonstrate periodic health checking. Add "nodh" option to
4026 fips_test_suite to skip very slow DH test.
4030 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4035 * More extensive health check for DRBG checking many more failure modes.
4036 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4037 combination: call this in fips_test_suite.
4041 * Add support for canonical generation of DSA parameter 'g'. See
4044 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4045 POST to handle HMAC cases.
4049 * Add functions FIPS_module_version() and FIPS_module_version_text()
4050 to return numerical and string versions of the FIPS module number.
4054 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4055 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4056 outside the validated module in the FIPS capable OpenSSL.
4060 * Minor change to DRBG entropy callback semantics. In some cases
4061 there is no multiple of the block length between min_len and
4062 max_len. Allow the callback to return more than max_len bytes
4063 of entropy but discard any extra: it is the callback's responsibility
4064 to ensure that the extra data discarded does not impact the
4065 requested amount of entropy.
4069 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4070 information in FIPS186-3, SP800-57 and SP800-131A.
4074 * CCM support via EVP. Interface is very similar to GCM case except we
4075 must supply all data in one chunk (i.e. no update, final) and the
4076 message length must be supplied if AAD is used. Add algorithm test
4081 * Initial version of POST overhaul. Add POST callback to allow the status
4082 of POST to be monitored and/or failures induced. Modify fips_test_suite
4083 to use callback. Always run all selftests even if one fails.
4087 * XTS support including algorithm test driver in the fips_gcmtest program.
4088 Note: this does increase the maximum key length from 32 to 64 bytes but
4089 there should be no binary compatibility issues as existing applications
4090 will never use XTS mode.
4094 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4095 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4096 performs algorithm blocking for unapproved PRNG types. Also do not
4097 set PRNG type in FIPS_mode_set(): leave this to the application.
4098 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4099 the standard OpenSSL PRNG: set additional data to a date time vector.
4103 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4104 This shouldn't present any incompatibility problems because applications
4105 shouldn't be using these directly and any that are will need to rethink
4106 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4110 * Extensive self tests and health checking required by SP800-90 DRBG.
4111 Remove strength parameter from FIPS_drbg_instantiate and always
4112 instantiate at maximum supported strength.
4116 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4120 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4124 * New function DH_compute_key_padded() to compute a DH key and pad with
4125 leading zeroes if needed: this complies with SP800-56A et al.
4129 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4130 anything, incomplete, subject to change and largely untested at present.
4134 * Modify fipscanisteronly build option to only build the necessary object
4135 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4139 * Add experimental option FIPSSYMS to give all symbols in
4140 fipscanister.o and FIPS or fips prefix. This will avoid
4141 conflicts with future versions of OpenSSL. Add perl script
4142 util/fipsas.pl to preprocess assembly language source files
4143 and rename any affected symbols.
4147 * Add selftest checks and algorithm block of non-fips algorithms in
4148 FIPS mode. Remove DES2 from selftests.
4152 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4153 return internal method without any ENGINE dependencies. Add new
4154 tiny fips sign and verify functions.
4158 * New build option no-ec2m to disable characteristic 2 code.
4162 * New build option "fipscanisteronly". This only builds fipscanister.o
4163 and (currently) associated fips utilities. Uses the file Makefile.fips
4164 instead of Makefile.org as the prototype.
4168 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4169 Update fips_gcmtest to use IV generator.
4173 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4174 setting output buffer to NULL. The `*Final` function must be
4175 called although it will not retrieve any additional data. The tag
4176 can be set or retrieved with a ctrl. The IV length is by default 12
4177 bytes (96 bits) but can be set to an alternative value. If the IV
4178 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4183 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4184 underlying do_cipher function handles all cipher semantics itself
4185 including padding and finalisation. This is useful if (for example)
4186 an ENGINE cipher handles block padding itself. The behaviour of
4187 do_cipher is subtly changed if this flag is set: the return value
4188 is the number of characters written to the output buffer (zero is
4189 no longer an error code) or a negative error code. Also if the
4190 input buffer is NULL and length 0 finalisation should be performed.
4194 * If a candidate issuer certificate is already part of the constructed
4195 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4199 * Improve forward-security support: add functions
4201 void SSL_CTX_set_not_resumable_session_callback(
4202 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4203 void SSL_set_not_resumable_session_callback(
4204 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4206 for use by SSL/TLS servers; the callback function will be called whenever a
4207 new session is created, and gets to decide whether the session may be
4208 cached to make it resumable (return 0) or not (return 1). (As by the
4209 SSL/TLS protocol specifications, the session_id sent by the server will be
4210 empty to indicate that the session is not resumable; also, the server will
4211 not generate RFC 4507 (RFC 5077) session tickets.)
4213 A simple reasonable callback implementation is to return is_forward_secure.
4214 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4215 by the SSL/TLS server library, indicating whether it can provide forward
4218 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4220 * New -verify_name option in command line utilities to set verification
4225 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4226 Add CMAC pkey methods.
4230 * Experimental renegotiation in s_server -www mode. If the client
4231 browses /reneg connection is renegotiated. If /renegcert it is
4232 renegotiated requesting a certificate.
4236 * Add an "external" session cache for debugging purposes to s_server. This
4237 should help trace issues which normally are only apparent in deployed
4238 multi-process servers.
4242 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4243 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4244 BIO_set_cipher() and some obscure PEM functions were changed so they
4245 can now return an error. The RAND changes required a change to the
4246 RAND_METHOD structure.
4250 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4251 a gcc attribute to warn if the result of a function is ignored. This
4252 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4253 whose return value is often ignored.
4257 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4258 These allow SCTs (signed certificate timestamps) to be requested and
4259 validated when establishing a connection.
4261 *Rob Percival <robpercival@google.com>*
4266 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4268 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4269 used even when parsing explicit parameters, when loading a encoded key
4270 or calling `EC_GROUP_new_from_ecpkparameters()`/
4271 `EC_GROUP_new_from_ecparameters()`.
4272 This prevents bypass of security hardening and performance gains,
4273 especially for curves with specialized EC_METHODs.
4274 By default, if a key encoded with explicit parameters is loaded and later
4275 encoded, the output is still encoded with explicit parameters, even if
4276 internally a "named" EC_GROUP is used for computation.
4280 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4281 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4282 NULL. After this change, only the cofactor parameter can be NULL. It also
4283 does some minimal sanity checks on the passed order.
4288 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4289 An attack is simple, if the first CMS_recipientInfo is valid but the
4290 second CMS_recipientInfo is chosen ciphertext. If the second
4291 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4292 encryption key will be replaced by garbage, and the message cannot be
4293 decoded, but if the RSA decryption fails, the correct encryption key is
4294 used and the recipient will not notice the attack.
4295 As a work around for this potential attack the length of the decrypted
4296 key must be equal to the cipher default key length, in case the
4297 certifiate is not given and all recipientInfo are tried out.
4298 The old behaviour can be re-enabled in the CMS code by setting the
4299 CMS_DEBUG_DECRYPT flag.
4304 * Document issue with installation paths in diverse Windows builds
4306 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4307 binaries and run-time config file.
4312 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4314 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4315 This changes the size when using the `genpkey` command when no size is given.
4316 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4317 generation commands to use 2048 bits by default.
4321 * Add FIPS support for Android Arm 64-bit
4323 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4324 Module in Version 2.0.10. For some reason, the corresponding target
4325 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4326 built with FIPS support on Android Arm 64-bit. This omission has been
4329 *Matthias St. Pierre*
4331 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4333 * 0-byte record padding oracle
4335 If an application encounters a fatal protocol error and then calls
4336 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4337 then OpenSSL can respond differently to the calling application if a 0 byte
4338 record is received with invalid padding compared to if a 0 byte record is
4339 received with an invalid MAC. If the application then behaves differently
4340 based on that in a way that is detectable to the remote peer, then this
4341 amounts to a padding oracle that could be used to decrypt data.
4343 In order for this to be exploitable "non-stitched" ciphersuites must be in
4344 use. Stitched ciphersuites are optimised implementations of certain
4345 commonly used ciphersuites. Also the application must call SSL_shutdown()
4346 twice even if a protocol error has occurred (applications should not do
4347 this but some do anyway).
4349 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4350 Aviram, with additional investigation by Steven Collison and Andrew
4351 Hourselt. It was reported to OpenSSL on 10th December 2018.
4356 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4360 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4362 * Microarchitecture timing vulnerability in ECC scalar multiplication
4364 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4365 shown to be vulnerable to a microarchitecture timing side channel attack.
4366 An attacker with sufficient access to mount local timing attacks during
4367 ECDSA signature generation could recover the private key.
4369 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4370 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4376 * Timing vulnerability in DSA signature generation
4378 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4379 timing side channel attack. An attacker could use variations in the signing
4380 algorithm to recover the private key.
4382 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4387 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4388 Module, accidentally introduced while backporting security fixes from the
4389 development branch and hindering the use of ECC in FIPS mode.
4393 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4395 * Client DoS due to large DH parameter
4397 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4398 malicious server can send a very large prime value to the client. This will
4399 cause the client to spend an unreasonably long period of time generating a
4400 key for this prime resulting in a hang until the client has finished. This
4401 could be exploited in a Denial Of Service attack.
4403 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4408 * Cache timing vulnerability in RSA Key Generation
4410 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4411 a cache timing side channel attack. An attacker with sufficient access to
4412 mount cache timing attacks during the RSA key generation process could
4413 recover the private key.
4415 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4416 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4421 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4422 parameter is no longer accepted, as it leads to a corrupt table. NULL
4423 pem_str is reserved for alias entries only.
4427 * Revert blinding in ECDSA sign and instead make problematic addition
4428 length-invariant. Switch even to fixed-length Montgomery multiplication.
4432 * Change generating and checking of primes so that the error rate of not
4433 being prime depends on the intended use based on the size of the input.
4434 For larger primes this will result in more rounds of Miller-Rabin.
4435 The maximal error rate for primes with more than 1080 bits is lowered
4438 *Kurt Roeckx, Annie Yousar*
4440 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4444 * Add blinding to ECDSA and DSA signatures to protect against side channel
4445 attacks discovered by Keegan Ryan (NCC Group).
4449 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4450 now allow empty (zero character) pass phrases.
4454 * Certificate time validation (X509_cmp_time) enforces stricter
4455 compliance with RFC 5280. Fractional seconds and timezone offsets
4456 are no longer allowed.
4460 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4462 * Constructed ASN.1 types with a recursive definition could exceed the stack
4464 Constructed ASN.1 types with a recursive definition (such as can be found
4465 in PKCS7) could eventually exceed the stack given malicious input with
4466 excessive recursion. This could result in a Denial Of Service attack. There
4467 are no such structures used within SSL/TLS that come from untrusted sources
4468 so this is considered safe.
4470 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4476 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4478 * Read/write after SSL object in error state
4480 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4481 mechanism. The intent was that if a fatal error occurred during a handshake
4482 then OpenSSL would move into the error state and would immediately fail if
4483 you attempted to continue the handshake. This works as designed for the
4484 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4485 SSL_connect()), however due to a bug it does not work correctly if
4486 SSL_read() or SSL_write() is called directly. In that scenario, if the
4487 handshake fails then a fatal error will be returned in the initial function
4488 call. If SSL_read()/SSL_write() is subsequently called by the application
4489 for the same SSL object then it will succeed and the data is passed without
4490 being decrypted/encrypted directly from the SSL/TLS record layer.
4492 In order to exploit this issue an application bug would have to be present
4493 that resulted in a call to SSL_read()/SSL_write() being issued after having
4494 already received a fatal error.
4496 This issue was reported to OpenSSL by David Benjamin (Google).
4501 * rsaz_1024_mul_avx2 overflow bug on x86_64
4503 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4504 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4505 Analysis suggests that attacks against RSA and DSA as a result of this
4506 defect would be very difficult to perform and are not believed likely.
4507 Attacks against DH1024 are considered just feasible, because most of the
4508 work necessary to deduce information about a private key may be performed
4509 offline. The amount of resources required for such an attack would be
4510 significant. However, for an attack on TLS to be meaningful, the server
4511 would have to share the DH1024 private key among multiple clients, which is
4512 no longer an option since CVE-2016-0701.
4514 This only affects processors that support the AVX2 but not ADX extensions
4515 like Intel Haswell (4th generation).
4517 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4518 was originally found via the OSS-Fuzz project.
4523 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4525 * bn_sqrx8x_internal carry bug on x86_64
4527 There is a carry propagating bug in the x86_64 Montgomery squaring
4528 procedure. No EC algorithms are affected. Analysis suggests that attacks
4529 against RSA and DSA as a result of this defect would be very difficult to
4530 perform and are not believed likely. Attacks against DH are considered just
4531 feasible (although very difficult) because most of the work necessary to
4532 deduce information about a private key may be performed offline. The amount
4533 of resources required for such an attack would be very significant and
4534 likely only accessible to a limited number of attackers. An attacker would
4535 additionally need online access to an unpatched system using the target
4536 private key in a scenario with persistent DH parameters and a private
4537 key that is shared between multiple clients.
4539 This only affects processors that support the BMI1, BMI2 and ADX extensions
4540 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4542 This issue was reported to OpenSSL by the OSS-Fuzz project.
4547 * Malformed X.509 IPAddressFamily could cause OOB read
4549 If an X.509 certificate has a malformed IPAddressFamily extension,
4550 OpenSSL could do a one-byte buffer overread. The most likely result
4551 would be an erroneous display of the certificate in text format.
4553 This issue was reported to OpenSSL by the OSS-Fuzz project.
4558 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4560 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4561 platform rather than 'mingw'.
4565 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4567 * Truncated packet could crash via OOB read
4569 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4570 cipher is being used, then a truncated packet can cause that host to
4571 perform an out-of-bounds read, usually resulting in a crash.
4573 This issue was reported to OpenSSL by Robert Święcki of Google.
4578 * BN_mod_exp may produce incorrect results on x86_64
4580 There is a carry propagating bug in the x86_64 Montgomery squaring
4581 procedure. No EC algorithms are affected. Analysis suggests that attacks
4582 against RSA and DSA as a result of this defect would be very difficult to
4583 perform and are not believed likely. Attacks against DH are considered just
4584 feasible (although very difficult) because most of the work necessary to
4585 deduce information about a private key may be performed offline. The amount
4586 of resources required for such an attack would be very significant and
4587 likely only accessible to a limited number of attackers. An attacker would
4588 additionally need online access to an unpatched system using the target
4589 private key in a scenario with persistent DH parameters and a private
4590 key that is shared between multiple clients. For example this can occur by
4591 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4592 similar to CVE-2015-3193 but must be treated as a separate problem.
4594 This issue was reported to OpenSSL by the OSS-Fuzz project.
4599 * Montgomery multiplication may produce incorrect results
4601 There is a carry propagating bug in the Broadwell-specific Montgomery
4602 multiplication procedure that handles input lengths divisible by, but
4603 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4604 and DH private keys are impossible. This is because the subroutine in
4605 question is not used in operations with the private key itself and an input
4606 of the attacker's direct choice. Otherwise the bug can manifest itself as
4607 transient authentication and key negotiation failures or reproducible
4608 erroneous outcome of public-key operations with specially crafted input.
4609 Among EC algorithms only Brainpool P-512 curves are affected and one
4610 presumably can attack ECDH key negotiation. Impact was not analyzed in
4611 detail, because pre-requisites for attack are considered unlikely. Namely
4612 multiple clients have to choose the curve in question and the server has to
4613 share the private key among them, neither of which is default behaviour.
4614 Even then only clients that chose the curve will be affected.
4616 This issue was publicly reported as transient failures and was not
4617 initially recognized as a security issue. Thanks to Richard Morgan for
4618 providing reproducible case.
4623 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4624 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4625 prevent issues where no progress is being made and the peer continually
4626 sends unrecognised record types, using up resources processing them.
4630 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4632 * Missing CRL sanity check
4634 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4635 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4636 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4638 This issue only affects the OpenSSL 1.0.2i
4643 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4645 * OCSP Status Request extension unbounded memory growth
4647 A malicious client can send an excessively large OCSP Status Request
4648 extension. If that client continually requests renegotiation, sending a
4649 large OCSP Status Request extension each time, then there will be unbounded
4650 memory growth on the server. This will eventually lead to a Denial Of
4651 Service attack through memory exhaustion. Servers with a default
4652 configuration are vulnerable even if they do not support OCSP. Builds using
4653 the "no-ocsp" build time option are not affected.
4655 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4660 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4663 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4669 * OOB write in MDC2_Update()
4671 An overflow can occur in MDC2_Update() either if called directly or
4672 through the EVP_DigestUpdate() function using MDC2. If an attacker
4673 is able to supply very large amounts of input data after a previous
4674 call to EVP_EncryptUpdate() with a partial block then a length check
4675 can overflow resulting in a heap corruption.
4677 The amount of data needed is comparable to SIZE_MAX which is impractical
4680 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4685 * Malformed SHA512 ticket DoS
4687 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4688 DoS attack where a malformed ticket will result in an OOB read which will
4691 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4692 a custom server callback and ticket lookup mechanism.
4694 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4699 * OOB write in BN_bn2dec()
4701 The function BN_bn2dec() does not check the return value of BN_div_word().
4702 This can cause an OOB write if an application uses this function with an
4703 overly large BIGNUM. This could be a problem if an overly large certificate
4704 or CRL is printed out from an untrusted source. TLS is not affected because
4705 record limits will reject an oversized certificate before it is parsed.
4707 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4712 * OOB read in TS_OBJ_print_bio()
4714 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4715 the total length the OID text representation would use and not the amount
4716 of data written. This will result in OOB reads when large OIDs are
4719 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4724 * Pointer arithmetic undefined behaviour
4726 Avoid some undefined pointer arithmetic
4728 A common idiom in the codebase is to check limits in the following manner:
4731 Where "p" points to some malloc'd data of SIZE bytes and
4734 "len" here could be from some externally supplied data (e.g. from a TLS
4737 The rules of C pointer arithmetic are such that "p + len" is only well
4738 defined where len <= SIZE. Therefore the above idiom is actually
4739 undefined behaviour.
4741 For example this could cause problems if some malloc implementation
4742 provides an address for "p" such that "p + len" actually overflows for
4743 values of len that are too big and therefore p + len < limit.
4745 This issue was reported to OpenSSL by Guido Vranken
4750 * Constant time flag not preserved in DSA signing
4752 Operations in the DSA signing algorithm should run in constant time in
4753 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4754 implementation means that a non-constant time codepath is followed for
4755 certain operations. This has been demonstrated through a cache-timing
4756 attack to be sufficient for an attacker to recover the private DSA key.
4758 This issue was reported by César Pereida (Aalto University), Billy Brumley
4759 (Tampere University of Technology), and Yuval Yarom (The University of
4760 Adelaide and NICTA).
4765 * DTLS buffered message DoS
4767 In a DTLS connection where handshake messages are delivered out-of-order
4768 those messages that OpenSSL is not yet ready to process will be buffered
4769 for later use. Under certain circumstances, a flaw in the logic means that
4770 those messages do not get removed from the buffer even though the handshake
4771 has been completed. An attacker could force up to approx. 15 messages to
4772 remain in the buffer when they are no longer required. These messages will
4773 be cleared when the DTLS connection is closed. The default maximum size for
4774 a message is 100k. Therefore the attacker could force an additional 1500k
4775 to be consumed per connection. By opening many simulataneous connections an
4776 attacker could cause a DoS attack through memory exhaustion.
4778 This issue was reported to OpenSSL by Quan Luo.
4783 * DTLS replay protection DoS
4785 A flaw in the DTLS replay attack protection mechanism means that records
4786 that arrive for future epochs update the replay protection "window" before
4787 the MAC for the record has been validated. This could be exploited by an
4788 attacker by sending a record for the next epoch (which does not have to
4789 decrypt or have a valid MAC), with a very large sequence number. This means
4790 that all subsequent legitimate packets are dropped causing a denial of
4791 service for a specific DTLS connection.
4793 This issue was reported to OpenSSL by the OCAP audit team.
4798 * Certificate message OOB reads
4800 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4801 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4802 theoretical DoS risk but this has not been observed in practice on common
4805 The messages affected are client certificate, client certificate request
4806 and server certificate. As a result the attack can only be performed
4807 against a client or a server which enables client authentication.
4809 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4814 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4816 * Prevent padding oracle in AES-NI CBC MAC check
4818 A MITM attacker can use a padding oracle attack to decrypt traffic
4819 when the connection uses an AES CBC cipher and the server support
4822 This issue was introduced as part of the fix for Lucky 13 padding
4823 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4824 constant time by making sure that always the same bytes are read and
4825 compared against either the MAC or padding bytes. But it no longer
4826 checked that there was enough data to have both the MAC and padding
4829 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4834 * Fix EVP_EncodeUpdate overflow
4836 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4837 Base64 encoding of binary data. If an attacker is able to supply very large
4838 amounts of input data then a length check can overflow resulting in a heap
4841 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4842 the `PEM_write_bio*` family of functions. These are mainly used within the
4843 OpenSSL command line applications, so any application which processes data
4844 from an untrusted source and outputs it as a PEM file should be considered
4845 vulnerable to this issue. User applications that call these APIs directly
4846 with large amounts of untrusted data may also be vulnerable.
4848 This issue was reported by Guido Vranken.
4853 * Fix EVP_EncryptUpdate overflow
4855 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4856 is able to supply very large amounts of input data after a previous call to
4857 EVP_EncryptUpdate() with a partial block then a length check can overflow
4858 resulting in a heap corruption. Following an analysis of all OpenSSL
4859 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4860 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4861 the first called function after an EVP_EncryptInit(), and therefore that
4862 specific call must be safe. The second form is where the length passed to
4863 EVP_EncryptUpdate() can be seen from the code to be some small value and
4864 therefore there is no possibility of an overflow. Since all instances are
4865 one of these two forms, it is believed that there can be no overflows in
4866 internal code due to this problem. It should be noted that
4867 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4868 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4869 of these calls have also been analysed too and it is believed there are no
4870 instances in internal usage where an overflow could occur.
4872 This issue was reported by Guido Vranken.
4877 * Prevent ASN.1 BIO excessive memory allocation
4879 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4880 a short invalid encoding can cause allocation of large amounts of memory
4881 potentially consuming excessive resources or exhausting memory.
4883 Any application parsing untrusted data through d2i BIO functions is
4884 affected. The memory based functions such as d2i_X509() are *not* affected.
4885 Since the memory based functions are used by the TLS library, TLS
4886 applications are not affected.
4888 This issue was reported by Brian Carpenter.
4895 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4896 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4897 in arbitrary stack data being returned in the buffer.
4899 This issue was reported by Guido Vranken.
4904 * Modify behavior of ALPN to invoke callback after SNI/servername
4905 callback, such that updates to the SSL_CTX affect ALPN.
4909 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4914 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4915 methods are enabled and ssl2 is disabled the methods return NULL.
4919 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4921 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4922 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4923 provide any "EXPORT" or "LOW" strength ciphers.
4927 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4928 is by default disabled at build-time. Builds that are not configured with
4929 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4930 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4931 will need to explicitly call either of:
4933 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4935 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4937 as appropriate. Even if either of those is used, or the application
4938 explicitly uses the version-specific SSLv2_method() or its client and
4939 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4940 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4941 ciphers, and SSLv2 56-bit DES are no longer available.
4946 * Fix a double-free in DSA code
4948 A double free bug was discovered when OpenSSL parses malformed DSA private
4949 keys and could lead to a DoS attack or memory corruption for applications
4950 that receive DSA private keys from untrusted sources. This scenario is
4953 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4959 * Disable SRP fake user seed to address a server memory leak.
4961 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4963 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4964 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4965 was changed to ignore the "fake user" SRP seed, even if the seed
4968 Users should use SRP_VBASE_get1_by_user instead. Note that in
4969 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4970 also that even though configuring the SRP seed attempts to hide
4971 invalid usernames by continuing the handshake with fake
4972 credentials, this behaviour is not constant time and no strong
4973 guarantees are made that the handshake is indistinguishable from
4974 that of a valid user.
4979 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4981 In the BN_hex2bn function the number of hex digits is calculated using an
4982 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4983 large values of `i` this can result in `bn_expand` not allocating any
4984 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4985 field as NULL leading to a subsequent NULL ptr deref. For very large values
4986 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4987 In this case memory is allocated to the internal BIGNUM data field, but it
4988 is insufficiently sized leading to heap corruption. A similar issue exists
4989 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4990 is ever called by user applications with very large untrusted hex/dec data.
4991 This is anticipated to be a rare occurrence.
4993 All OpenSSL internal usage of these functions use data that is not expected
4994 to be untrusted, e.g. config file data or application command line
4995 arguments. If user developed applications generate config file data based
4996 on untrusted data then it is possible that this could also lead to security
4997 consequences. This is also anticipated to be rare.
4999 This issue was reported to OpenSSL by Guido Vranken.
5004 * Fix memory issues in `BIO_*printf` functions
5006 The internal `fmtstr` function used in processing a "%s" format string in
5007 the `BIO_*printf` functions could overflow while calculating the length of a
5008 string and cause an OOB read when printing very long strings.
5010 Additionally the internal `doapr_outch` function can attempt to write to an
5011 OOB memory location (at an offset from the NULL pointer) in the event of a
5012 memory allocation failure. In 1.0.2 and below this could be caused where
5013 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5014 could be in processing a very long "%s" format string. Memory leaks can
5017 The first issue may mask the second issue dependent on compiler behaviour.
5018 These problems could enable attacks where large amounts of untrusted data
5019 is passed to the `BIO_*printf` functions. If applications use these functions
5020 in this way then they could be vulnerable. OpenSSL itself uses these
5021 functions when printing out human-readable dumps of ASN.1 data. Therefore
5022 applications that print this data could be vulnerable if the data is from
5023 untrusted sources. OpenSSL command line applications could also be
5024 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5025 as command line arguments.
5027 Libssl is not considered directly vulnerable. Additionally certificates etc
5028 received via remote connections via libssl are also unlikely to be able to
5029 trigger these issues because of message size limits enforced within libssl.
5031 This issue was reported to OpenSSL Guido Vranken.
5036 * Side channel attack on modular exponentiation
5038 A side-channel attack was found which makes use of cache-bank conflicts on
5039 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5040 of RSA keys. The ability to exploit this issue is limited as it relies on
5041 an attacker who has control of code in a thread running on the same
5042 hyper-threaded core as the victim thread which is performing decryptions.
5044 This issue was reported to OpenSSL by Yuval Yarom, The University of
5045 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5046 Nadia Heninger, University of Pennsylvania with more information at
5047 <http://cachebleed.info>.
5052 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5053 if no keysize is specified with default_bits. This fixes an
5054 omission in an earlier change that changed all RSA/DSA key generation
5055 commands to use 2048 bits by default.
5059 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5061 * DH small subgroups
5063 Historically OpenSSL only ever generated DH parameters based on "safe"
5064 primes. More recently (in version 1.0.2) support was provided for
5065 generating X9.42 style parameter files such as those required for RFC 5114
5066 support. The primes used in such files may not be "safe". Where an
5067 application is using DH configured with parameters based on primes that are
5068 not "safe" then an attacker could use this fact to find a peer's private
5069 DH exponent. This attack requires that the attacker complete multiple
5070 handshakes in which the peer uses the same private DH exponent. For example
5071 this could be used to discover a TLS server's private DH exponent if it's
5072 reusing the private DH exponent or it's using a static DH ciphersuite.
5074 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5075 TLS. It is not on by default. If the option is not set then the server
5076 reuses the same private DH exponent for the life of the server process and
5077 would be vulnerable to this attack. It is believed that many popular
5078 applications do set this option and would therefore not be at risk.
5080 The fix for this issue adds an additional check where a "q" parameter is
5081 available (as is the case in X9.42 based parameters). This detects the
5082 only known attack, and is the only possible defense for static DH
5083 ciphersuites. This could have some performance impact.
5085 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5086 default and cannot be disabled. This could have some performance impact.
5088 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5093 * SSLv2 doesn't block disabled ciphers
5095 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5096 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5097 been disabled, provided that the SSLv2 protocol was not also disabled via
5100 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5101 and Sebastian Schinzel.
5106 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5108 * BN_mod_exp may produce incorrect results on x86_64
5110 There is a carry propagating bug in the x86_64 Montgomery squaring
5111 procedure. No EC algorithms are affected. Analysis suggests that attacks
5112 against RSA and DSA as a result of this defect would be very difficult to
5113 perform and are not believed likely. Attacks against DH are considered just
5114 feasible (although very difficult) because most of the work necessary to
5115 deduce information about a private key may be performed offline. The amount
5116 of resources required for such an attack would be very significant and
5117 likely only accessible to a limited number of attackers. An attacker would
5118 additionally need online access to an unpatched system using the target
5119 private key in a scenario with persistent DH parameters and a private
5120 key that is shared between multiple clients. For example this can occur by
5121 default in OpenSSL DHE based SSL/TLS ciphersuites.
5123 This issue was reported to OpenSSL by Hanno Böck.
5128 * Certificate verify crash with missing PSS parameter
5130 The signature verification routines will crash with a NULL pointer
5131 dereference if presented with an ASN.1 signature using the RSA PSS
5132 algorithm and absent mask generation function parameter. Since these
5133 routines are used to verify certificate signature algorithms this can be
5134 used to crash any certificate verification operation and exploited in a
5135 DoS attack. Any application which performs certificate verification is
5136 vulnerable including OpenSSL clients and servers which enable client
5139 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5144 * X509_ATTRIBUTE memory leak
5146 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5147 memory. This structure is used by the PKCS#7 and CMS routines so any
5148 application which reads PKCS#7 or CMS data from untrusted sources is
5149 affected. SSL/TLS is not affected.
5151 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5157 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5158 This changes the decoding behaviour for some invalid messages,
5159 though the change is mostly in the more lenient direction, and
5160 legacy behaviour is preserved as much as possible.
5164 * In DSA_generate_parameters_ex, if the provided seed is too short,
5167 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5169 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5171 * Alternate chains certificate forgery
5173 During certificate verification, OpenSSL will attempt to find an
5174 alternative certificate chain if the first attempt to build such a chain
5175 fails. An error in the implementation of this logic can mean that an
5176 attacker could cause certain checks on untrusted certificates to be
5177 bypassed, such as the CA flag, enabling them to use a valid leaf
5178 certificate to act as a CA and "issue" an invalid certificate.
5180 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5185 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5187 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5188 incompatibility in the handling of HMAC. The previous ABI has now been
5193 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5195 * Malformed ECParameters causes infinite loop
5197 When processing an ECParameters structure OpenSSL enters an infinite loop
5198 if the curve specified is over a specially malformed binary polynomial
5201 This can be used to perform denial of service against any
5202 system which processes public keys, certificate requests or
5203 certificates. This includes TLS clients and TLS servers with
5204 client authentication enabled.
5206 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5211 * Exploitable out-of-bounds read in X509_cmp_time
5213 X509_cmp_time does not properly check the length of the ASN1_TIME
5214 string and can read a few bytes out of bounds. In addition,
5215 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5218 An attacker can use this to craft malformed certificates and CRLs of
5219 various sizes and potentially cause a segmentation fault, resulting in
5220 a DoS on applications that verify certificates or CRLs. TLS clients
5221 that verify CRLs are affected. TLS clients and servers with client
5222 authentication enabled may be affected if they use custom verification
5225 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5226 independently by Hanno Böck.
5231 * PKCS7 crash with missing EnvelopedContent
5233 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5234 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5235 with missing content and trigger a NULL pointer dereference on parsing.
5237 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5238 structures from untrusted sources are affected. OpenSSL clients and
5239 servers are not affected.
5241 This issue was reported to OpenSSL by Michal Zalewski (Google).
5246 * CMS verify infinite loop with unknown hash function
5248 When verifying a signedData message the CMS code can enter an infinite loop
5249 if presented with an unknown hash function OID. This can be used to perform
5250 denial of service against any system which verifies signedData messages using
5252 This issue was reported to OpenSSL by Johannes Bauer.
5257 * Race condition handling NewSessionTicket
5259 If a NewSessionTicket is received by a multi-threaded client when attempting to
5260 reuse a previous ticket then a race condition can occur potentially leading to
5261 a double free of the ticket data.
5266 * Only support 256-bit or stronger elliptic curves with the
5267 'ecdh_auto' setting (server) or by default (client). Of supported
5268 curves, prefer P-256 (both).
5272 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5274 * ClientHello sigalgs DoS fix
5276 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5277 invalid signature algorithms extension a NULL pointer dereference will
5278 occur. This can be exploited in a DoS attack against the server.
5280 This issue was was reported to OpenSSL by David Ramos of Stanford
5284 *Stephen Henson and Matt Caswell*
5286 * Multiblock corrupted pointer fix
5288 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5289 feature only applies on 64 bit x86 architecture platforms that support AES
5290 NI instructions. A defect in the implementation of "multiblock" can cause
5291 OpenSSL's internal write buffer to become incorrectly set to NULL when
5292 using non-blocking IO. Typically, when the user application is using a
5293 socket BIO for writing, this will only result in a failed connection.
5294 However if some other BIO is used then it is likely that a segmentation
5295 fault will be triggered, thus enabling a potential DoS attack.
5297 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5302 * Segmentation fault in DTLSv1_listen fix
5304 The DTLSv1_listen function is intended to be stateless and processes the
5305 initial ClientHello from many peers. It is common for user code to loop
5306 over the call to DTLSv1_listen until a valid ClientHello is received with
5307 an associated cookie. A defect in the implementation of DTLSv1_listen means
5308 that state is preserved in the SSL object from one invocation to the next
5309 that can lead to a segmentation fault. Errors processing the initial
5310 ClientHello can trigger this scenario. An example of such an error could be
5311 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5314 This issue was reported to OpenSSL by Per Allansson.
5319 * Segmentation fault in ASN1_TYPE_cmp fix
5321 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5322 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5323 certificate signature algorithm consistency this can be used to crash any
5324 certificate verification operation and exploited in a DoS attack. Any
5325 application which performs certificate verification is vulnerable including
5326 OpenSSL clients and servers which enable client authentication.
5331 * Segmentation fault for invalid PSS parameters fix
5333 The signature verification routines will crash with a NULL pointer
5334 dereference if presented with an ASN.1 signature using the RSA PSS
5335 algorithm and invalid parameters. Since these routines are used to verify
5336 certificate signature algorithms this can be used to crash any
5337 certificate verification operation and exploited in a DoS attack. Any
5338 application which performs certificate verification is vulnerable including
5339 OpenSSL clients and servers which enable client authentication.
5341 This issue was was reported to OpenSSL by Brian Carpenter.
5346 * ASN.1 structure reuse memory corruption fix
5348 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5349 memory corruption via an invalid write. Such reuse is and has been
5350 strongly discouraged and is believed to be rare.
5352 Applications that parse structures containing CHOICE or ANY DEFINED BY
5353 components may be affected. Certificate parsing (d2i_X509 and related
5354 functions) are however not affected. OpenSSL clients and servers are
5360 * PKCS7 NULL pointer dereferences fix
5362 The PKCS#7 parsing code does not handle missing outer ContentInfo
5363 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5364 missing content and trigger a NULL pointer dereference on parsing.
5366 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5367 otherwise parse PKCS#7 structures from untrusted sources are
5368 affected. OpenSSL clients and servers are not affected.
5370 This issue was reported to OpenSSL by Michal Zalewski (Google).
5375 * DoS via reachable assert in SSLv2 servers fix
5377 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5378 servers that both support SSLv2 and enable export cipher suites by sending
5379 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5381 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5382 (OpenSSL development team).
5387 * Empty CKE with client auth and DHE fix
5389 If client auth is used then a server can seg fault in the event of a DHE
5390 ciphersuite being selected and a zero length ClientKeyExchange message
5391 being sent by the client. This could be exploited in a DoS attack.
5396 * Handshake with unseeded PRNG fix
5398 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5399 with an unseeded PRNG. The conditions are:
5400 - The client is on a platform where the PRNG has not been seeded
5401 automatically, and the user has not seeded manually
5402 - A protocol specific client method version has been used (i.e. not
5403 SSL_client_methodv23)
5404 - A ciphersuite is used that does not require additional random data from
5405 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5407 If the handshake succeeds then the client random that has been used will
5408 have been generated from a PRNG with insufficient entropy and therefore the
5409 output may be predictable.
5411 For example using the following command with an unseeded openssl will
5412 succeed on an unpatched platform:
5414 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5419 * Use After Free following d2i_ECPrivatekey error fix
5421 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5422 could cause a use after free condition. This, in turn, could cause a double
5423 free in several private key parsing functions (such as d2i_PrivateKey
5424 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5425 for applications that receive EC private keys from untrusted
5426 sources. This scenario is considered rare.
5428 This issue was discovered by the BoringSSL project and fixed in their
5434 * X509_to_X509_REQ NULL pointer deref fix
5436 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5437 the certificate key is invalid. This function is rarely used in practice.
5439 This issue was discovered by Brian Carpenter.
5444 * Removed the export ciphers from the DEFAULT ciphers
5448 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5450 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5451 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5452 So far those who have to target multiple platforms would compromise
5453 and argue that binary targeting say ARMv5 would still execute on
5454 ARMv8. "Universal" build resolves this compromise by providing
5455 near-optimal performance even on newer platforms.
5459 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5460 (other platforms pending).
5462 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5464 * Add support for the SignedCertificateTimestampList certificate and
5465 OCSP response extensions from RFC6962.
5469 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5470 for corner cases. (Certain input points at infinity could lead to
5471 bogus results, with non-infinity inputs mapped to infinity too.)
5475 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5476 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5477 common cases are optimized and there still is room for further
5478 improvements. Vector Permutation AES for Altivec is also added.
5482 * Add support for little-endian ppc64 Linux target.
5484 *Marcelo Cerri (IBM)*
5486 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5487 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5488 are optimized and there still is room for further improvements.
5489 Both 32- and 64-bit modes are supported.
5491 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5493 * Improved ARMv7 NEON support.
5497 * Support for SPARC Architecture 2011 crypto extensions, first
5498 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5499 SHA256/512, MD5, GHASH and modular exponentiation.
5501 *Andy Polyakov, David Miller*
5503 * Accelerated modular exponentiation for Intel processors, a.k.a.
5506 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5508 * Support for new and upcoming Intel processors, including AVX2,
5509 BMI and SHA ISA extensions. This includes additional "stitched"
5510 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5513 This work was sponsored by Intel Corp.
5517 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5518 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5519 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5523 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5524 this fixes a limitation in previous versions of OpenSSL.
5528 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5529 MGF1 digest and OAEP label.
5533 * Add EVP support for key wrapping algorithms, to avoid problems with
5534 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5535 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5536 algorithms and include tests cases.
5540 * Add functions to allocate and set the fields of an ECDSA_METHOD
5543 *Douglas E. Engert, Steve Henson*
5545 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5546 difference in days and seconds between two tm or ASN1_TIME structures.
5550 * Add -rev test option to s_server to just reverse order of characters
5551 received by client and send back to server. Also prints an abbreviated
5552 summary of the connection parameters.
5556 * New option -brief for s_client and s_server to print out a brief summary
5557 of connection parameters.
5561 * Add callbacks for arbitrary TLS extensions.
5563 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5565 * New option -crl_download in several openssl utilities to download CRLs
5566 from CRLDP extension in certificates.
5570 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5574 * New function X509_CRL_diff to generate a delta CRL from the difference
5575 of two full CRLs. Add support to "crl" utility.
5579 * New functions to set lookup_crls function and to retrieve
5580 X509_STORE from X509_STORE_CTX.
5584 * Print out deprecated issuer and subject unique ID fields in
5589 * Extend OCSP I/O functions so they can be used for simple general purpose
5590 HTTP as well as OCSP. New wrapper function which can be used to download
5591 CRLs using the OCSP API.
5595 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5599 * `SSL_CONF*` functions. These provide a common framework for application
5600 configuration using configuration files or command lines.
5604 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5605 message callback and prints the results. Needs compile time option
5606 "enable-ssl-trace". New options to s_client and s_server to enable
5611 * New ctrl and macro to retrieve supported points extensions.
5612 Print out extension in s_server and s_client.
5616 * New functions to retrieve certificate signature and signature
5621 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5626 * New Suite B modes for TLS code. These use and enforce the requirements
5627 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5628 only use Suite B curves. The Suite B modes can be set by using the
5629 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5633 * New chain verification flags for Suite B levels of security. Check
5634 algorithms are acceptable when flags are set in X509_verify_cert.
5638 * Make tls1_check_chain return a set of flags indicating checks passed
5639 by a certificate chain. Add additional tests to handle client
5640 certificates: checks for matching certificate type and issuer name
5645 * If an attempt is made to use a signature algorithm not in the peer
5646 preference list abort the handshake. If client has no suitable
5647 signature algorithms in response to a certificate request do not
5648 use the certificate.
5652 * If server EC tmp key is not in client preference list abort handshake.
5656 * Add support for certificate stores in CERT structure. This makes it
5657 possible to have different stores per SSL structure or one store in
5658 the parent SSL_CTX. Include distinct stores for certificate chain
5659 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5660 to build and store a certificate chain in CERT structure: returning
5661 an error if the chain cannot be built: this will allow applications
5662 to test if a chain is correctly configured.
5664 Note: if the CERT based stores are not set then the parent SSL_CTX
5665 store is used to retain compatibility with existing behaviour.
5669 * New function ssl_set_client_disabled to set a ciphersuite disabled
5670 mask based on the current session, check mask when sending client
5671 hello and checking the requested ciphersuite.
5675 * New ctrls to retrieve and set certificate types in a certificate
5676 request message. Print out received values in s_client. If certificate
5677 types is not set with custom values set sensible values based on
5678 supported signature algorithms.
5682 * Support for distinct client and server supported signature algorithms.
5686 * Add certificate callback. If set this is called whenever a certificate
5687 is required by client or server. An application can decide which
5688 certificate chain to present based on arbitrary criteria: for example
5689 supported signature algorithms. Add very simple example to s_server.
5690 This fixes many of the problems and restrictions of the existing client
5691 certificate callback: for example you can now clear an existing
5692 certificate and specify the whole chain.
5696 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5697 the certificate can be used for (if anything). Set valid_flags field
5698 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5699 to have similar checks in it.
5701 Add new "cert_flags" field to CERT structure and include a "strict mode".
5702 This enforces some TLS certificate requirements (such as only permitting
5703 certificate signature algorithms contained in the supported algorithms
5704 extension) which some implementations ignore: this option should be used
5705 with caution as it could cause interoperability issues.
5709 * Update and tidy signature algorithm extension processing. Work out
5710 shared signature algorithms based on preferences and peer algorithms
5711 and print them out in s_client and s_server. Abort handshake if no
5712 shared signature algorithms.
5716 * Add new functions to allow customised supported signature algorithms
5717 for SSL and SSL_CTX structures. Add options to s_client and s_server
5722 * New function SSL_certs_clear() to delete all references to certificates
5723 from an SSL structure. Before this once a certificate had been added
5724 it couldn't be removed.
5728 * Integrate hostname, email address and IP address checking with certificate
5729 verification. New verify options supporting checking in openssl utility.
5733 * Fixes and wildcard matching support to hostname and email checking
5734 functions. Add manual page.
5736 *Florian Weimer (Red Hat Product Security Team)*
5738 * New functions to check a hostname email or IP address against a
5739 certificate. Add options x509 utility to print results of checks against
5744 * Fix OCSP checking.
5746 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5748 * Initial experimental support for explicitly trusted non-root CAs.
5749 OpenSSL still tries to build a complete chain to a root but if an
5750 intermediate CA has a trust setting included that is used. The first
5751 setting is used: whether to trust (e.g., -addtrust option to the x509
5756 * Add -trusted_first option which attempts to find certificates in the
5757 trusted store even if an untrusted chain is also supplied.
5761 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5762 platform support for Linux and Android.
5766 * Support for linux-x32, ILP32 environment in x86_64 framework.
5770 * Experimental multi-implementation support for FIPS capable OpenSSL.
5771 When in FIPS mode the approved implementations are used as normal,
5772 when not in FIPS mode the internal unapproved versions are used instead.
5773 This means that the FIPS capable OpenSSL isn't forced to use the
5774 (often lower performance) FIPS implementations outside FIPS mode.
5778 * Transparently support X9.42 DH parameters when calling
5779 PEM_read_bio_DHparameters. This means existing applications can handle
5780 the new parameter format automatically.
5784 * Initial experimental support for X9.42 DH parameter format: mainly
5785 to support use of 'q' parameter for RFC5114 parameters.
5789 * Add DH parameters from RFC5114 including test data to dhtest.
5793 * Support for automatic EC temporary key parameter selection. If enabled
5794 the most preferred EC parameters are automatically used instead of
5795 hardcoded fixed parameters. Now a server just has to call:
5796 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5797 support ECDH and use the most appropriate parameters.
5801 * Enhance and tidy EC curve and point format TLS extension code. Use
5802 static structures instead of allocation if default values are used.
5803 New ctrls to set curves we wish to support and to retrieve shared curves.
5804 Print out shared curves in s_server. New options to s_server and s_client
5805 to set list of supported curves.
5809 * New ctrls to retrieve supported signature algorithms and
5810 supported curve values as an array of NIDs. Extend openssl utility
5811 to print out received values.
5815 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5816 between NIDs and the more common NIST names such as "P-256". Enhance
5817 ecparam utility and ECC method to recognise the NIST names for curves.
5821 * Enhance SSL/TLS certificate chain handling to support different
5822 chains for each certificate instead of one chain in the parent SSL_CTX.
5826 * Support for fixed DH ciphersuite client authentication: where both
5827 server and client use DH certificates with common parameters.
5831 * Support for fixed DH ciphersuites: those requiring DH server
5836 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5838 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5839 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5840 X509_CINF_get_signature were reverted post internal team review.
5845 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5847 * OCSP Status Request extension unbounded memory growth
5849 A malicious client can send an excessively large OCSP Status Request
5850 extension. If that client continually requests renegotiation, sending a
5851 large OCSP Status Request extension each time, then there will be unbounded
5852 memory growth on the server. This will eventually lead to a Denial Of
5853 Service attack through memory exhaustion. Servers with a default
5854 configuration are vulnerable even if they do not support OCSP. Builds using
5855 the "no-ocsp" build time option are not affected.
5857 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5862 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5865 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5871 * OOB write in MDC2_Update()
5873 An overflow can occur in MDC2_Update() either if called directly or
5874 through the EVP_DigestUpdate() function using MDC2. If an attacker
5875 is able to supply very large amounts of input data after a previous
5876 call to EVP_EncryptUpdate() with a partial block then a length check
5877 can overflow resulting in a heap corruption.
5879 The amount of data needed is comparable to SIZE_MAX which is impractical
5882 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5887 * Malformed SHA512 ticket DoS
5889 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5890 DoS attack where a malformed ticket will result in an OOB read which will
5893 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5894 a custom server callback and ticket lookup mechanism.
5896 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5901 * OOB write in BN_bn2dec()
5903 The function BN_bn2dec() does not check the return value of BN_div_word().
5904 This can cause an OOB write if an application uses this function with an
5905 overly large BIGNUM. This could be a problem if an overly large certificate
5906 or CRL is printed out from an untrusted source. TLS is not affected because
5907 record limits will reject an oversized certificate before it is parsed.
5909 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5914 * OOB read in TS_OBJ_print_bio()
5916 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5917 the total length the OID text representation would use and not the amount
5918 of data written. This will result in OOB reads when large OIDs are
5921 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5926 * Pointer arithmetic undefined behaviour
5928 Avoid some undefined pointer arithmetic
5930 A common idiom in the codebase is to check limits in the following manner:
5933 Where "p" points to some malloc'd data of SIZE bytes and
5936 "len" here could be from some externally supplied data (e.g. from a TLS
5939 The rules of C pointer arithmetic are such that "p + len" is only well
5940 defined where len <= SIZE. Therefore the above idiom is actually
5941 undefined behaviour.
5943 For example this could cause problems if some malloc implementation
5944 provides an address for "p" such that "p + len" actually overflows for
5945 values of len that are too big and therefore p + len < limit.
5947 This issue was reported to OpenSSL by Guido Vranken
5952 * Constant time flag not preserved in DSA signing
5954 Operations in the DSA signing algorithm should run in constant time in
5955 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5956 implementation means that a non-constant time codepath is followed for
5957 certain operations. This has been demonstrated through a cache-timing
5958 attack to be sufficient for an attacker to recover the private DSA key.
5960 This issue was reported by César Pereida (Aalto University), Billy Brumley
5961 (Tampere University of Technology), and Yuval Yarom (The University of
5962 Adelaide and NICTA).
5967 * DTLS buffered message DoS
5969 In a DTLS connection where handshake messages are delivered out-of-order
5970 those messages that OpenSSL is not yet ready to process will be buffered
5971 for later use. Under certain circumstances, a flaw in the logic means that
5972 those messages do not get removed from the buffer even though the handshake
5973 has been completed. An attacker could force up to approx. 15 messages to
5974 remain in the buffer when they are no longer required. These messages will
5975 be cleared when the DTLS connection is closed. The default maximum size for
5976 a message is 100k. Therefore the attacker could force an additional 1500k
5977 to be consumed per connection. By opening many simulataneous connections an
5978 attacker could cause a DoS attack through memory exhaustion.
5980 This issue was reported to OpenSSL by Quan Luo.
5985 * DTLS replay protection DoS
5987 A flaw in the DTLS replay attack protection mechanism means that records
5988 that arrive for future epochs update the replay protection "window" before
5989 the MAC for the record has been validated. This could be exploited by an
5990 attacker by sending a record for the next epoch (which does not have to
5991 decrypt or have a valid MAC), with a very large sequence number. This means
5992 that all subsequent legitimate packets are dropped causing a denial of
5993 service for a specific DTLS connection.
5995 This issue was reported to OpenSSL by the OCAP audit team.
6000 * Certificate message OOB reads
6002 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6003 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6004 theoretical DoS risk but this has not been observed in practice on common
6007 The messages affected are client certificate, client certificate request
6008 and server certificate. As a result the attack can only be performed
6009 against a client or a server which enables client authentication.
6011 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6016 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6018 * Prevent padding oracle in AES-NI CBC MAC check
6020 A MITM attacker can use a padding oracle attack to decrypt traffic
6021 when the connection uses an AES CBC cipher and the server support
6024 This issue was introduced as part of the fix for Lucky 13 padding
6025 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6026 constant time by making sure that always the same bytes are read and
6027 compared against either the MAC or padding bytes. But it no longer
6028 checked that there was enough data to have both the MAC and padding
6031 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6036 * Fix EVP_EncodeUpdate overflow
6038 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6039 Base64 encoding of binary data. If an attacker is able to supply very large
6040 amounts of input data then a length check can overflow resulting in a heap
6043 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6044 the `PEM_write_bio*` family of functions. These are mainly used within the
6045 OpenSSL command line applications, so any application which processes data
6046 from an untrusted source and outputs it as a PEM file should be considered
6047 vulnerable to this issue. User applications that call these APIs directly
6048 with large amounts of untrusted data may also be vulnerable.
6050 This issue was reported by Guido Vranken.
6055 * Fix EVP_EncryptUpdate overflow
6057 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6058 is able to supply very large amounts of input data after a previous call to
6059 EVP_EncryptUpdate() with a partial block then a length check can overflow
6060 resulting in a heap corruption. Following an analysis of all OpenSSL
6061 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6062 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6063 the first called function after an EVP_EncryptInit(), and therefore that
6064 specific call must be safe. The second form is where the length passed to
6065 EVP_EncryptUpdate() can be seen from the code to be some small value and
6066 therefore there is no possibility of an overflow. Since all instances are
6067 one of these two forms, it is believed that there can be no overflows in
6068 internal code due to this problem. It should be noted that
6069 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6070 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6071 of these calls have also been analysed too and it is believed there are no
6072 instances in internal usage where an overflow could occur.
6074 This issue was reported by Guido Vranken.
6079 * Prevent ASN.1 BIO excessive memory allocation
6081 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6082 a short invalid encoding can casuse allocation of large amounts of memory
6083 potentially consuming excessive resources or exhausting memory.
6085 Any application parsing untrusted data through d2i BIO functions is
6086 affected. The memory based functions such as d2i_X509() are *not* affected.
6087 Since the memory based functions are used by the TLS library, TLS
6088 applications are not affected.
6090 This issue was reported by Brian Carpenter.
6097 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6098 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6099 in arbitrary stack data being returned in the buffer.
6101 This issue was reported by Guido Vranken.
6106 * Modify behavior of ALPN to invoke callback after SNI/servername
6107 callback, such that updates to the SSL_CTX affect ALPN.
6111 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6116 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6117 methods are enabled and ssl2 is disabled the methods return NULL.
6121 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6123 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6124 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6125 provide any "EXPORT" or "LOW" strength ciphers.
6129 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6130 is by default disabled at build-time. Builds that are not configured with
6131 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6132 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6133 will need to explicitly call either of:
6135 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6137 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6139 as appropriate. Even if either of those is used, or the application
6140 explicitly uses the version-specific SSLv2_method() or its client and
6141 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6142 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6143 ciphers, and SSLv2 56-bit DES are no longer available.
6148 * Fix a double-free in DSA code
6150 A double free bug was discovered when OpenSSL parses malformed DSA private
6151 keys and could lead to a DoS attack or memory corruption for applications
6152 that receive DSA private keys from untrusted sources. This scenario is
6155 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6161 * Disable SRP fake user seed to address a server memory leak.
6163 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6165 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6166 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6167 was changed to ignore the "fake user" SRP seed, even if the seed
6170 Users should use SRP_VBASE_get1_by_user instead. Note that in
6171 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6172 also that even though configuring the SRP seed attempts to hide
6173 invalid usernames by continuing the handshake with fake
6174 credentials, this behaviour is not constant time and no strong
6175 guarantees are made that the handshake is indistinguishable from
6176 that of a valid user.
6181 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6183 In the BN_hex2bn function the number of hex digits is calculated using an
6184 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6185 large values of `i` this can result in `bn_expand` not allocating any
6186 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6187 field as NULL leading to a subsequent NULL ptr deref. For very large values
6188 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6189 In this case memory is allocated to the internal BIGNUM data field, but it
6190 is insufficiently sized leading to heap corruption. A similar issue exists
6191 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6192 is ever called by user applications with very large untrusted hex/dec data.
6193 This is anticipated to be a rare occurrence.
6195 All OpenSSL internal usage of these functions use data that is not expected
6196 to be untrusted, e.g. config file data or application command line
6197 arguments. If user developed applications generate config file data based
6198 on untrusted data then it is possible that this could also lead to security
6199 consequences. This is also anticipated to be rare.
6201 This issue was reported to OpenSSL by Guido Vranken.
6206 * Fix memory issues in `BIO_*printf` functions
6208 The internal `fmtstr` function used in processing a "%s" format string in
6209 the `BIO_*printf` functions could overflow while calculating the length of a
6210 string and cause an OOB read when printing very long strings.
6212 Additionally the internal `doapr_outch` function can attempt to write to an
6213 OOB memory location (at an offset from the NULL pointer) in the event of a
6214 memory allocation failure. In 1.0.2 and below this could be caused where
6215 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6216 could be in processing a very long "%s" format string. Memory leaks can
6219 The first issue may mask the second issue dependent on compiler behaviour.
6220 These problems could enable attacks where large amounts of untrusted data
6221 is passed to the `BIO_*printf` functions. If applications use these functions
6222 in this way then they could be vulnerable. OpenSSL itself uses these
6223 functions when printing out human-readable dumps of ASN.1 data. Therefore
6224 applications that print this data could be vulnerable if the data is from
6225 untrusted sources. OpenSSL command line applications could also be
6226 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6227 as command line arguments.
6229 Libssl is not considered directly vulnerable. Additionally certificates etc
6230 received via remote connections via libssl are also unlikely to be able to
6231 trigger these issues because of message size limits enforced within libssl.
6233 This issue was reported to OpenSSL Guido Vranken.
6238 * Side channel attack on modular exponentiation
6240 A side-channel attack was found which makes use of cache-bank conflicts on
6241 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6242 of RSA keys. The ability to exploit this issue is limited as it relies on
6243 an attacker who has control of code in a thread running on the same
6244 hyper-threaded core as the victim thread which is performing decryptions.
6246 This issue was reported to OpenSSL by Yuval Yarom, The University of
6247 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6248 Nadia Heninger, University of Pennsylvania with more information at
6249 <http://cachebleed.info>.
6254 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6255 if no keysize is specified with default_bits. This fixes an
6256 omission in an earlier change that changed all RSA/DSA key generation
6257 commands to use 2048 bits by default.
6261 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6263 * Protection for DH small subgroup attacks
6265 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6266 switched on by default and cannot be disabled. This could have some
6271 * SSLv2 doesn't block disabled ciphers
6273 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6274 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6275 been disabled, provided that the SSLv2 protocol was not also disabled via
6278 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6279 and Sebastian Schinzel.
6284 * Reject DH handshakes with parameters shorter than 1024 bits.
6288 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6290 * Certificate verify crash with missing PSS parameter
6292 The signature verification routines will crash with a NULL pointer
6293 dereference if presented with an ASN.1 signature using the RSA PSS
6294 algorithm and absent mask generation function parameter. Since these
6295 routines are used to verify certificate signature algorithms this can be
6296 used to crash any certificate verification operation and exploited in a
6297 DoS attack. Any application which performs certificate verification is
6298 vulnerable including OpenSSL clients and servers which enable client
6301 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6306 * X509_ATTRIBUTE memory leak
6308 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6309 memory. This structure is used by the PKCS#7 and CMS routines so any
6310 application which reads PKCS#7 or CMS data from untrusted sources is
6311 affected. SSL/TLS is not affected.
6313 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6319 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6320 This changes the decoding behaviour for some invalid messages,
6321 though the change is mostly in the more lenient direction, and
6322 legacy behaviour is preserved as much as possible.
6326 * In DSA_generate_parameters_ex, if the provided seed is too short,
6327 use a random seed, as already documented.
6329 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6331 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6333 * Alternate chains certificate forgery
6335 During certificate verfification, OpenSSL will attempt to find an
6336 alternative certificate chain if the first attempt to build such a chain
6337 fails. An error in the implementation of this logic can mean that an
6338 attacker could cause certain checks on untrusted certificates to be
6339 bypassed, such as the CA flag, enabling them to use a valid leaf
6340 certificate to act as a CA and "issue" an invalid certificate.
6342 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6348 * Race condition handling PSK identify hint
6350 If PSK identity hints are received by a multi-threaded client then
6351 the values are wrongly updated in the parent SSL_CTX structure. This can
6352 result in a race condition potentially leading to a double free of the
6358 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6360 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6361 incompatibility in the handling of HMAC. The previous ABI has now been
6364 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6366 * Malformed ECParameters causes infinite loop
6368 When processing an ECParameters structure OpenSSL enters an infinite loop
6369 if the curve specified is over a specially malformed binary polynomial
6372 This can be used to perform denial of service against any
6373 system which processes public keys, certificate requests or
6374 certificates. This includes TLS clients and TLS servers with
6375 client authentication enabled.
6377 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6382 * Exploitable out-of-bounds read in X509_cmp_time
6384 X509_cmp_time does not properly check the length of the ASN1_TIME
6385 string and can read a few bytes out of bounds. In addition,
6386 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6389 An attacker can use this to craft malformed certificates and CRLs of
6390 various sizes and potentially cause a segmentation fault, resulting in
6391 a DoS on applications that verify certificates or CRLs. TLS clients
6392 that verify CRLs are affected. TLS clients and servers with client
6393 authentication enabled may be affected if they use custom verification
6396 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6397 independently by Hanno Böck.
6402 * PKCS7 crash with missing EnvelopedContent
6404 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6405 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6406 with missing content and trigger a NULL pointer dereference on parsing.
6408 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6409 structures from untrusted sources are affected. OpenSSL clients and
6410 servers are not affected.
6412 This issue was reported to OpenSSL by Michal Zalewski (Google).
6417 * CMS verify infinite loop with unknown hash function
6419 When verifying a signedData message the CMS code can enter an infinite loop
6420 if presented with an unknown hash function OID. This can be used to perform
6421 denial of service against any system which verifies signedData messages using
6423 This issue was reported to OpenSSL by Johannes Bauer.
6428 * Race condition handling NewSessionTicket
6430 If a NewSessionTicket is received by a multi-threaded client when attempting to
6431 reuse a previous ticket then a race condition can occur potentially leading to
6432 a double free of the ticket data.
6437 * Reject DH handshakes with parameters shorter than 768 bits.
6439 *Kurt Roeckx and Emilia Kasper*
6441 * dhparam: generate 2048-bit parameters by default.
6443 *Kurt Roeckx and Emilia Kasper*
6445 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6447 * Segmentation fault in ASN1_TYPE_cmp fix
6449 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6450 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6451 certificate signature algorithm consistency this can be used to crash any
6452 certificate verification operation and exploited in a DoS attack. Any
6453 application which performs certificate verification is vulnerable including
6454 OpenSSL clients and servers which enable client authentication.
6459 * ASN.1 structure reuse memory corruption fix
6461 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6462 memory corruption via an invalid write. Such reuse is and has been
6463 strongly discouraged and is believed to be rare.
6465 Applications that parse structures containing CHOICE or ANY DEFINED BY
6466 components may be affected. Certificate parsing (d2i_X509 and related
6467 functions) are however not affected. OpenSSL clients and servers are
6473 * PKCS7 NULL pointer dereferences fix
6475 The PKCS#7 parsing code does not handle missing outer ContentInfo
6476 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6477 missing content and trigger a NULL pointer dereference on parsing.
6479 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6480 otherwise parse PKCS#7 structures from untrusted sources are
6481 affected. OpenSSL clients and servers are not affected.
6483 This issue was reported to OpenSSL by Michal Zalewski (Google).
6488 * DoS via reachable assert in SSLv2 servers fix
6490 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6491 servers that both support SSLv2 and enable export cipher suites by sending
6492 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6494 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6495 (OpenSSL development team).
6500 * Use After Free following d2i_ECPrivatekey error fix
6502 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6503 could cause a use after free condition. This, in turn, could cause a double
6504 free in several private key parsing functions (such as d2i_PrivateKey
6505 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6506 for applications that receive EC private keys from untrusted
6507 sources. This scenario is considered rare.
6509 This issue was discovered by the BoringSSL project and fixed in their
6515 * X509_to_X509_REQ NULL pointer deref fix
6517 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6518 the certificate key is invalid. This function is rarely used in practice.
6520 This issue was discovered by Brian Carpenter.
6525 * Removed the export ciphers from the DEFAULT ciphers
6529 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6531 * Build fixes for the Windows and OpenVMS platforms
6533 *Matt Caswell and Richard Levitte*
6535 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6537 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6538 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6539 dereference. This could lead to a Denial Of Service attack. Thanks to
6540 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6545 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6546 dtls1_buffer_record function under certain conditions. In particular this
6547 could occur if an attacker sent repeated DTLS records with the same
6548 sequence number but for the next epoch. The memory leak could be exploited
6549 by an attacker in a Denial of Service attack through memory exhaustion.
6550 Thanks to Chris Mueller for reporting this issue.
6555 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6556 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6557 method would be set to NULL which could later result in a NULL pointer
6558 dereference. Thanks to Frank Schmirler for reporting this issue.
6563 * Abort handshake if server key exchange message is omitted for ephemeral
6566 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6567 reporting this issue.
6572 * Remove non-export ephemeral RSA code on client and server. This code
6573 violated the TLS standard by allowing the use of temporary RSA keys in
6574 non-export ciphersuites and could be used by a server to effectively
6575 downgrade the RSA key length used to a value smaller than the server
6576 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6577 INRIA or reporting this issue.
6582 * Fixed issue where DH client certificates are accepted without verification.
6583 An OpenSSL server will accept a DH certificate for client authentication
6584 without the certificate verify message. This effectively allows a client to
6585 authenticate without the use of a private key. This only affects servers
6586 which trust a client certificate authority which issues certificates
6587 containing DH keys: these are extremely rare and hardly ever encountered.
6588 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6594 * Ensure that the session ID context of an SSL is updated when its
6595 SSL_CTX is updated via SSL_set_SSL_CTX.
6597 The session ID context is typically set from the parent SSL_CTX,
6598 and can vary with the CTX.
6602 * Fix various certificate fingerprint issues.
6604 By using non-DER or invalid encodings outside the signed portion of a
6605 certificate the fingerprint can be changed without breaking the signature.
6606 Although no details of the signed portion of the certificate can be changed
6607 this can cause problems with some applications: e.g. those using the
6608 certificate fingerprint for blacklists.
6610 1. Reject signatures with non zero unused bits.
6612 If the BIT STRING containing the signature has non zero unused bits reject
6613 the signature. All current signature algorithms require zero unused bits.
6615 2. Check certificate algorithm consistency.
6617 Check the AlgorithmIdentifier inside TBS matches the one in the
6618 certificate signature. NB: this will result in signature failure
6619 errors for some broken certificates.
6621 Thanks to Konrad Kraszewski from Google for reporting this issue.
6623 3. Check DSA/ECDSA signatures use DER.
6625 Re-encode DSA/ECDSA signatures and compare with the original received
6626 signature. Return an error if there is a mismatch.
6628 This will reject various cases including garbage after signature
6629 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6630 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6631 (negative or with leading zeroes).
6633 Further analysis was conducted and fixes were developed by Stephen Henson
6634 of the OpenSSL core team.
6640 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6641 results on some platforms, including x86_64. This bug occurs at random
6642 with a very low probability, and is not known to be exploitable in any
6643 way, though its exact impact is difficult to determine. Thanks to Pieter
6644 Wuille (Blockstream) who reported this issue and also suggested an initial
6645 fix. Further analysis was conducted by the OpenSSL development team and
6646 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6647 the OpenSSL core team.
6652 * Do not resume sessions on the server if the negotiated protocol
6653 version does not match the session's version. Resuming with a different
6654 version, while not strictly forbidden by the RFC, is of questionable
6655 sanity and breaks all known clients.
6657 *David Benjamin, Emilia Käsper*
6659 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6660 early CCS messages during renegotiation. (Note that because
6661 renegotiation is encrypted, this early CCS was not exploitable.)
6665 * Tighten client-side session ticket handling during renegotiation:
6666 ensure that the client only accepts a session ticket if the server sends
6667 the extension anew in the ServerHello. Previously, a TLS client would
6668 reuse the old extension state and thus accept a session ticket if one was
6669 announced in the initial ServerHello.
6671 Similarly, ensure that the client requires a session ticket if one
6672 was advertised in the ServerHello. Previously, a TLS client would
6673 ignore a missing NewSessionTicket message.
6677 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6681 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6682 sends a carefully crafted handshake message, to cause OpenSSL to fail
6683 to free up to 64k of memory causing a memory leak. This could be
6684 exploited in a Denial Of Service attack. This issue affects OpenSSL
6685 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6686 whether SRTP is used or configured. Implementations of OpenSSL that
6687 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6689 The fix was developed by the OpenSSL team.
6694 * Session Ticket Memory Leak.
6696 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6697 integrity of that ticket is first verified. In the event of a session
6698 ticket integrity check failing, OpenSSL will fail to free memory
6699 causing a memory leak. By sending a large number of invalid session
6700 tickets an attacker could exploit this issue in a Denial Of Service
6706 * Build option no-ssl3 is incomplete.
6708 When OpenSSL is configured with "no-ssl3" as a build option, servers
6709 could accept and complete a SSL 3.0 handshake, and clients could be
6710 configured to send them.
6713 *Akamai and the OpenSSL team*
6715 * Add support for TLS_FALLBACK_SCSV.
6716 Client applications doing fallback retries should call
6717 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6720 *Adam Langley, Bodo Moeller*
6722 * Add additional DigestInfo checks.
6724 Re-encode DigestInto in DER and check against the original when
6725 verifying RSA signature: this will reject any improperly encoded
6726 DigestInfo structures.
6728 Note: this is a precautionary measure and no attacks are currently known.
6732 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6734 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6735 SRP code can be overrun an internal buffer. Add sanity check that
6736 g, A, B < N to SRP code.
6738 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6739 Group for discovering this issue.
6744 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6745 TLS 1.0 instead of higher protocol versions when the ClientHello message
6746 is badly fragmented. This allows a man-in-the-middle attacker to force a
6747 downgrade to TLS 1.0 even if both the server and the client support a
6748 higher protocol version, by modifying the client's TLS records.
6750 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6751 researching this issue.
6756 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6757 to a denial of service attack. A malicious server can crash the client
6758 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6759 ciphersuite and sending carefully crafted handshake messages.
6761 Thanks to Felix Gröbert (Google) for discovering and researching this
6767 * By sending carefully crafted DTLS packets an attacker could cause openssl
6768 to leak memory. This can be exploited through a Denial of Service attack.
6769 Thanks to Adam Langley for discovering and researching this issue.
6774 * An attacker can force openssl to consume large amounts of memory whilst
6775 processing DTLS handshake messages. This can be exploited through a
6776 Denial of Service attack.
6777 Thanks to Adam Langley for discovering and researching this issue.
6782 * An attacker can force an error condition which causes openssl to crash
6783 whilst processing DTLS packets due to memory being freed twice. This
6784 can be exploited through a Denial of Service attack.
6785 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6791 * If a multithreaded client connects to a malicious server using a resumed
6792 session and the server sends an ec point format extension it could write
6793 up to 255 bytes to freed memory.
6795 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6801 * A malicious server can crash an OpenSSL client with a null pointer
6802 dereference (read) by specifying an SRP ciphersuite even though it was not
6803 properly negotiated with the client. This can be exploited through a
6804 Denial of Service attack.
6806 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6807 discovering and researching this issue.
6812 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6813 X509_name_oneline, X509_name_print_ex et al. to leak some information
6814 from the stack. Applications may be affected if they echo pretty printing
6815 output to the attacker.
6817 Thanks to Ivan Fratric (Google) for discovering this issue.
6820 *Emilia Käsper, and Steve Henson*
6822 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6823 for corner cases. (Certain input points at infinity could lead to
6824 bogus results, with non-infinity inputs mapped to infinity too.)
6828 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6830 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6831 handshake can force the use of weak keying material in OpenSSL
6832 SSL/TLS clients and servers.
6834 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6835 researching this issue. ([CVE-2014-0224])
6837 *KIKUCHI Masashi, Steve Henson*
6839 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6840 OpenSSL DTLS client the code can be made to recurse eventually crashing
6843 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6846 *Imre Rad, Steve Henson*
6848 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6849 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6850 client or server. This is potentially exploitable to run arbitrary
6851 code on a vulnerable client or server.
6853 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6855 *Jüri Aedla, Steve Henson*
6857 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6858 are subject to a denial of service attack.
6860 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6861 this issue. ([CVE-2014-3470])
6863 *Felix Gröbert, Ivan Fratric, Steve Henson*
6865 * Harmonize version and its documentation. -f flag is used to display
6868 *mancha <mancha1@zoho.com>*
6870 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6871 in i2d_ECPrivateKey.
6873 *mancha <mancha1@zoho.com>*
6875 * Fix some double frees. These are not thought to be exploitable.
6877 *mancha <mancha1@zoho.com>*
6879 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6881 * A missing bounds check in the handling of the TLS heartbeat extension
6882 can be used to reveal up to 64k of memory to a connected client or
6885 Thanks for Neel Mehta of Google Security for discovering this bug and to
6886 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6887 preparing the fix ([CVE-2014-0160])
6889 *Adam Langley, Bodo Moeller*
6891 * Fix for the attack described in the paper "Recovering OpenSSL
6892 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6893 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6894 <http://eprint.iacr.org/2014/140>
6896 Thanks to Yuval Yarom and Naomi Benger for discovering this
6897 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6899 *Yuval Yarom and Naomi Benger*
6901 * TLS pad extension: draft-agl-tls-padding-03
6903 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6904 TLS client Hello record length value would otherwise be > 255 and
6905 less that 512 pad with a dummy extension containing zeroes so it
6906 is at least 512 bytes long.
6908 *Adam Langley, Steve Henson*
6910 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6912 * Fix for TLS record tampering bug. A carefully crafted invalid
6913 handshake could crash OpenSSL with a NULL pointer exception.
6914 Thanks to Anton Johansson for reporting this issues.
6917 * Keep original DTLS digest and encryption contexts in retransmission
6918 structures so we can use the previous session parameters if they need
6919 to be resent. ([CVE-2013-6450])
6923 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6924 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6925 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6926 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6927 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6928 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6930 *Rob Stradling, Adam Langley*
6932 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6934 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6935 supporting platforms or when small records were transferred.
6937 *Andy Polyakov, Steve Henson*
6939 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6941 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6943 This addresses the flaw in CBC record processing discovered by
6944 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6945 at: <http://www.isg.rhul.ac.uk/tls/>
6947 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6948 Security Group at Royal Holloway, University of London
6949 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6950 Emilia Käsper for the initial patch.
6953 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6955 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6956 ciphersuites which can be exploited in a denial of service attack.
6957 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6958 and detecting this bug and to Wolfgang Ettlinger
6959 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6964 * Return an error when checking OCSP signatures when key is NULL.
6965 This fixes a DoS attack. ([CVE-2013-0166])
6969 * Make openssl verify return errors.
6971 *Chris Palmer <palmer@google.com> and Ben Laurie*
6973 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6974 the right response is stapled. Also change SSL_get_certificate()
6975 so it returns the certificate actually sent.
6976 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6978 *Rob Stradling <rob.stradling@comodo.com>*
6980 * Fix possible deadlock when decoding public keys.
6984 * Don't use TLS 1.0 record version number in initial client hello
6989 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6991 * Sanity check record length before skipping explicit IV in TLS
6992 1.2, 1.1 and DTLS to fix DoS attack.
6994 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6995 fuzzing as a service testing platform.
7000 * Initialise tkeylen properly when encrypting CMS messages.
7001 Thanks to Solar Designer of Openwall for reporting this issue.
7005 * In FIPS mode don't try to use composite ciphers as they are not
7010 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7012 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7013 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7014 mean any application compiled against OpenSSL 1.0.0 headers setting
7015 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7016 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7017 0x10000000L Any application which was previously compiled against
7018 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7019 will need to be recompiled as a result. Letting be results in
7020 inability to disable specifically TLS 1.1 and in client context,
7021 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7025 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7026 disable just protocol X, but all protocols above X *if* there are
7027 protocols *below* X still enabled. In more practical terms it means
7028 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7029 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7030 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7035 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7037 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7038 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7039 in CRYPTO_realloc_clean.
7041 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7042 issue and to Adam Langley <agl@chromium.org> for fixing it.
7045 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7047 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7051 * Workarounds for some broken servers that "hang" if a client hello
7052 record length exceeds 255 bytes.
7054 1. Do not use record version number > TLS 1.0 in initial client
7055 hello: some (but not all) hanging servers will now work.
7056 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7057 the number of ciphers sent in the client hello. This should be
7058 set to an even number, such as 50, for example by passing:
7059 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7060 Most broken servers should now work.
7061 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7062 TLS 1.2 client support entirely.
7066 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7070 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7072 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7073 STRING form instead of a DigestInfo.
7077 * The format used for MDC2 RSA signatures is inconsistent between EVP
7078 and the RSA_sign/RSA_verify functions. This was made more apparent when
7079 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7080 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7081 the correct format in RSA_verify so both forms transparently work.
7085 * Some servers which support TLS 1.0 can choke if we initially indicate
7086 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7087 encrypted premaster secret. As a workaround use the maximum permitted
7088 client version in client hello, this should keep such servers happy
7089 and still work with previous versions of OpenSSL.
7093 * Add support for TLS/DTLS heartbeats.
7095 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7097 * Add support for SCTP.
7099 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7101 * Improved PRNG seeding for VOS.
7103 *Paul Green <Paul.Green@stratus.com>*
7105 * Extensive assembler packs updates, most notably:
7107 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7108 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7109 - x86_64: bit-sliced AES implementation;
7110 - ARM: NEON support, contemporary platforms optimizations;
7111 - s390x: z196 support;
7112 - `*`: GHASH and GF(2^m) multiplication implementations;
7116 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7117 (removal of unnecessary code)
7119 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7121 * Add TLS key material exporter from RFC 5705.
7125 * Add DTLS-SRTP negotiation from RFC 5764.
7129 * Add Next Protocol Negotiation,
7130 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7131 disabled with a no-npn flag to config or Configure. Code donated
7134 *Adam Langley <agl@google.com> and Ben Laurie*
7136 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7137 NIST-P256, NIST-P521, with constant-time single point multiplication on
7138 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7139 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7140 Code made available under Apache License version 2.0.
7142 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7143 line to include this in your build of OpenSSL, and run "make depend" (or
7144 "make update"). This enables the following EC_METHODs:
7146 EC_GFp_nistp224_method()
7147 EC_GFp_nistp256_method()
7148 EC_GFp_nistp521_method()
7150 EC_GROUP_new_by_curve_name() will automatically use these (while
7151 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7154 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7156 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7157 all platforms. Move ssize_t definition from e_os.h to the public
7158 header file e_os2.h as it now appears in public header file cms.h
7162 * New -sigopt option to the ca, req and x509 utilities. Additional
7163 signature parameters can be passed using this option and in
7168 * Add RSA PSS signing function. This will generate and set the
7169 appropriate AlgorithmIdentifiers for PSS based on those in the
7170 corresponding EVP_MD_CTX structure. No application support yet.
7174 * Support for companion algorithm specific ASN1 signing routines.
7175 New function ASN1_item_sign_ctx() signs a pre-initialised
7176 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7177 the appropriate parameters.
7181 * Add new algorithm specific ASN1 verification initialisation function
7182 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7183 handling will be the same no matter what EVP_PKEY_METHOD is used.
7184 Add a PSS handler to support verification of PSS signatures: checked
7185 against a number of sample certificates.
7189 * Add signature printing for PSS. Add PSS OIDs.
7191 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7193 * Add algorithm specific signature printing. An individual ASN1 method
7194 can now print out signatures instead of the standard hex dump.
7196 More complex signatures (e.g. PSS) can print out more meaningful
7197 information. Include DSA version that prints out the signature
7202 * Password based recipient info support for CMS library: implementing
7207 * Split password based encryption into PBES2 and PBKDF2 functions. This
7208 neatly separates the code into cipher and PBE sections and is required
7209 for some algorithms that split PBES2 into separate pieces (such as
7210 password based CMS).
7214 * Session-handling fixes:
7215 - Fix handling of connections that are resuming with a session ID,
7216 but also support Session Tickets.
7217 - Fix a bug that suppressed issuing of a new ticket if the client
7218 presented a ticket with an expired session.
7219 - Try to set the ticket lifetime hint to something reasonable.
7220 - Make tickets shorter by excluding irrelevant information.
7221 - On the client side, don't ignore renewed tickets.
7223 *Adam Langley, Bodo Moeller (Google)*
7225 * Fix PSK session representation.
7229 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7231 This work was sponsored by Intel.
7235 * Add GCM support to TLS library. Some custom code is needed to split
7236 the IV between the fixed (from PRF) and explicit (from TLS record)
7237 portions. This adds all GCM ciphersuites supported by RFC5288 and
7238 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7239 add a special AESGCM string for GCM only.
7243 * Expand range of ctrls for AES GCM. Permit setting invocation
7244 field on decrypt and retrieval of invocation field only on encrypt.
7248 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7249 As required by RFC5289 these ciphersuites cannot be used if for
7250 versions of TLS earlier than 1.2.
7254 * For FIPS capable OpenSSL interpret a NULL default public key method
7255 as unset and return the appropriate default but do *not* set the default.
7256 This means we can return the appropriate method in applications that
7257 switch between FIPS and non-FIPS modes.
7261 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7262 ENGINE is used then we cannot handle that in the FIPS module so we
7263 keep original code iff non-FIPS operations are allowed.
7267 * Add -attime option to openssl utilities.
7269 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7271 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7275 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7276 FIPS EC methods unconditionally for now.
7280 * New build option no-ec2m to disable characteristic 2 code.
7284 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7285 all cases can be covered as some introduce binary incompatibilities.
7289 * Redirect RSA operations to FIPS module including keygen,
7290 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7294 * Add similar low level API blocking to ciphers.
7298 * Low level digest APIs are not approved in FIPS mode: any attempt
7299 to use these will cause a fatal error. Applications that *really* want
7300 to use them can use the `private_*` version instead.
7304 * Redirect cipher operations to FIPS module for FIPS builds.
7308 * Redirect digest operations to FIPS module for FIPS builds.
7312 * Update build system to add "fips" flag which will link in fipscanister.o
7313 for static and shared library builds embedding a signature if needed.
7317 * Output TLS supported curves in preference order instead of numerical
7318 order. This is currently hardcoded for the highest order curves first.
7319 This should be configurable so applications can judge speed vs strength.
7323 * Add TLS v1.2 server support for client authentication.
7327 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7332 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7333 FIPS modules versions.
7337 * Add TLS v1.2 client side support for client authentication. Keep cache
7338 of handshake records longer as we don't know the hash algorithm to use
7339 until after the certificate request message is received.
7343 * Initial TLS v1.2 client support. Add a default signature algorithms
7344 extension including all the algorithms we support. Parse new signature
7345 format in client key exchange. Relax some ECC signing restrictions for
7346 TLS v1.2 as indicated in RFC5246.
7350 * Add server support for TLS v1.2 signature algorithms extension. Switch
7351 to new signature format when needed using client digest preference.
7352 All server ciphersuites should now work correctly in TLS v1.2. No client
7353 support yet and no support for client certificates.
7357 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7358 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7359 ciphersuites. At present only RSA key exchange ciphersuites work with
7360 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7361 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7362 and version checking.
7366 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7367 with this defined it will not be affected by any changes to ssl internal
7368 structures. Add several utility functions to allow openssl application
7369 to work with OPENSSL_NO_SSL_INTERN defined.
7373 * A long standing patch to add support for SRP from EdelWeb (Peter
7374 Sylvester and Christophe Renou) was integrated.
7375 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7376 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7379 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7383 * Permit abbreviated handshakes when renegotiating using the function
7384 SSL_renegotiate_abbreviated().
7386 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7388 * Add call to ENGINE_register_all_complete() to
7389 ENGINE_load_builtin_engines(), so some implementations get used
7390 automatically instead of needing explicit application support.
7394 * Add support for TLS key exporter as described in RFC5705.
7396 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7398 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7399 a few changes are required:
7401 Add SSL_OP_NO_TLSv1_1 flag.
7402 Add TLSv1_1 methods.
7403 Update version checking logic to handle version 1.1.
7404 Add explicit IV handling (ported from DTLS code).
7405 Add command line options to s_client/s_server.
7412 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7414 * X509_ATTRIBUTE memory leak
7416 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7417 memory. This structure is used by the PKCS#7 and CMS routines so any
7418 application which reads PKCS#7 or CMS data from untrusted sources is
7419 affected. SSL/TLS is not affected.
7421 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7427 * Race condition handling PSK identify hint
7429 If PSK identity hints are received by a multi-threaded client then
7430 the values are wrongly updated in the parent SSL_CTX structure. This can
7431 result in a race condition potentially leading to a double free of the
7437 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7439 * Malformed ECParameters causes infinite loop
7441 When processing an ECParameters structure OpenSSL enters an infinite loop
7442 if the curve specified is over a specially malformed binary polynomial
7445 This can be used to perform denial of service against any
7446 system which processes public keys, certificate requests or
7447 certificates. This includes TLS clients and TLS servers with
7448 client authentication enabled.
7450 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7455 * Exploitable out-of-bounds read in X509_cmp_time
7457 X509_cmp_time does not properly check the length of the ASN1_TIME
7458 string and can read a few bytes out of bounds. In addition,
7459 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7462 An attacker can use this to craft malformed certificates and CRLs of
7463 various sizes and potentially cause a segmentation fault, resulting in
7464 a DoS on applications that verify certificates or CRLs. TLS clients
7465 that verify CRLs are affected. TLS clients and servers with client
7466 authentication enabled may be affected if they use custom verification
7469 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7470 independently by Hanno Böck.
7475 * PKCS7 crash with missing EnvelopedContent
7477 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7478 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7479 with missing content and trigger a NULL pointer dereference on parsing.
7481 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7482 structures from untrusted sources are affected. OpenSSL clients and
7483 servers are not affected.
7485 This issue was reported to OpenSSL by Michal Zalewski (Google).
7490 * CMS verify infinite loop with unknown hash function
7492 When verifying a signedData message the CMS code can enter an infinite loop
7493 if presented with an unknown hash function OID. This can be used to perform
7494 denial of service against any system which verifies signedData messages using
7496 This issue was reported to OpenSSL by Johannes Bauer.
7501 * Race condition handling NewSessionTicket
7503 If a NewSessionTicket is received by a multi-threaded client when attempting to
7504 reuse a previous ticket then a race condition can occur potentially leading to
7505 a double free of the ticket data.
7510 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7512 * Segmentation fault in ASN1_TYPE_cmp fix
7514 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7515 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7516 certificate signature algorithm consistency this can be used to crash any
7517 certificate verification operation and exploited in a DoS attack. Any
7518 application which performs certificate verification is vulnerable including
7519 OpenSSL clients and servers which enable client authentication.
7524 * ASN.1 structure reuse memory corruption fix
7526 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7527 memory corruption via an invalid write. Such reuse is and has been
7528 strongly discouraged and is believed to be rare.
7530 Applications that parse structures containing CHOICE or ANY DEFINED BY
7531 components may be affected. Certificate parsing (d2i_X509 and related
7532 functions) are however not affected. OpenSSL clients and servers are
7538 * PKCS7 NULL pointer dereferences fix
7540 The PKCS#7 parsing code does not handle missing outer ContentInfo
7541 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7542 missing content and trigger a NULL pointer dereference on parsing.
7544 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7545 otherwise parse PKCS#7 structures from untrusted sources are
7546 affected. OpenSSL clients and servers are not affected.
7548 This issue was reported to OpenSSL by Michal Zalewski (Google).
7553 * DoS via reachable assert in SSLv2 servers fix
7555 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7556 servers that both support SSLv2 and enable export cipher suites by sending
7557 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7559 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7560 (OpenSSL development team).
7565 * Use After Free following d2i_ECPrivatekey error fix
7567 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7568 could cause a use after free condition. This, in turn, could cause a double
7569 free in several private key parsing functions (such as d2i_PrivateKey
7570 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7571 for applications that receive EC private keys from untrusted
7572 sources. This scenario is considered rare.
7574 This issue was discovered by the BoringSSL project and fixed in their
7580 * X509_to_X509_REQ NULL pointer deref fix
7582 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7583 the certificate key is invalid. This function is rarely used in practice.
7585 This issue was discovered by Brian Carpenter.
7590 * Removed the export ciphers from the DEFAULT ciphers
7594 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7596 * Build fixes for the Windows and OpenVMS platforms
7598 *Matt Caswell and Richard Levitte*
7600 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7602 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7603 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7604 dereference. This could lead to a Denial Of Service attack. Thanks to
7605 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7610 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7611 dtls1_buffer_record function under certain conditions. In particular this
7612 could occur if an attacker sent repeated DTLS records with the same
7613 sequence number but for the next epoch. The memory leak could be exploited
7614 by an attacker in a Denial of Service attack through memory exhaustion.
7615 Thanks to Chris Mueller for reporting this issue.
7620 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7621 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7622 method would be set to NULL which could later result in a NULL pointer
7623 dereference. Thanks to Frank Schmirler for reporting this issue.
7628 * Abort handshake if server key exchange message is omitted for ephemeral
7631 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7632 reporting this issue.
7637 * Remove non-export ephemeral RSA code on client and server. This code
7638 violated the TLS standard by allowing the use of temporary RSA keys in
7639 non-export ciphersuites and could be used by a server to effectively
7640 downgrade the RSA key length used to a value smaller than the server
7641 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7642 INRIA or reporting this issue.
7647 * Fixed issue where DH client certificates are accepted without verification.
7648 An OpenSSL server will accept a DH certificate for client authentication
7649 without the certificate verify message. This effectively allows a client to
7650 authenticate without the use of a private key. This only affects servers
7651 which trust a client certificate authority which issues certificates
7652 containing DH keys: these are extremely rare and hardly ever encountered.
7653 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7659 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7660 results on some platforms, including x86_64. This bug occurs at random
7661 with a very low probability, and is not known to be exploitable in any
7662 way, though its exact impact is difficult to determine. Thanks to Pieter
7663 Wuille (Blockstream) who reported this issue and also suggested an initial
7664 fix. Further analysis was conducted by the OpenSSL development team and
7665 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7666 the OpenSSL core team.
7671 * Fix various certificate fingerprint issues.
7673 By using non-DER or invalid encodings outside the signed portion of a
7674 certificate the fingerprint can be changed without breaking the signature.
7675 Although no details of the signed portion of the certificate can be changed
7676 this can cause problems with some applications: e.g. those using the
7677 certificate fingerprint for blacklists.
7679 1. Reject signatures with non zero unused bits.
7681 If the BIT STRING containing the signature has non zero unused bits reject
7682 the signature. All current signature algorithms require zero unused bits.
7684 2. Check certificate algorithm consistency.
7686 Check the AlgorithmIdentifier inside TBS matches the one in the
7687 certificate signature. NB: this will result in signature failure
7688 errors for some broken certificates.
7690 Thanks to Konrad Kraszewski from Google for reporting this issue.
7692 3. Check DSA/ECDSA signatures use DER.
7694 Reencode DSA/ECDSA signatures and compare with the original received
7695 signature. Return an error if there is a mismatch.
7697 This will reject various cases including garbage after signature
7698 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7699 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7700 (negative or with leading zeroes).
7702 Further analysis was conducted and fixes were developed by Stephen Henson
7703 of the OpenSSL core team.
7709 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7711 * Session Ticket Memory Leak.
7713 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7714 integrity of that ticket is first verified. In the event of a session
7715 ticket integrity check failing, OpenSSL will fail to free memory
7716 causing a memory leak. By sending a large number of invalid session
7717 tickets an attacker could exploit this issue in a Denial Of Service
7723 * Build option no-ssl3 is incomplete.
7725 When OpenSSL is configured with "no-ssl3" as a build option, servers
7726 could accept and complete a SSL 3.0 handshake, and clients could be
7727 configured to send them.
7730 *Akamai and the OpenSSL team*
7732 * Add support for TLS_FALLBACK_SCSV.
7733 Client applications doing fallback retries should call
7734 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7737 *Adam Langley, Bodo Moeller*
7739 * Add additional DigestInfo checks.
7741 Reencode DigestInto in DER and check against the original when
7742 verifying RSA signature: this will reject any improperly encoded
7743 DigestInfo structures.
7745 Note: this is a precautionary measure and no attacks are currently known.
7749 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7751 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7752 to a denial of service attack. A malicious server can crash the client
7753 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7754 ciphersuite and sending carefully crafted handshake messages.
7756 Thanks to Felix Gröbert (Google) for discovering and researching this
7762 * By sending carefully crafted DTLS packets an attacker could cause openssl
7763 to leak memory. This can be exploited through a Denial of Service attack.
7764 Thanks to Adam Langley for discovering and researching this issue.
7769 * An attacker can force openssl to consume large amounts of memory whilst
7770 processing DTLS handshake messages. This can be exploited through a
7771 Denial of Service attack.
7772 Thanks to Adam Langley for discovering and researching this issue.
7777 * An attacker can force an error condition which causes openssl to crash
7778 whilst processing DTLS packets due to memory being freed twice. This
7779 can be exploited through a Denial of Service attack.
7780 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7786 * If a multithreaded client connects to a malicious server using a resumed
7787 session and the server sends an ec point format extension it could write
7788 up to 255 bytes to freed memory.
7790 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7796 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7797 X509_name_oneline, X509_name_print_ex et al. to leak some information
7798 from the stack. Applications may be affected if they echo pretty printing
7799 output to the attacker.
7801 Thanks to Ivan Fratric (Google) for discovering this issue.
7804 *Emilia Käsper, and Steve Henson*
7806 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7807 for corner cases. (Certain input points at infinity could lead to
7808 bogus results, with non-infinity inputs mapped to infinity too.)
7812 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7814 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7815 handshake can force the use of weak keying material in OpenSSL
7816 SSL/TLS clients and servers.
7818 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7819 researching this issue. ([CVE-2014-0224])
7821 *KIKUCHI Masashi, Steve Henson*
7823 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7824 OpenSSL DTLS client the code can be made to recurse eventually crashing
7827 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7830 *Imre Rad, Steve Henson*
7832 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7833 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7834 client or server. This is potentially exploitable to run arbitrary
7835 code on a vulnerable client or server.
7837 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7839 *Jüri Aedla, Steve Henson*
7841 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7842 are subject to a denial of service attack.
7844 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7845 this issue. ([CVE-2014-3470])
7847 *Felix Gröbert, Ivan Fratric, Steve Henson*
7849 * Harmonize version and its documentation. -f flag is used to display
7852 *mancha <mancha1@zoho.com>*
7854 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7855 in i2d_ECPrivateKey.
7857 *mancha <mancha1@zoho.com>*
7859 * Fix some double frees. These are not thought to be exploitable.
7861 *mancha <mancha1@zoho.com>*
7863 * Fix for the attack described in the paper "Recovering OpenSSL
7864 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7865 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7866 <http://eprint.iacr.org/2014/140>
7868 Thanks to Yuval Yarom and Naomi Benger for discovering this
7869 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7871 *Yuval Yarom and Naomi Benger*
7873 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7875 * Keep original DTLS digest and encryption contexts in retransmission
7876 structures so we can use the previous session parameters if they need
7877 to be resent. ([CVE-2013-6450])
7881 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7882 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7883 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7884 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7885 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7886 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7888 *Rob Stradling, Adam Langley*
7890 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7892 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7894 This addresses the flaw in CBC record processing discovered by
7895 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7896 at: <http://www.isg.rhul.ac.uk/tls/>
7898 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7899 Security Group at Royal Holloway, University of London
7900 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7901 Emilia Käsper for the initial patch.
7904 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7906 * Return an error when checking OCSP signatures when key is NULL.
7907 This fixes a DoS attack. ([CVE-2013-0166])
7911 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7912 the right response is stapled. Also change SSL_get_certificate()
7913 so it returns the certificate actually sent.
7914 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7915 (This is a backport)
7917 *Rob Stradling <rob.stradling@comodo.com>*
7919 * Fix possible deadlock when decoding public keys.
7923 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7925 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7928 * Sanity check record length before skipping explicit IV in DTLS
7931 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7932 fuzzing as a service testing platform.
7937 * Initialise tkeylen properly when encrypting CMS messages.
7938 Thanks to Solar Designer of Openwall for reporting this issue.
7942 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7944 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7945 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7946 in CRYPTO_realloc_clean.
7948 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7949 issue and to Adam Langley <agl@chromium.org> for fixing it.
7952 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7954 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7956 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7957 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7958 content decryption and always return the same error. Note: this attack
7959 needs on average 2^20 messages so it only affects automated senders. The
7960 old behaviour can be re-enabled in the CMS code by setting the
7961 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7962 an MMA defence is not necessary.
7963 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7964 this issue. ([CVE-2012-0884])
7968 * Fix CVE-2011-4619: make sure we really are receiving a
7969 client hello before rejecting multiple SGC restarts. Thanks to
7970 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7974 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7976 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7977 Thanks to Antonio Martin, Enterprise Secure Access Research and
7978 Development, Cisco Systems, Inc. for discovering this bug and
7979 preparing a fix. ([CVE-2012-0050])
7983 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7985 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7986 of the Vaudenay padding oracle attack on CBC mode encryption
7987 which enables an efficient plaintext recovery attack against
7988 the OpenSSL implementation of DTLS. Their attack exploits timing
7989 differences arising during decryption processing. A research
7990 paper describing this attack can be found at:
7991 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7992 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7993 Security Group at Royal Holloway, University of London
7994 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7995 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7996 for preparing the fix. ([CVE-2011-4108])
7998 *Robin Seggelmann, Michael Tuexen*
8000 * Clear bytes used for block padding of SSL 3.0 records.
8003 *Adam Langley (Google)*
8005 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8006 Kadianakis <desnacked@gmail.com> for discovering this issue and
8007 Adam Langley for preparing the fix. ([CVE-2011-4619])
8009 *Adam Langley (Google)*
8011 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8013 *Andrey Kulikov <amdeich@gmail.com>*
8015 * Prevent malformed RFC3779 data triggering an assertion failure.
8016 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8017 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8019 *Rob Austein <sra@hactrn.net>*
8021 * Improved PRNG seeding for VOS.
8023 *Paul Green <Paul.Green@stratus.com>*
8025 * Fix ssl_ciph.c set-up race.
8027 *Adam Langley (Google)*
8029 * Fix spurious failures in ecdsatest.c.
8031 *Emilia Käsper (Google)*
8033 * Fix the BIO_f_buffer() implementation (which was mixing different
8034 interpretations of the `..._len` fields).
8036 *Adam Langley (Google)*
8038 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8039 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8040 threads won't reuse the same blinding coefficients.
8042 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8043 lock to call BN_BLINDING_invert_ex, and avoids one use of
8044 BN_BLINDING_update for each BN_BLINDING structure (previously,
8045 the last update always remained unused).
8047 *Emilia Käsper (Google)*
8049 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8051 *Bob Buckholz (Google)*
8053 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8055 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8056 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8058 *Kaspar Brand <ossl@velox.ch>*
8060 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8061 for multi-threaded use of ECDH. ([CVE-2011-3210])
8063 *Adam Langley (Google)*
8065 * Fix x509_name_ex_d2i memory leak on bad inputs.
8069 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8070 signature public key algorithm by using OID xref utilities instead.
8071 Before this you could only use some ECC ciphersuites with SHA1 only.
8075 * Add protection against ECDSA timing attacks as mentioned in the paper
8076 by Billy Bob Brumley and Nicola Tuveri, see:
8077 <http://eprint.iacr.org/2011/232.pdf>
8079 *Billy Bob Brumley and Nicola Tuveri*
8081 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8083 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8085 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8087 * Fix bug in string printing code: if *any* escaping is enabled we must
8088 escape the escape character (backslash) or the resulting string is
8093 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8095 * Disable code workaround for ancient and obsolete Netscape browsers
8096 and servers: an attacker can use it in a ciphersuite downgrade attack.
8097 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8101 * Fixed J-PAKE implementation error, originally discovered by
8102 Sebastien Martini, further info and confirmation from Stefan
8103 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8107 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8109 * Fix extension code to avoid race conditions which can result in a buffer
8110 overrun vulnerability: resumed sessions must not be modified as they can
8111 be shared by multiple threads. CVE-2010-3864
8115 * Fix WIN32 build system to correctly link an ENGINE directory into
8120 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8122 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8125 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8127 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8129 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8130 context. The operation can be customised via the ctrl mechanism in
8131 case ENGINEs want to include additional functionality.
8135 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8139 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8140 output hashes compatible with older versions of OpenSSL.
8142 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8144 * Fix compression algorithm handling: if resuming a session use the
8145 compression algorithm of the resumed session instead of determining
8146 it from client hello again. Don't allow server to change algorithm.
8150 * Add load_crls() function to commands tidying load_certs() too. Add option
8151 to verify utility to allow additional CRLs to be included.
8155 * Update OCSP request code to permit adding custom headers to the request:
8156 some responders need this.
8160 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8163 *Julia Lawall <julia@diku.dk>*
8165 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8166 needlessly dereferenced structures, used obsolete functions and
8167 didn't handle all updated verify codes correctly.
8171 * Disable MD2 in the default configuration.
8175 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8176 indicate the initial BIO being pushed or popped. This makes it possible
8177 to determine whether the BIO is the one explicitly called or as a result
8178 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8179 it handles reference counts correctly and doesn't zero out the I/O bio
8180 when it is not being explicitly popped. WARNING: applications which
8181 included workarounds for the old buggy behaviour will need to be modified
8182 or they could free up already freed BIOs.
8186 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8187 renaming to all platforms (within the 0.9.8 branch, this was
8188 done conditionally on Netware platforms to avoid a name clash).
8190 *Guenter <lists@gknw.net>*
8192 * Add ECDHE and PSK support to DTLS.
8194 *Michael Tuexen <tuexen@fh-muenster.de>*
8196 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8201 * Add "missing" function EVP_MD_flags() (without this the only way to
8202 retrieve a digest flags is by accessing the structure directly. Update
8203 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8204 or cipher is registered as in the "from" argument. Print out all
8205 registered digests in the dgst usage message instead of manually
8206 attempting to work them out.
8210 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8211 this allows the use of compression and extensions. Change default cipher
8212 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8213 by default unless an application cipher string requests it.
8217 * Alter match criteria in PKCS12_parse(). It used to try to use local
8218 key ids to find matching certificates and keys but some PKCS#12 files
8219 don't follow the (somewhat unwritten) rules and this strategy fails.
8220 Now just gather all certificates together and the first private key
8221 then look for the first certificate that matches the key.
8225 * Support use of registered digest and cipher names for dgst and cipher
8226 commands instead of having to add each one as a special case. So now
8233 openssl dgst -sha256 foo
8235 and this works for ENGINE based algorithms too.
8239 * Update Gost ENGINE to support parameter files.
8241 *Victor B. Wagner <vitus@cryptocom.ru>*
8243 * Support GeneralizedTime in ca utility.
8245 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8247 * Enhance the hash format used for certificate directory links. The new
8248 form uses the canonical encoding (meaning equivalent names will work
8249 even if they aren't identical) and uses SHA1 instead of MD5. This form
8250 is incompatible with the older format and as a result c_rehash should
8251 be used to rebuild symbolic links.
8255 * Make PKCS#8 the default write format for private keys, replacing the
8256 traditional format. This form is standardised, more secure and doesn't
8257 include an implicit MD5 dependency.
8261 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8262 committed to OpenSSL should pass this lot as a minimum.
8266 * Add session ticket override functionality for use by EAP-FAST.
8268 *Jouni Malinen <j@w1.fi>*
8270 * Modify HMAC functions to return a value. Since these can be implemented
8271 in an ENGINE errors can occur.
8275 * Type-checked OBJ_bsearch_ex.
8279 * Type-checked OBJ_bsearch. Also some constification necessitated
8280 by type-checking. Still to come: TXT_DB, bsearch(?),
8281 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8286 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8287 seconds to a tm structure directly, instead of going through OS
8288 specific date routines. This avoids any issues with OS routines such
8289 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8290 and X509_time_adj_ex() to cover the extended range. The existing
8291 X509_time_adj() is still usable and will no longer have any date issues.
8295 * Delta CRL support. New use deltas option which will attempt to locate
8296 and search any appropriate delta CRLs available.
8298 This work was sponsored by Google.
8302 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8303 code and add additional score elements. Validate alternate CRL paths
8304 as part of the CRL checking and indicate a new error "CRL path validation
8305 error" in this case. Applications wanting additional details can use
8306 the verify callback and check the new "parent" field. If this is not
8307 NULL CRL path validation is taking place. Existing applications won't
8308 see this because it requires extended CRL support which is off by
8311 This work was sponsored by Google.
8315 * Support for freshest CRL extension.
8317 This work was sponsored by Google.
8321 * Initial indirect CRL support. Currently only supported in the CRLs
8322 passed directly and not via lookup. Process certificate issuer
8323 CRL entry extension and lookup CRL entries by bother issuer name
8324 and serial number. Check and process CRL issuer entry in IDP extension.
8326 This work was sponsored by Google.
8330 * Add support for distinct certificate and CRL paths. The CRL issuer
8331 certificate is validated separately in this case. Only enabled if
8332 an extended CRL support flag is set: this flag will enable additional
8333 CRL functionality in future.
8335 This work was sponsored by Google.
8339 * Add support for policy mappings extension.
8341 This work was sponsored by Google.
8345 * Fixes to pathlength constraint, self issued certificate handling,
8346 policy processing to align with RFC3280 and PKITS tests.
8348 This work was sponsored by Google.
8352 * Support for name constraints certificate extension. DN, email, DNS
8353 and URI types are currently supported.
8355 This work was sponsored by Google.
8359 * To cater for systems that provide a pointer-based thread ID rather
8360 than numeric, deprecate the current numeric thread ID mechanism and
8361 replace it with a structure and associated callback type. This
8362 mechanism allows a numeric "hash" to be extracted from a thread ID in
8363 either case, and on platforms where pointers are larger than 'long',
8364 mixing is done to help ensure the numeric 'hash' is usable even if it
8365 can't be guaranteed unique. The default mechanism is to use "&errno"
8366 as a pointer-based thread ID to distinguish between threads.
8368 Applications that want to provide their own thread IDs should now use
8369 CRYPTO_THREADID_set_callback() to register a callback that will call
8370 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8372 Note that ERR_remove_state() is now deprecated, because it is tied
8373 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8374 to free the current thread's error state should be replaced by
8375 ERR_remove_thread_state(NULL).
8377 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8378 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8379 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8380 application was previously providing a numeric thread callback that
8381 was inappropriate for distinguishing threads, then uniqueness might
8382 have been obtained with &errno that happened immediately in the
8383 intermediate development versions of OpenSSL; this is no longer the
8384 case, the numeric thread callback will now override the automatic use
8387 *Geoff Thorpe, with help from Bodo Moeller*
8389 * Initial support for different CRL issuing certificates. This covers a
8390 simple case where the self issued certificates in the chain exist and
8391 the real CRL issuer is higher in the existing chain.
8393 This work was sponsored by Google.
8397 * Removed effectively defunct crypto/store from the build.
8401 * Revamp of STACK to provide stronger type-checking. Still to come:
8402 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8403 ASN1_STRING, CONF_VALUE.
8407 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8408 RAM on SSL connections. This option can save about 34k per idle SSL.
8412 * Revamp of LHASH to provide stronger type-checking. Still to come:
8413 STACK, TXT_DB, bsearch, qsort.
8417 * Initial support for Cryptographic Message Syntax (aka CMS) based
8418 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8419 support for data, signedData, compressedData, digestedData and
8420 encryptedData, envelopedData types included. Scripts to check against
8421 RFC4134 examples draft and interop and consistency checks of many
8422 content types and variants.
8426 * Add options to enc utility to support use of zlib compression BIO.
8430 * Extend mk1mf to support importing of options and assembly language
8431 files from Configure script, currently only included in VC-WIN32.
8432 The assembly language rules can now optionally generate the source
8433 files from the associated perl scripts.
8437 * Implement remaining functionality needed to support GOST ciphersuites.
8438 Interop testing has been performed using CryptoPro implementations.
8440 *Victor B. Wagner <vitus@cryptocom.ru>*
8442 * s390x assembler pack.
8446 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8451 * Implement Opaque PRF Input TLS extension as specified in
8452 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8453 official specification yet and no extension type assignment by
8454 IANA exists, this extension (for now) will have to be explicitly
8455 enabled when building OpenSSL by providing the extension number
8456 to use. For example, specify an option
8458 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8460 to the "config" or "Configure" script to enable the extension,
8461 assuming extension number 0x9527 (which is a completely arbitrary
8462 and unofficial assignment based on the MD5 hash of the Internet
8463 Draft). Note that by doing so, you potentially lose
8464 interoperability with other TLS implementations since these might
8465 be using the same extension number for other purposes.
8467 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8468 opaque PRF input value to use in the handshake. This will create
8469 an internal copy of the length-'len' string at 'src', and will
8470 return non-zero for success.
8472 To get more control and flexibility, provide a callback function
8475 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8476 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8480 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8483 Callback function 'cb' will be called in handshakes, and is
8484 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8485 Argument 'arg' is for application purposes (the value as given to
8486 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8487 be provided to the callback function). The callback function
8488 has to return non-zero to report success: usually 1 to use opaque
8489 PRF input just if possible, or 2 to enforce use of the opaque PRF
8490 input. In the latter case, the library will abort the handshake
8491 if opaque PRF input is not successfully negotiated.
8493 Arguments 'peerinput' and 'len' given to the callback function
8494 will always be NULL and 0 in the case of a client. A server will
8495 see the client's opaque PRF input through these variables if
8496 available (NULL and 0 otherwise). Note that if the server
8497 provides an opaque PRF input, the length must be the same as the
8498 length of the client's opaque PRF input.
8500 Note that the callback function will only be called when creating
8501 a new session (session resumption can resume whatever was
8502 previously negotiated), and will not be called in SSL 2.0
8503 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8504 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8505 for applications that need to enforce opaque PRF input.
8509 * Update ssl code to support digests other than SHA1+MD5 for handshake
8512 *Victor B. Wagner <vitus@cryptocom.ru>*
8514 * Add RFC4507 support to OpenSSL. This includes the corrections in
8515 RFC4507bis. The encrypted ticket format is an encrypted encoded
8516 SSL_SESSION structure, that way new session features are automatically
8519 If a client application caches session in an SSL_SESSION structure
8520 support is transparent because tickets are now stored in the encoded
8523 The SSL_CTX structure automatically generates keys for ticket
8524 protection in servers so again support should be possible
8525 with no application modification.
8527 If a client or server wishes to disable RFC4507 support then the option
8528 SSL_OP_NO_TICKET can be set.
8530 Add a TLS extension debugging callback to allow the contents of any client
8531 or server extensions to be examined.
8533 This work was sponsored by Google.
8537 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8538 OpenSSL should now compile cleanly on gcc 4.2
8540 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8542 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8543 support including streaming MAC support: this is required for GOST
8544 ciphersuite support.
8546 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8548 * Add option -stream to use PKCS#7 streaming in smime utility. New
8549 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8550 to output in BER and PEM format.
8554 * Experimental support for use of HMAC via EVP_PKEY interface. This
8555 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8556 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8557 ENGINE support for HMAC keys which are unextractable. New -mac and
8558 -macopt options to dgst utility.
8562 * New option -sigopt to dgst utility. Update dgst to use
8563 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8564 alternative signing parameters such as X9.31 or PSS in the dgst
8569 * Change ssl_cipher_apply_rule(), the internal function that does
8570 the work each time a ciphersuite string requests enabling
8571 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8572 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8573 the order of disabled ciphersuites such that those ciphersuites
8574 that most recently went from enabled to disabled not only stay
8575 in order with respect to each other, but also have higher priority
8576 than other disabled ciphersuites the next time ciphersuites are
8579 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8580 the same ciphersuites as with "HIGH" alone, but in a specific
8581 order where the PSK ciphersuites come first (since they are the
8582 most recently disabled ciphersuites when "HIGH" is parsed).
8584 Also, change ssl_create_cipher_list() (using this new
8585 functionality) such that between otherwise identical
8586 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8591 * Change ssl_create_cipher_list() so that it automatically
8592 arranges the ciphersuites in reasonable order before starting
8593 to process the rule string. Thus, the definition for "DEFAULT"
8594 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8595 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8596 This makes it much easier to arrive at a reasonable default order
8597 in applications for which anonymous ciphers are OK (meaning
8598 that you can't actually use DEFAULT).
8600 *Bodo Moeller; suggested by Victor Duchovni*
8602 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8603 processing) into multiple integers instead of setting
8604 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8605 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8606 (These masks as well as the individual bit definitions are hidden
8607 away into the non-exported interface ssl/ssl_locl.h, so this
8608 change to the definition of the SSL_CIPHER structure shouldn't
8609 affect applications.) This give us more bits for each of these
8610 categories, so there is no longer a need to coagulate AES128 and
8611 AES256 into a single algorithm bit, and to coagulate Camellia128
8612 and Camellia256 into a single algorithm bit, which has led to all
8615 Thus, among other things, the kludge introduced in 0.9.7m and
8616 0.9.8e for masking out AES256 independently of AES128 or masking
8617 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8619 With the change, we also introduce new ciphersuite aliases that
8620 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8625 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8626 Use the leftmost N bytes of the signature input if the input is
8627 larger than the prime q (with N being the size in bytes of q).
8631 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8632 it yet and it is largely untested.
8636 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8640 * Initial incomplete changes to avoid need for function casts in OpenSSL
8641 some compilers (gcc 4.2 and later) reject their use. Safestack is
8642 reimplemented. Update ASN1 to avoid use of legacy functions.
8646 * Win32/64 targets are linked with Winsock2.
8650 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8651 to external functions. This can be used to increase CRL handling
8652 efficiency especially when CRLs are very large by (for example) storing
8653 the CRL revoked certificates in a database.
8657 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8658 new CRLs added to a directory can be used. New command line option
8659 -verify_return_error to s_client and s_server. This causes real errors
8660 to be returned by the verify callback instead of carrying on no matter
8661 what. This reflects the way a "real world" verify callback would behave.
8665 * GOST engine, supporting several GOST algorithms and public key formats.
8666 Kindly donated by Cryptocom.
8670 * Partial support for Issuing Distribution Point CRL extension. CRLs
8671 partitioned by DP are handled but no indirect CRL or reason partitioning
8672 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8673 selected via a scoring technique which handles IDP and AKID in CRLs.
8677 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8678 will ultimately be used for all verify operations: this will remove the
8679 X509_STORE dependency on certificate verification and allow alternative
8680 lookup methods. X509_STORE based implementations of these two callbacks.
8684 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8685 Modify get_crl() to find a valid (unexpired) CRL if possible.
8689 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8690 this would be called X509_CRL_cmp() but that name is already used by
8691 a function that just compares CRL issuer names. Cache several CRL
8692 extensions in X509_CRL structure and cache CRLDP in X509.
8696 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8697 this maps equivalent X509_NAME structures into a consistent structure.
8698 Name comparison can then be performed rapidly using memcmp().
8702 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8707 * Allow digests to supply their own micalg string for S/MIME type using
8708 the ctrl EVP_MD_CTRL_MICALG.
8712 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8713 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8714 ctrl. It can then customise the structure before and/or after signing
8719 * New function OBJ_add_sigid() to allow application defined signature OIDs
8720 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8721 to free up any added signature OIDs.
8725 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8726 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8727 digest and cipher tables. New options added to openssl utility:
8728 list-message-digest-algorithms and list-cipher-algorithms.
8732 * Change the array representation of binary polynomials: the list
8733 of degrees of non-zero coefficients is now terminated with -1.
8734 Previously it was terminated with 0, which was also part of the
8735 value; thus, the array representation was not applicable to
8736 polynomials where t^0 has coefficient zero. This change makes
8737 the array representation useful in a more general context.
8741 * Various modifications and fixes to SSL/TLS cipher string
8742 handling. For ECC, the code now distinguishes between fixed ECDH
8743 with RSA certificates on the one hand and with ECDSA certificates
8744 on the other hand, since these are separate ciphersuites. The
8745 unused code for Fortezza ciphersuites has been removed.
8747 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8748 (not "ECDHE"). For consistency with the code for DH
8749 certificates, use of ECDH certificates is now considered ECDH
8750 authentication, not RSA or ECDSA authentication (the latter is
8751 merely the CA's signing algorithm and not actively used in the
8754 The temporary ciphersuite alias "ECCdraft" is no longer
8755 available, and ECC ciphersuites are no longer excluded from "ALL"
8756 and "DEFAULT". The following aliases now exist for RFC 4492
8757 ciphersuites, most of these by analogy with the DH case:
8759 kECDHr - ECDH cert, signed with RSA
8760 kECDHe - ECDH cert, signed with ECDSA
8761 kECDH - ECDH cert (signed with either RSA or ECDSA)
8762 kEECDH - ephemeral ECDH
8763 ECDH - ECDH cert or ephemeral ECDH
8769 AECDH - anonymous ECDH
8770 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8774 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8775 Use correct micalg parameters depending on digest(s) in signed message.
8779 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8780 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8784 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8785 an engine to register a method. Add ENGINE lookups for methods and
8786 functional reference processing.
8790 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8791 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8796 * New -resign option to smime utility. This adds one or more signers
8797 to an existing PKCS#7 signedData structure. Also -md option to use an
8798 alternative message digest algorithm for signing.
8802 * Tidy up PKCS#7 routines and add new functions to make it easier to
8803 create PKCS7 structures containing multiple signers. Update smime
8804 application to support multiple signers.
8808 * New -macalg option to pkcs12 utility to allow setting of an alternative
8813 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8814 Reorganize PBE internals to lookup from a static table using NIDs,
8815 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8816 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8817 PRF which will be automatically used with PBES2.
8821 * Replace the algorithm specific calls to generate keys in "req" with the
8826 * Update PKCS#7 enveloped data routines to use new API. This is now
8827 supported by any public key method supporting the encrypt operation. A
8828 ctrl is added to allow the public key algorithm to examine or modify
8829 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8834 * Add a ctrl to asn1 method to allow a public key algorithm to express
8835 a default digest type to use. In most cases this will be SHA1 but some
8836 algorithms (such as GOST) need to specify an alternative digest. The
8837 return value indicates how strong the preference is 1 means optional and
8838 2 is mandatory (that is it is the only supported type). Modify
8839 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8840 use the default md. Update openssl utilities to use the default digest
8841 type for signing if it is not explicitly indicated.
8845 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8846 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8847 signing method from the key type. This effectively removes the link
8848 between digests and public key types.
8852 * Add an OID cross reference table and utility functions. Its purpose is to
8853 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8854 rsaEncryption. This will allow some of the algorithm specific hackery
8855 needed to use the correct OID to be removed.
8859 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8860 structures for PKCS7_sign(). They are now set up by the relevant public
8865 * Add provisional EC pkey method with support for ECDSA and ECDH.
8869 * Add support for key derivation (agreement) in the API, DH method and
8874 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8875 public and private key formats. As a side effect these add additional
8876 command line functionality not previously available: DSA signatures can be
8877 generated and verified using pkeyutl and DH key support and generation in
8884 *Oliver Tappe <zooey@hirschkaefer.de>*
8886 * New make target "install_html_docs" installs HTML renditions of the
8889 *Oliver Tappe <zooey@hirschkaefer.de>*
8891 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8892 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8893 support key and parameter generation and add initial key generation
8894 functionality for RSA.
8898 * Add functions for main EVP_PKEY_method operations. The undocumented
8899 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8900 `EVP_PKEY_{encrypt,decrypt}_old`.
8904 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8905 key API, doesn't do much yet.
8909 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8910 public key algorithms. New option to openssl utility:
8911 "list-public-key-algorithms" to print out info.
8915 * Implement the Supported Elliptic Curves Extension for
8916 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8920 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8921 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8925 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8926 utilities such as rsa, dsa, dsaparam etc except they process any key
8931 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8932 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8933 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8938 * Initial support for pluggable public key ASN1.
8939 De-spaghettify the public key ASN1 handling. Move public and private
8940 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8941 algorithm specific handling to a single module within the relevant
8942 algorithm directory. Add functions to allow (near) opaque processing
8943 of public and private key structures.
8947 * Implement the Supported Point Formats Extension for
8948 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8952 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8953 for the psk identity [hint] and the psk callback functions to the
8954 SSL_SESSION, SSL and SSL_CTX structure.
8957 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8961 SSL_CTX_use_psk_identity_hint
8962 SSL_get_psk_identity_hint
8963 SSL_get_psk_identity
8964 SSL_use_psk_identity_hint
8966 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8968 * Add RFC 3161 compliant time stamp request creation, response generation
8969 and response verification functionality.
8971 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8973 * Add initial support for TLS extensions, specifically for the server_name
8974 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8975 have new members for a host name. The SSL data structure has an
8976 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8977 stored in that context to allow for session resumption, even after the
8978 SSL has been switched to a new SSL_CTX in reaction to a client's
8979 server_name extension.
8981 New functions (subject to change):
8983 SSL_get_servername()
8984 SSL_get_servername_type()
8987 New CTRL codes and macros (subject to change):
8989 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8990 - SSL_CTX_set_tlsext_servername_callback()
8991 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8992 - SSL_CTX_set_tlsext_servername_arg()
8993 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8995 openssl s_client has a new '-servername ...' option.
8997 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8998 '-key2 ...', '-servername_fatal' (subject to change). This allows
8999 testing the HostName extension for a specific single host name ('-cert'
9000 and '-key' remain fallbacks for handshakes without HostName
9001 negotiation). If the unrecognized_name alert has to be sent, this by
9002 default is a warning; it becomes fatal with the '-servername_fatal'
9005 *Peter Sylvester, Remy Allais, Christophe Renou*
9007 * Whirlpool hash implementation is added.
9011 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9012 bn(64,32). Because of instruction set limitations it doesn't have
9013 any negative impact on performance. This was done mostly in order
9014 to make it possible to share assembler modules, such as bn_mul_mont
9015 implementations, between 32- and 64-bit builds without hassle.
9019 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9020 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9025 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9026 dedicated Montgomery multiplication procedure, is introduced.
9027 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9028 "64-bit" performance on certain 32-bit targets.
9032 * New option SSL_OP_NO_COMP to disable use of compression selectively
9033 in SSL structures. New SSL ctrl to set maximum send fragment size.
9034 Save memory by setting the I/O buffer sizes dynamically instead of
9035 using the maximum available value.
9039 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9040 in addition to the text details.
9044 * Very, very preliminary EXPERIMENTAL support for printing of general
9045 ASN1 structures. This currently produces rather ugly output and doesn't
9046 handle several customised structures at all.
9050 * Integrated support for PVK file format and some related formats such
9051 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9052 these in the 'rsa' and 'dsa' utilities.
9056 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9060 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9061 place for the (very old) "NETSCAPE" format certificates which are now
9062 handled using new ASN1 code equivalents.
9066 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9067 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9068 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9072 * Modify CRL distribution points extension code to print out previously
9073 unsupported fields. Enhance extension setting code to allow setting of
9078 * Add print and set support for Issuing Distribution Point CRL extension.
9082 * Change 'Configure' script to enable Camellia by default.
9089 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9091 * When rejecting SSL/TLS records due to an incorrect version number, never
9092 update s->server with a new major version number. As of
9093 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9094 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9095 the previous behavior could result in a read attempt at NULL when
9096 receiving specific incorrect SSL/TLS records once record payload
9097 protection is active. ([CVE-2010-0740])
9099 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9101 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9102 could be crashed if the relevant tables were not present (e.g. chrooted).
9104 *Tomas Hoger <thoger@redhat.com>*
9106 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9108 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9110 *Martin Olsson, Neel Mehta*
9112 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9113 accommodate for stack sorting, always a write lock!).
9117 * On some versions of WIN32 Heap32Next is very slow. This can cause
9118 excessive delays in the RAND_poll(): over a minute. As a workaround
9119 include a time check in the inner Heap32Next loop too.
9123 * The code that handled flushing of data in SSL/TLS originally used the
9124 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9125 the problem outlined in PR#1949. The fix suggested there however can
9126 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9127 of Apache). So instead simplify the code to flush unconditionally.
9128 This should be fine since flushing with no data to flush is a no op.
9132 * Handle TLS versions 2.0 and later properly and correctly use the
9133 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9134 off ancient servers have a habit of sticking around for a while...
9138 * Modify compression code so it frees up structures without using the
9139 ex_data callbacks. This works around a problem where some applications
9140 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9141 restarting) then use compression (e.g. SSL with compression) later.
9142 This results in significant per-connection memory leaks and
9143 has caused some security issues including CVE-2008-1678 and
9148 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9149 change when encrypting or decrypting.
9153 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9154 connect and renegotiate with servers which do not support RI.
9155 Until RI is more widely deployed this option is enabled by default.
9159 * Add "missing" ssl ctrls to clear options and mode.
9163 * If client attempts to renegotiate and doesn't support RI respond with
9164 a no_renegotiation alert as required by RFC5746. Some renegotiating
9165 TLS clients will continue a connection gracefully when they receive
9166 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9167 waiting for a server hello which it will never receive. Now we treat a
9168 received no_renegotiation alert as a fatal error. This is because
9169 applications requesting a renegotiation might well expect it to succeed
9170 and would have no code in place to handle the server denying it so the
9171 only safe thing to do is to terminate the connection.
9175 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9176 peer supports secure renegotiation and 0 otherwise. Print out peer
9177 renegotiation support in s_client/s_server.
9181 * Replace the highly broken and deprecated SPKAC certification method with
9182 the updated NID creation version. This should correctly handle UTF8.
9186 * Implement RFC5746. Re-enable renegotiation but require the extension
9187 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9188 turns out to be a bad idea. It has been replaced by
9189 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9190 SSL_CTX_set_options(). This is really not recommended unless you
9191 know what you are doing.
9193 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9195 * Fixes to stateless session resumption handling. Use initial_ctx when
9196 issuing and attempting to decrypt tickets in case it has changed during
9197 servername handling. Use a non-zero length session ID when attempting
9198 stateless session resumption: this makes it possible to determine if
9199 a resumption has occurred immediately after receiving server hello
9200 (several places in OpenSSL subtly assume this) instead of later in
9205 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9206 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9207 fixes for a few places where the return code is not checked
9210 *Julia Lawall <julia@diku.dk>*
9212 * Add --strict-warnings option to Configure script to include devteam
9213 warnings in other configurations.
9217 * Add support for --libdir option and LIBDIR variable in makefiles. This
9218 makes it possible to install openssl libraries in locations which
9219 have names other than "lib", for example "/usr/lib64" which some
9222 *Steve Henson, based on patch from Jeremy Utley*
9224 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9225 X690 8.9.12 and can produce some misleading textual output of OIDs.
9227 *Steve Henson, reported by Dan Kaminsky*
9229 * Delete MD2 from algorithm tables. This follows the recommendation in
9230 several standards that it is not used in new applications due to
9231 several cryptographic weaknesses. For binary compatibility reasons
9232 the MD2 API is still compiled in by default.
9236 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9241 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9242 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9245 *Guenter <lists@gknw.net>*
9247 * Fix the server certificate chain building code to use X509_verify_cert(),
9248 it used to have an ad-hoc builder which was unable to cope with anything
9249 other than a simple chain.
9251 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9253 * Don't check self signed certificate signatures in X509_verify_cert()
9254 by default (a flag can override this): it just wastes time without
9255 adding any security. As a useful side effect self signed root CAs
9256 with non-FIPS digests are now usable in FIPS mode.
9260 * In dtls1_process_out_of_seq_message() the check if the current message
9261 is already buffered was missing. For every new message was memory
9262 allocated, allowing an attacker to perform an denial of service attack
9263 with sending out of seq handshake messages until there is no memory
9264 left. Additionally every future message was buffered, even if the
9265 sequence number made no sense and would be part of another handshake.
9266 So only messages with sequence numbers less than 10 in advance will be
9267 buffered. ([CVE-2009-1378])
9269 *Robin Seggelmann, discovered by Daniel Mentz*
9271 * Records are buffered if they arrive with a future epoch to be
9272 processed after finishing the corresponding handshake. There is
9273 currently no limitation to this buffer allowing an attacker to perform
9274 a DOS attack with sending records with future epochs until there is no
9275 memory left. This patch adds the pqueue_size() function to determine
9276 the size of a buffer and limits the record buffer to 100 entries.
9279 *Robin Seggelmann, discovered by Daniel Mentz*
9281 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9282 parent structure is freed. ([CVE-2009-1379])
9286 * Handle non-blocking I/O properly in SSL_shutdown() call.
9288 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9290 * Add `2.5.4.*` OIDs
9292 *Ilya O. <vrghost@gmail.com>*
9294 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9296 * Disable renegotiation completely - this fixes a severe security
9297 problem ([CVE-2009-3555]) at the cost of breaking all
9298 renegotiation. Renegotiation can be re-enabled by setting
9299 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9300 run-time. This is really not recommended unless you know what
9305 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9307 * Don't set val to NULL when freeing up structures, it is freed up by
9308 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9309 zeroing past the valid field. ([CVE-2009-0789])
9311 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9313 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9314 checked correctly. This would allow some invalid signed attributes to
9315 appear to verify correctly. ([CVE-2009-0591])
9317 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9319 * Reject UniversalString and BMPString types with invalid lengths. This
9320 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9321 a legal length. ([CVE-2009-0590])
9325 * Set S/MIME signing as the default purpose rather than setting it
9326 unconditionally. This allows applications to override it at the store
9331 * Permit restricted recursion of ASN1 strings. This is needed in practice
9332 to handle some structures.
9336 * Improve efficiency of mem_gets: don't search whole buffer each time
9339 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9341 * New -hex option for openssl rand.
9345 * Print out UTF8String and NumericString when parsing ASN1.
9349 * Support NumericString type for name components.
9353 * Allow CC in the environment to override the automatically chosen
9354 compiler. Note that nothing is done to ensure flags work with the
9359 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9361 * Properly check EVP_VerifyFinal() and similar return values
9364 *Ben Laurie, Bodo Moeller, Google Security Team*
9366 * Enable TLS extensions by default.
9370 * Allow the CHIL engine to be loaded, whether the application is
9371 multithreaded or not. (This does not release the developer from the
9372 obligation to set up the dynamic locking callbacks.)
9374 *Sander Temme <sander@temme.net>*
9376 * Use correct exit code if there is an error in dgst command.
9378 *Steve Henson; problem pointed out by Roland Dirlewanger*
9380 * Tweak Configure so that you need to say "experimental-jpake" to enable
9381 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9385 * Add experimental JPAKE support, including demo authentication in
9386 s_client and s_server.
9390 * Set the comparison function in v3_addr_canonize().
9392 *Rob Austein <sra@hactrn.net>*
9394 * Add support for XMPP STARTTLS in s_client.
9396 *Philip Paeps <philip@freebsd.org>*
9398 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9399 to ensure that even with this option, only ciphersuites in the
9400 server's preference list will be accepted. (Note that the option
9401 applies only when resuming a session, so the earlier behavior was
9402 just about the algorithm choice for symmetric cryptography.)
9406 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9408 * Fix NULL pointer dereference if a DTLS server received
9409 ChangeCipherSpec as first record ([CVE-2009-1386]).
9413 * Fix a state transition in s3_srvr.c and d1_srvr.c
9414 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9418 * The fix in 0.9.8c that supposedly got rid of unsafe
9419 double-checked locking was incomplete for RSA blinding,
9420 addressing just one layer of what turns out to have been
9421 doubly unsafe triple-checked locking.
9423 So now fix this for real by retiring the MONT_HELPER macro
9424 in crypto/rsa/rsa_eay.c.
9426 *Bodo Moeller; problem pointed out by Marius Schilder*
9428 * Various precautionary measures:
9430 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9432 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9433 (NB: This would require knowledge of the secret session ticket key
9434 to exploit, in which case you'd be SOL either way.)
9436 - Change bn_nist.c so that it will properly handle input BIGNUMs
9437 outside the expected range.
9439 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9442 *Neel Mehta, Bodo Moeller*
9444 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9445 the load fails. Useful for distros.
9447 *Ben Laurie and the FreeBSD team*
9449 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9453 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9457 * Expand ENGINE to support engine supplied SSL client certificate functions.
9459 This work was sponsored by Logica.
9463 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9464 keystores. Support for SSL/TLS client authentication too.
9465 Not compiled unless enable-capieng specified to Configure.
9467 This work was sponsored by Logica.
9471 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9472 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9473 attribute creation routines such as certificate requests and PKCS#12
9478 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9480 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9481 handshake which could lead to a client crash as found using the
9482 Codenomicon TLS test suite ([CVE-2008-1672])
9484 *Steve Henson, Mark Cox*
9486 * Fix double free in TLS server name extensions which could lead to
9487 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9491 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9493 Clear the error queue to ensure that error entries left from
9494 older function calls do not interfere with the correct operation.
9496 *Lutz Jaenicke, Erik de Castro Lopo*
9498 * Remove root CA certificates of commercial CAs:
9500 The OpenSSL project does not recommend any specific CA and does not
9501 have any policy with respect to including or excluding any CA.
9502 Therefore it does not make any sense to ship an arbitrary selection
9503 of root CA certificates with the OpenSSL software.
9507 * RSA OAEP patches to fix two separate invalid memory reads.
9508 The first one involves inputs when 'lzero' is greater than
9509 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9510 before the beginning of from). The second one involves inputs where
9511 the 'db' section contains nothing but zeroes (there is a one-byte
9512 invalid read after the end of 'db').
9514 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9516 * Partial backport from 0.9.9-dev:
9518 Introduce bn_mul_mont (dedicated Montgomery multiplication
9519 procedure) as a candidate for BIGNUM assembler implementation.
9520 While 0.9.9-dev uses assembler for various architectures, only
9521 x86_64 is available by default here in the 0.9.8 branch, and
9522 32-bit x86 is available through a compile-time setting.
9524 To try the 32-bit x86 assembler implementation, use Configure
9525 option "enable-montasm" (which exists only for this backport).
9527 As "enable-montasm" for 32-bit x86 disclaims code stability
9528 anyway, in this constellation we activate additional code
9529 backported from 0.9.9-dev for further performance improvements,
9530 namely BN_from_montgomery_word. (To enable this otherwise,
9531 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9533 *Andy Polyakov (backport partially by Bodo Moeller)*
9535 * Add TLS session ticket callback. This allows an application to set
9536 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9537 values. This is useful for key rollover for example where several key
9538 sets may exist with different names.
9542 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9543 This was broken until now in 0.9.8 releases, such that the only way
9544 a registered ENGINE could be used (assuming it initialises
9545 successfully on the host) was to explicitly set it as the default
9546 for the relevant algorithms. This is in contradiction with 0.9.7
9547 behaviour and the documentation. With this fix, when an ENGINE is
9548 registered into a given algorithm's table of implementations, the
9549 'uptodate' flag is reset so that auto-discovery will be used next
9550 time a new context for that algorithm attempts to select an
9553 *Ian Lister (tweaked by Geoff Thorpe)*
9555 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9556 implementation in the following ways:
9558 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9561 Lack of BER streaming support means one pass streaming processing is
9562 only supported if data is detached: setting the streaming flag is
9563 ignored for embedded content.
9565 CMS support is disabled by default and must be explicitly enabled
9566 with the enable-cms configuration option.
9570 * Update the GMP engine glue to do direct copies between BIGNUM and
9571 mpz_t when openssl and GMP use the same limb size. Otherwise the
9572 existing "conversion via a text string export" trick is still used.
9574 *Paul Sheer <paulsheer@gmail.com>*
9576 * Zlib compression BIO. This is a filter BIO which compressed and
9577 uncompresses any data passed through it.
9581 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9582 RFC3394 compatible AES key wrapping.
9586 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9587 sets string data without copying. X509_ALGOR_set0() and
9588 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9589 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9590 from an X509_ATTRIBUTE structure optionally checking it occurs only
9591 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9596 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9597 to get the expected BN_FLG_CONSTTIME behavior.
9599 *Bodo Moeller (Google)*
9603 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9604 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9605 - added some more tests to do_tests.pl
9606 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9607 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9608 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9609 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9610 - various changes to netware.pl to enable gcc-cross builds on Win32
9612 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9613 - various changes to fix missing prototype warnings
9614 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9615 - added AES, WHIRLPOOL and CPUID assembler code to build files
9616 - added missing AES assembler make rules to mk1mf.pl
9617 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9619 *Guenter Knauf <eflash@gmx.net>*
9621 * Implement certificate status request TLS extension defined in RFC3546.
9622 A client can set the appropriate parameters and receive the encoded
9623 OCSP response via a callback. A server can query the supplied parameters
9624 and set the encoded OCSP response in the callback. Add simplified examples
9625 to s_client and s_server.
9629 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9632 + Binary incompatibility of ssl_ctx_st structure
9633 + DTLS interoperation with non-compliant servers
9634 + Don't call get_session_cb() without proposed session
9635 + Fix ia64 assembler code
9637 *Andy Polyakov, Steve Henson*
9639 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9641 * DTLS Handshake overhaul. There were longstanding issues with
9642 OpenSSL DTLS implementation, which were making it impossible for
9643 RFC 4347 compliant client to communicate with OpenSSL server.
9644 Unfortunately just fixing these incompatibilities would "cut off"
9645 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9646 server keeps tolerating non RFC compliant syntax. The opposite is
9647 not true, 0.9.8f client can not communicate with earlier server.
9648 This update even addresses CVE-2007-4995.
9652 * Changes to avoid need for function casts in OpenSSL: some compilers
9653 (gcc 4.2 and later) reject their use.
9654 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9657 * Add RFC4507 support to OpenSSL. This includes the corrections in
9658 RFC4507bis. The encrypted ticket format is an encrypted encoded
9659 SSL_SESSION structure, that way new session features are automatically
9662 If a client application caches session in an SSL_SESSION structure
9663 support is transparent because tickets are now stored in the encoded
9666 The SSL_CTX structure automatically generates keys for ticket
9667 protection in servers so again support should be possible
9668 with no application modification.
9670 If a client or server wishes to disable RFC4507 support then the option
9671 SSL_OP_NO_TICKET can be set.
9673 Add a TLS extension debugging callback to allow the contents of any client
9674 or server extensions to be examined.
9676 This work was sponsored by Google.
9680 * Add initial support for TLS extensions, specifically for the server_name
9681 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9682 have new members for a host name. The SSL data structure has an
9683 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9684 stored in that context to allow for session resumption, even after the
9685 SSL has been switched to a new SSL_CTX in reaction to a client's
9686 server_name extension.
9688 New functions (subject to change):
9690 SSL_get_servername()
9691 SSL_get_servername_type()
9694 New CTRL codes and macros (subject to change):
9696 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9697 - SSL_CTX_set_tlsext_servername_callback()
9698 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9699 - SSL_CTX_set_tlsext_servername_arg()
9700 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9702 openssl s_client has a new '-servername ...' option.
9704 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9705 '-key2 ...', '-servername_fatal' (subject to change). This allows
9706 testing the HostName extension for a specific single host name ('-cert'
9707 and '-key' remain fallbacks for handshakes without HostName
9708 negotiation). If the unrecognized_name alert has to be sent, this by
9709 default is a warning; it becomes fatal with the '-servername_fatal'
9712 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9714 * Add AES and SSE2 assembly language support to VC++ build.
9718 * Mitigate attack on final subtraction in Montgomery reduction.
9722 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9723 (which previously caused an internal error).
9727 * Squeeze another 10% out of IGE mode when in != out.
9731 * AES IGE mode speedup.
9733 *Dean Gaudet (Google)*
9735 * Add the Korean symmetric 128-bit cipher SEED (see
9736 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9737 add SEED ciphersuites from RFC 4162:
9739 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9740 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9741 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9742 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9744 To minimize changes between patchlevels in the OpenSSL 0.9.8
9745 series, SEED remains excluded from compilation unless OpenSSL
9746 is configured with 'enable-seed'.
9748 *KISA, Bodo Moeller*
9750 * Mitigate branch prediction attacks, which can be practical if a
9751 single processor is shared, allowing a spy process to extract
9752 information. For detailed background information, see
9753 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9754 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9755 and Necessary Software Countermeasures"). The core of the change
9756 are new versions BN_div_no_branch() and
9757 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9758 respectively, which are slower, but avoid the security-relevant
9759 conditional branches. These are automatically called by BN_div()
9760 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9761 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9762 remove a conditional branch.
9764 BN_FLG_CONSTTIME is the new name for the previous
9765 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9766 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9767 in the exponent causes BN_mod_exp_mont() to use the alternative
9768 implementation in BN_mod_exp_mont_consttime().) The old name
9769 remains as a deprecated alias.
9771 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9772 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9773 constant-time implementations for more than just exponentiation.
9774 Here too the old name is kept as a deprecated alias.
9776 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9777 the BN_BLINDING structure gets an independent copy of the
9778 modulus. This means that the previous `BIGNUM *m` argument to
9779 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9780 essentially becomes `const BIGNUM *m`, although we can't actually
9781 change this in the header file before 0.9.9. It allows
9782 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9783 enable BN_FLG_CONSTTIME.
9785 *Matthew D Wood (Intel Corp)*
9787 * In the SSL/TLS server implementation, be strict about session ID
9788 context matching (which matters if an application uses a single
9789 external cache for different purposes). Previously,
9790 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9791 set. This did ensure strict client verification, but meant that,
9792 with applications using a single external cache for quite
9793 different requirements, clients could circumvent ciphersuite
9794 restrictions for a given session ID context by starting a session
9795 in a different context.
9799 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9800 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9801 authentication-only ciphersuites.
9805 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9806 not complete and could lead to a possible single byte overflow
9807 ([CVE-2007-5135]) [Ben Laurie]
9809 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9811 * Since AES128 and AES256 (and similarly Camellia128 and
9812 Camellia256) share a single mask bit in the logic of
9813 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9814 kludge to work properly if AES128 is available and AES256 isn't
9815 (or if Camellia128 is available and Camellia256 isn't).
9819 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9820 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9821 When a point or a seed is encoded in a BIT STRING, we need to
9822 prevent the removal of trailing zero bits to get the proper DER
9823 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9824 of a NamedBitList, for which trailing 0 bits need to be removed.)
9828 * Have SSL/TLS server implementation tolerate "mismatched" record
9829 protocol version while receiving ClientHello even if the
9830 ClientHello is fragmented. (The server can't insist on the
9831 particular protocol version it has chosen before the ServerHello
9832 message has informed the client about his choice.)
9836 * Add RFC 3779 support.
9838 *Rob Austein for ARIN, Ben Laurie*
9840 * Load error codes if they are not already present instead of using a
9841 static variable. This allows them to be cleanly unloaded and reloaded.
9842 Improve header file function name parsing.
9846 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9847 or CAPABILITY handshake as required by RFCs.
9851 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9853 * Introduce limits to prevent malicious keys being able to
9854 cause a denial of service. ([CVE-2006-2940])
9856 *Steve Henson, Bodo Moeller*
9858 * Fix ASN.1 parsing of certain invalid structures that can result
9859 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9861 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9862 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9864 * Fix SSL client code which could crash if connecting to a
9865 malicious SSLv2 server. ([CVE-2006-4343])
9867 *Tavis Ormandy and Will Drewry, Google Security Team*
9869 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9870 match only those. Before that, "AES256-SHA" would be interpreted
9871 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9872 the same strength classification in 0.9.7h) as we currently only
9873 have a single AES bit in the ciphersuite description bitmap.
9874 That change, however, also applied to ciphersuite strings such as
9875 "RC4-MD5" that intentionally matched multiple ciphersuites --
9876 namely, SSL 2.0 ciphersuites in addition to the more common ones
9877 from SSL 3.0/TLS 1.0.
9879 So we change the selection algorithm again: Naming an explicit
9880 ciphersuite selects this one ciphersuite, and any other similar
9881 ciphersuite (same bitmap) from *other* protocol versions.
9882 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9883 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9885 Since SSL 2.0 does not have any ciphersuites for which the
9886 128/256 bit distinction would be relevant, this works for now.
9887 The proper fix will be to use different bits for AES128 and
9888 AES256, which would have avoided the problems from the beginning;
9889 however, bits are scarce, so we can only do this in a new release
9890 (not just a patchlevel) when we can change the SSL_CIPHER
9891 definition to split the single 'unsigned long mask' bitmap into
9892 multiple values to extend the available space.
9896 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9898 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9899 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9901 * Add AES IGE and biIGE modes.
9905 * Change the Unix randomness entropy gathering to use poll() when
9906 possible instead of select(), since the latter has some
9907 undesirable limitations.
9909 *Darryl Miles via Richard Levitte and Bodo Moeller*
9911 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9912 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9913 cannot be implicitly activated as part of, e.g., the "AES" alias.
9914 However, please upgrade to OpenSSL 0.9.9[-dev] for
9915 non-experimental use of the ECC ciphersuites to get TLS extension
9916 support, which is required for curve and point format negotiation
9917 to avoid potential handshake problems.
9921 * Disable rogue ciphersuites:
9923 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9924 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9925 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9927 The latter two were purportedly from
9928 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9931 Also deactivate the remaining ciphersuites from
9932 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9933 unofficial, and the ID has long expired.
9937 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9938 dual-core machines) and other potential thread-safety issues.
9942 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9943 versions), which is now available for royalty-free use
9944 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9945 Also, add Camellia TLS ciphersuites from RFC 4132.
9947 To minimize changes between patchlevels in the OpenSSL 0.9.8
9948 series, Camellia remains excluded from compilation unless OpenSSL
9949 is configured with 'enable-camellia'.
9953 * Disable the padding bug check when compression is in use. The padding
9954 bug check assumes the first packet is of even length, this is not
9955 necessarily true if compression is enabled and can result in false
9956 positives causing handshake failure. The actual bug test is ancient
9957 code so it is hoped that implementations will either have fixed it by
9958 now or any which still have the bug do not support compression.
9962 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9964 * When applying a cipher rule check to see if string match is an explicit
9965 cipher suite and only match that one cipher suite if it is.
9969 * Link in manifests for VC++ if needed.
9971 *Austin Ziegler <halostatue@gmail.com>*
9973 * Update support for ECC-based TLS ciphersuites according to
9974 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9975 TLS extensions, which are supported starting with the 0.9.9
9976 branch, not in the OpenSSL 0.9.8 branch).
9980 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9981 opaque EVP_CIPHER_CTX handling.
9985 * Fixes and enhancements to zlib compression code. We now only use
9986 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9987 to conform with the standards mentioned here:
9988 <http://www.zlib.net/DLL_FAQ.txt>
9989 Static zlib linking now works on Windows and the new --with-zlib-include
9990 --with-zlib-lib options to Configure can be used to supply the location
9991 of the headers and library. Gracefully handle case where zlib library
9996 * Several fixes and enhancements to the OID generation code. The old code
9997 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9998 handle numbers larger than ULONG_MAX, truncated printing and had a
9999 non standard OBJ_obj2txt() behaviour.
10003 * Add support for building of engines under engine/ as shared libraries
10004 under VC++ build system.
10008 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10009 Hopefully, we will not see any false combination of paths any more.
10013 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10015 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10016 (part of SSL_OP_ALL). This option used to disable the
10017 countermeasure against man-in-the-middle protocol-version
10018 rollback in the SSL 2.0 server implementation, which is a bad
10019 idea. ([CVE-2005-2969])
10021 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10022 for Information Security, National Institute of Advanced Industrial
10023 Science and Technology [AIST], Japan)*
10025 * Add two function to clear and return the verify parameter flags.
10029 * Keep cipherlists sorted in the source instead of sorting them at
10030 runtime, thus removing the need for a lock.
10034 * Avoid some small subgroup attacks in Diffie-Hellman.
10036 *Nick Mathewson and Ben Laurie*
10038 * Add functions for well-known primes.
10042 * Extended Windows CE support.
10044 *Satoshi Nakamura and Andy Polyakov*
10046 * Initialize SSL_METHOD structures at compile time instead of during
10047 runtime, thus removing the need for a lock.
10051 * Make PKCS7_decrypt() work even if no certificate is supplied by
10052 attempting to decrypt each encrypted key in turn. Add support to
10057 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10059 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10062 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10066 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10067 key into the same file any more.
10071 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10075 * Add -utf8 command line and config file option to 'ca'.
10077 *Stefan <stf@udoma.org*
10079 * Removed the macro des_crypt(), as it seems to conflict with some
10080 libraries. Use DES_crypt().
10084 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10085 involves renaming the source and generated shared-libs for
10086 both. The engines will accept the corrected or legacy ids
10087 ('ncipher' and '4758_cca' respectively) when binding. NB,
10088 this only applies when building 'shared'.
10090 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10092 * Add attribute functions to EVP_PKEY structure. Modify
10093 PKCS12_create() to recognize a CSP name attribute and
10094 use it. Make -CSP option work again in pkcs12 utility.
10098 * Add new functionality to the bn blinding code:
10099 - automatic re-creation of the BN_BLINDING parameters after
10100 a fixed number of uses (currently 32)
10101 - add new function for parameter creation
10102 - introduce flags to control the update behaviour of the
10103 BN_BLINDING parameters
10104 - hide BN_BLINDING structure
10105 Add a second BN_BLINDING slot to the RSA structure to improve
10106 performance when a single RSA object is shared among several
10111 * Add support for DTLS.
10113 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10115 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10116 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10120 * Remove buggy and incomplete DH cert support from
10121 ssl/ssl_rsa.c and ssl/s3_both.c
10125 * Use SHA-1 instead of MD5 as the default digest algorithm for
10126 the `apps/openssl` commands.
10130 * Compile clean with "-Wall -Wmissing-prototypes
10131 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10132 DEBUG_SAFESTACK must also be set.
10136 * Change ./Configure so that certain algorithms can be disabled by default.
10137 The new counterpiece to "no-xxx" is "enable-xxx".
10139 The patented RC5 and MDC2 algorithms will now be disabled unless
10140 "enable-rc5" and "enable-mdc2", respectively, are specified.
10142 (IDEA remains enabled despite being patented. This is because IDEA
10143 is frequently required for interoperability, and there is no license
10144 fee for non-commercial use. As before, "no-idea" can be used to
10145 avoid this algorithm.)
10149 * Add processing of proxy certificates (see RFC 3820). This work was
10150 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10151 EGEE (Enabling Grids for E-science in Europe).
10155 * RC4 performance overhaul on modern architectures/implementations, such
10156 as Intel P4, IA-64 and AMD64.
10160 * New utility extract-section.pl. This can be used specify an alternative
10161 section number in a pod file instead of having to treat each file as
10162 a separate case in Makefile. This can be done by adding two lines to the
10165 =for comment openssl_section:XXX
10167 The blank line is mandatory.
10171 * New arguments -certform, -keyform and -pass for s_client and s_server
10172 to allow alternative format key and certificate files and passphrase
10177 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10178 update associated structures and add various utility functions.
10180 Add new policy related verify parameters, include policy checking in
10181 standard verify code. Enhance 'smime' application with extra parameters
10182 to support policy checking and print out.
10186 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10187 Nehemiah processors. These extensions support AES encryption in hardware
10188 as well as RNG (though RNG support is currently disabled).
10190 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10192 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10196 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10198 *Andy Polyakov and a number of other people*
10200 * Improved PowerPC platform support. Most notably BIGNUM assembler
10201 implementation contributed by IBM.
10203 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10205 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10206 exponent rather than 'unsigned long'. There is a corresponding change to
10207 the new 'rsa_keygen' element of the RSA_METHOD structure.
10209 *Jelte Jansen, Geoff Thorpe*
10211 * Functionality for creating the initial serial number file is now
10212 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10214 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10215 number file to 1, which is bound to cause problems. To avoid
10216 the problems while respecting compatibility between different 0.9.7
10217 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10218 CA.pl for serial number initialization. With the new release 0.9.8,
10219 we can fix the problem directly in the 'ca' utility.)
10223 * Reduced header interdependencies by declaring more opaque objects in
10224 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10225 give fewer recursive includes, which could break lazy source code - so
10226 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10227 developers should define this symbol when building and using openssl to
10228 ensure they track the recommended behaviour, interfaces, [etc], but
10229 backwards-compatible behaviour prevails when this isn't defined.
10233 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10237 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10238 This will generate a random key of the appropriate length based on the
10239 cipher context. The EVP_CIPHER can provide its own random key generation
10240 routine to support keys of a specific form. This is used in the des and
10241 3des routines to generate a key of the correct parity. Update S/MIME
10242 code to use new functions and hence generate correct parity DES keys.
10243 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10244 valid (weak or incorrect parity).
10248 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10249 as looking them up. This is useful when the verified structure may contain
10250 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10251 present unless the new PKCS7_NO_CRL flag is asserted.
10255 * Extend ASN1 oid configuration module. It now additionally accepts the
10258 shortName = some long name, 1.2.3.4
10262 * Reimplemented the BN_CTX implementation. There is now no more static
10263 limitation on the number of variables it can handle nor the depth of the
10264 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10265 information can now expand as required, and rather than having a single
10266 static array of bignums, BN_CTX now uses a linked-list of such arrays
10267 allowing it to expand on demand whilst maintaining the usefulness of
10268 BN_CTX's "bundling".
10272 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10273 to allow all RSA operations to function using a single BN_CTX.
10277 * Preliminary support for certificate policy evaluation and checking. This
10278 is initially intended to pass the tests outlined in "Conformance Testing
10279 of Relying Party Client Certificate Path Processing Logic" v1.07.
10283 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10284 remained unused and not that useful. A variety of other little bignum
10285 tweaks and fixes have also been made continuing on from the audit (see
10290 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10291 associated ASN1, EVP and SSL functions and old ASN1 macros.
10295 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10296 and this should never fail. So the return value from the use of
10297 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10298 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10302 * BN_CTX_get() should return zero-valued bignums, providing the same
10303 initialised value as BN_new().
10305 *Geoff Thorpe, suggested by Ulf Möller*
10307 * Support for inhibitAnyPolicy certificate extension.
10311 * An audit of the BIGNUM code is underway, for which debugging code is
10312 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10313 is considered valid when processing BIGNUMs, and causes execution to
10314 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10315 further steps are taken to deliberately pollute unused data in BIGNUM
10316 structures to try and expose faulty code further on. For now, openssl will
10317 (in its default mode of operation) continue to tolerate the inconsistent
10318 forms that it has tolerated in the past, but authors and packagers should
10319 consider trying openssl and their own applications when compiled with
10320 these debugging symbols defined. It will help highlight potential bugs in
10321 their own code, and will improve the test coverage for OpenSSL itself. At
10322 some point, these tighter rules will become openssl's default to improve
10323 maintainability, though the assert()s and other overheads will remain only
10324 in debugging configurations. See bn.h for more details.
10326 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10328 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10329 that can only be obtained through BN_CTX_new() (which implicitly
10330 initialises it). The presence of this function only made it possible
10331 to overwrite an existing structure (and cause memory leaks).
10335 * Because of the callback-based approach for implementing LHASH as a
10336 template type, lh_insert() adds opaque objects to hash-tables and
10337 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10338 to clean up those corresponding objects before destroying the hash table
10339 (and losing the object pointers). So some over-zealous constifications in
10340 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10341 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10342 prototyped to have "const" restrictions on the object pointers they are
10343 given (and so aren't required to cast them away any more).
10347 * The tmdiff.h API was so ugly and minimal that our own timing utility
10348 (speed) prefers to use its own implementation. The two implementations
10349 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10350 its object type properly exposed (MS_TM) instead of casting to/from
10351 `char *`. This may still change yet if someone realises MS_TM and
10353 aren't necessarily the greatest nomenclatures - but this is what was used
10354 internally to the implementation so I've used that for now.
10358 * Ensure that deprecated functions do not get compiled when
10359 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10360 the self-tests were still using deprecated key-generation functions so
10361 these have been updated also.
10365 * Reorganise PKCS#7 code to separate the digest location functionality
10366 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10367 New function PKCS7_set_digest() to set the digest type for PKCS#7
10368 digestedData type. Add additional code to correctly generate the
10369 digestedData type and add support for this type in PKCS7 initialization
10374 * New function PKCS7_set0_type_other() this initializes a PKCS7
10375 structure of type "other".
10379 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10380 sure the loop does correctly stop and breaking ("division by zero")
10381 modulus operations are not performed. The (pre-generated) prime
10382 table crypto/bn/bn_prime.h was already correct, but it could not be
10383 re-generated on some platforms because of the "division by zero"
10384 situation in the script.
10386 *Ralf S. Engelschall*
10388 * Update support for ECC-based TLS ciphersuites according to
10389 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10390 SHA-1 now is only used for "small" curves (where the
10391 representation of a field element takes up to 24 bytes); for
10392 larger curves, the field element resulting from ECDH is directly
10393 used as premaster secret.
10395 *Douglas Stebila (Sun Microsystems Laboratories)*
10397 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10398 curve secp160r1 to the tests.
10400 *Douglas Stebila (Sun Microsystems Laboratories)*
10402 * Add the possibility to load symbols globally with DSO.
10404 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10406 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10407 control of the error stack.
10411 * Add support for STORE in ENGINE.
10415 * Add the STORE type. The intention is to provide a common interface
10416 to certificate and key stores, be they simple file-based stores, or
10417 HSM-type store, or LDAP stores, or...
10418 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10422 * Add a generic structure called OPENSSL_ITEM. This can be used to
10423 pass a list of arguments to any function as well as provide a way
10424 for a function to pass data back to the caller.
10428 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10429 works like BUF_strdup() but can be used to duplicate a portion of
10430 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10435 * Add the function sk_find_ex() which works like sk_find(), but will
10436 return an index to an element even if an exact match couldn't be
10437 found. The index is guaranteed to point at the element where the
10438 searched-for key would be inserted to preserve sorting order.
10442 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10443 takes an extra flags argument for optional functionality. Currently,
10444 the following flags are defined:
10446 OBJ_BSEARCH_VALUE_ON_NOMATCH
10447 This one gets OBJ_bsearch_ex() to return a pointer to the first
10448 element where the comparing function returns a negative or zero
10451 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10452 This one gets OBJ_bsearch_ex() to return a pointer to the first
10453 element where the comparing function returns zero. This is useful
10454 if there are more than one element where the comparing function
10459 * Make it possible to create self-signed certificates with 'openssl ca'
10460 in such a way that the self-signed certificate becomes part of the
10461 CA database and uses the same mechanisms for serial number generation
10462 as all other certificate signing. The new flag '-selfsign' enables
10463 this functionality. Adapt CA.sh and CA.pl.in.
10467 * Add functionality to check the public key of a certificate request
10468 against a given private. This is useful to check that a certificate
10469 request can be signed by that key (self-signing).
10473 * Make it possible to have multiple active certificates with the same
10474 subject in the CA index file. This is done only if the keyword
10475 'unique_subject' is set to 'no' in the main CA section (default
10476 if 'CA_default') of the configuration file. The value is saved
10477 with the database itself in a separate index attribute file,
10478 named like the index file with '.attr' appended to the name.
10482 * Generate multi-valued AVAs using '+' notation in config files for
10487 * Support for nameConstraints certificate extension.
10491 * Support for policyConstraints certificate extension.
10495 * Support for policyMappings certificate extension.
10499 * Make sure the default DSA_METHOD implementation only uses its
10500 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10501 and change its own handlers to be NULL so as to remove unnecessary
10502 indirection. This lets alternative implementations fallback to the
10503 default implementation more easily.
10507 * Support for directoryName in GeneralName related extensions
10512 * Make it possible to link applications using Makefile.shared.
10513 Make that possible even when linking against static libraries!
10517 * Support for single pass processing for S/MIME signing. This now
10518 means that S/MIME signing can be done from a pipe, in addition
10519 cleartext signing (multipart/signed type) is effectively streaming
10520 and the signed data does not need to be all held in memory.
10522 This is done with a new flag PKCS7_STREAM. When this flag is set
10523 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10524 is done after the data is output (and digests calculated) in
10525 SMIME_write_PKCS7().
10529 * Add full support for -rpath/-R, both in shared libraries and
10530 applications, at least on the platforms where it's known how
10535 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10536 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10537 will now compute a table of multiples of the generator that
10538 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10539 faster (notably in the case of a single point multiplication,
10540 scalar * generator).
10542 *Nils Larsch, Bodo Moeller*
10544 * IPv6 support for certificate extensions. The various extensions
10545 which use the IP:a.b.c.d can now take IPv6 addresses using the
10546 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10551 * Added an ENGINE that implements RSA by performing private key
10552 exponentiations with the GMP library. The conversions to and from
10553 GMP's mpz_t format aren't optimised nor are any montgomery forms
10554 cached, and on x86 it appears OpenSSL's own performance has caught up.
10555 However there are likely to be other architectures where GMP could
10556 provide a boost. This ENGINE is not built in by default, but it can be
10557 specified at Configure time and should be accompanied by the necessary
10558 linker additions, eg;
10559 ./config -DOPENSSL_USE_GMP -lgmp
10563 * "openssl engine" will not display ENGINE/DSO load failure errors when
10564 testing availability of engines with "-t" - the old behaviour is
10565 produced by increasing the feature's verbosity with "-tt".
10569 * ECDSA routines: under certain error conditions uninitialized BN objects
10570 could be freed. Solution: make sure initialization is performed early
10571 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10576 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10577 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10578 software implementations. For DSA and DH, parameter generation can
10579 also be overridden by providing the appropriate method callbacks.
10583 * Change the "progress" mechanism used in key-generation and
10584 primality testing to functions that take a new BN_GENCB pointer in
10585 place of callback/argument pairs. The new API functions have `_ex`
10586 postfixes and the older functions are reimplemented as wrappers for
10587 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10588 declarations of the old functions to help (graceful) attempts to
10589 migrate to the new functions. Also, the new key-generation API
10590 functions operate on a caller-supplied key-structure and return
10591 success/failure rather than returning a key or NULL - this is to
10592 help make "keygen" another member function of RSA_METHOD etc.
10594 Example for using the new callback interface:
10596 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10597 void *my_arg = ...;
10600 BN_GENCB_set(&my_cb, my_callback, my_arg);
10602 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10603 /* For the meaning of a, b in calls to my_callback(), see the
10604 * documentation of the function that calls the callback.
10605 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10606 * my_callback should return 1 if it wants BN_is_prime_ex()
10607 * to continue, or 0 to stop.
10612 * Change the ZLIB compression method to be stateful, and make it
10613 available to TLS with the number defined in
10614 draft-ietf-tls-compression-04.txt.
10618 * Add the ASN.1 structures and functions for CertificatePair, which
10619 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10621 CertificatePair ::= SEQUENCE {
10622 forward [0] Certificate OPTIONAL,
10623 reverse [1] Certificate OPTIONAL,
10624 -- at least one of the pair shall be present -- }
10626 Also implement the PEM functions to read and write certificate
10627 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10629 This needed to be defined, mostly for the sake of the LDAP
10630 attribute crossCertificatePair, but may prove useful elsewhere as
10635 * Make it possible to inhibit symlinking of shared libraries in
10636 Makefile.shared, for Cygwin's sake.
10640 * Extend the BIGNUM API by creating a function
10641 void BN_set_negative(BIGNUM *a, int neg);
10642 and a macro that behave like
10643 int BN_is_negative(const BIGNUM *a);
10645 to avoid the need to access 'a->neg' directly in applications.
10649 * Implement fast modular reduction for pseudo-Mersenne primes
10650 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10651 EC_GROUP_new_curve_GFp() will now automatically use this
10654 *Nils Larsch <nla@trustcenter.de>*
10656 * Add new lock type (CRYPTO_LOCK_BN).
10660 * Change the ENGINE framework to automatically load engines
10661 dynamically from specific directories unless they could be
10662 found to already be built in or loaded. Move all the
10663 current engines except for the cryptodev one to a new
10664 directory engines/.
10665 The engines in engines/ are built as shared libraries if
10666 the "shared" options was given to ./Configure or ./config.
10667 Otherwise, they are inserted in libcrypto.a.
10668 /usr/local/ssl/engines is the default directory for dynamic
10669 engines, but that can be overridden at configure time through
10670 the usual use of --prefix and/or --openssldir, and at run
10671 time with the environment variable OPENSSL_ENGINES.
10673 *Geoff Thorpe and Richard Levitte*
10675 * Add Makefile.shared, a helper makefile to build shared
10676 libraries. Adapt Makefile.org.
10680 * Add version info to Win32 DLLs.
10682 *Peter 'Luna' Runestig" <peter@runestig.com>*
10684 * Add new 'medium level' PKCS#12 API. Certificates and keys
10685 can be added using this API to created arbitrary PKCS#12
10686 files while avoiding the low level API.
10688 New options to PKCS12_create(), key or cert can be NULL and
10689 will then be omitted from the output file. The encryption
10690 algorithm NIDs can be set to -1 for no encryption, the mac
10691 iteration count can be set to 0 to omit the mac.
10693 Enhance pkcs12 utility by making the -nokeys and -nocerts
10694 options work when creating a PKCS#12 file. New option -nomac
10695 to omit the mac, NONE can be set for an encryption algorithm.
10696 New code is modified to use the enhanced PKCS12_create()
10697 instead of the low level API.
10701 * Extend ASN1 encoder to support indefinite length constructed
10702 encoding. This can output sequences tags and octet strings in
10703 this form. Modify pk7_asn1.c to support indefinite length
10704 encoding. This is experimental and needs additional code to
10705 be useful, such as an ASN1 bio and some enhanced streaming
10708 Extend template encode functionality so that tagging is passed
10709 down to the template encoder.
10713 * Let 'openssl req' fail if an argument to '-newkey' is not
10714 recognized instead of using RSA as a default.
10718 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10719 As these are not official, they are not included in "ALL";
10720 the "ECCdraft" ciphersuite group alias can be used to select them.
10722 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10724 * Add ECDH engine support.
10726 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10728 * Add ECDH in new directory crypto/ecdh/.
10730 *Douglas Stebila (Sun Microsystems Laboratories)*
10732 * Let BN_rand_range() abort with an error after 100 iterations
10733 without success (which indicates a broken PRNG).
10737 * Change BN_mod_sqrt() so that it verifies that the input value
10738 is really the square of the return value. (Previously,
10739 BN_mod_sqrt would show GIGO behaviour.)
10743 * Add named elliptic curves over binary fields from X9.62, SECG,
10744 and WAP/WTLS; add OIDs that were still missing.
10746 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10748 * Extend the EC library for elliptic curves over binary fields
10749 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10752 EC_GF2m_simple_method
10756 EC_GROUP_new_curve_GF2m
10757 EC_GROUP_set_curve_GF2m
10758 EC_GROUP_get_curve_GF2m
10759 EC_POINT_set_affine_coordinates_GF2m
10760 EC_POINT_get_affine_coordinates_GF2m
10761 EC_POINT_set_compressed_coordinates_GF2m
10763 Point compression for binary fields is disabled by default for
10764 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10767 As binary polynomials are represented as BIGNUMs, various members
10768 of the EC_GROUP and EC_POINT data structures can be shared
10769 between the implementations for prime fields and binary fields;
10770 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10771 are essentially identical to their `..._GFp` counterparts.
10772 (For simplicity, the `..._GFp` prefix has been dropped from
10773 various internal method names.)
10775 An internal 'field_div' method (similar to 'field_mul' and
10776 'field_sqr') has been added; this is used only for binary fields.
10778 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10780 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10781 through methods ('mul', 'precompute_mult').
10783 The generic implementations (now internally called 'ec_wNAF_mul'
10784 and 'ec_wNAF_precomputed_mult') remain the default if these
10785 methods are undefined.
10787 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10789 * New function EC_GROUP_get_degree, which is defined through
10790 EC_METHOD. For curves over prime fields, this returns the bit
10791 length of the modulus.
10793 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10795 * New functions EC_GROUP_dup, EC_POINT_dup.
10796 (These simply call ..._new and ..._copy).
10798 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10800 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10801 Polynomials are represented as BIGNUMs (where the sign bit is not
10802 used) in the following functions [macros]:
10805 BN_GF2m_sub [= BN_GF2m_add]
10806 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10807 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10808 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10810 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10811 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10812 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10813 BN_GF2m_cmp [= BN_ucmp]
10815 (Note that only the 'mod' functions are actually for fields GF(2^m).
10816 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10818 For some functions, an the irreducible polynomial defining a
10819 field can be given as an 'unsigned int[]' with strictly
10820 decreasing elements giving the indices of those bits that are set;
10821 i.e., p[] represents the polynomial
10822 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10824 p[0] > p[1] > ... > p[k] = 0.
10825 This applies to the following functions:
10828 BN_GF2m_mod_mul_arr
10829 BN_GF2m_mod_sqr_arr
10830 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10831 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10832 BN_GF2m_mod_exp_arr
10833 BN_GF2m_mod_sqrt_arr
10834 BN_GF2m_mod_solve_quad_arr
10838 Conversion can be performed by the following functions:
10843 bntest.c has additional tests for binary polynomial arithmetic.
10845 Two implementations for BN_GF2m_mod_div() are available.
10846 The default algorithm simply uses BN_GF2m_mod_inv() and
10847 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10848 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10849 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10851 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10853 * Add new error code 'ERR_R_DISABLED' that can be used when some
10854 functionality is disabled at compile-time.
10856 *Douglas Stebila <douglas.stebila@sun.com>*
10858 * Change default behaviour of 'openssl asn1parse' so that more
10859 information is visible when viewing, e.g., a certificate:
10861 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10862 mode the content of non-printable OCTET STRINGs is output in a
10863 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10864 avoid the appearance of a printable string.
10866 *Nils Larsch <nla@trustcenter.de>*
10868 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10870 EC_GROUP_set_asn1_flag()
10871 EC_GROUP_get_asn1_flag()
10872 EC_GROUP_set_point_conversion_form()
10873 EC_GROUP_get_point_conversion_form()
10874 These control ASN1 encoding details:
10875 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10876 has been set to OPENSSL_EC_NAMED_CURVE.
10877 - Points are encoded in uncompressed form by default; options for
10878 asn1_for are as for point2oct, namely
10879 POINT_CONVERSION_COMPRESSED
10880 POINT_CONVERSION_UNCOMPRESSED
10881 POINT_CONVERSION_HYBRID
10883 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10885 EC_GROUP_set_seed()
10886 EC_GROUP_get0_seed()
10887 EC_GROUP_get_seed_len()
10888 This is used only for ASN1 purposes (so far).
10890 *Nils Larsch <nla@trustcenter.de>*
10892 * Add 'field_type' member to EC_METHOD, which holds the NID
10893 of the appropriate field type OID. The new function
10894 EC_METHOD_get_field_type() returns this value.
10896 *Nils Larsch <nla@trustcenter.de>*
10899 EC_POINT_point2bn()
10900 EC_POINT_bn2point()
10901 EC_POINT_point2hex()
10902 EC_POINT_hex2point()
10903 providing useful interfaces to EC_POINT_point2oct() and
10904 EC_POINT_oct2point().
10906 *Nils Larsch <nla@trustcenter.de>*
10908 * Change internals of the EC library so that the functions
10909 EC_GROUP_set_generator()
10910 EC_GROUP_get_generator()
10911 EC_GROUP_get_order()
10912 EC_GROUP_get_cofactor()
10913 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10914 to methods, which would lead to unnecessary code duplication when
10915 adding different types of curves.
10917 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10919 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10920 arithmetic, and such that modified wNAFs are generated
10921 (which avoid length expansion in many cases).
10925 * Add a function EC_GROUP_check_discriminant() (defined via
10926 EC_METHOD) that verifies that the curve discriminant is non-zero.
10928 Add a function EC_GROUP_check() that makes some sanity tests
10929 on a EC_GROUP, its generator and order. This includes
10930 EC_GROUP_check_discriminant().
10932 *Nils Larsch <nla@trustcenter.de>*
10934 * Add ECDSA in new directory crypto/ecdsa/.
10936 Add applications 'openssl ecparam' and 'openssl ecdsa'
10937 (these are based on 'openssl dsaparam' and 'openssl dsa').
10939 ECDSA support is also included in various other files across the
10940 library. Most notably,
10941 - 'openssl req' now has a '-newkey ecdsa:file' option;
10942 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10943 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10944 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10945 them suitable for ECDSA where domain parameters must be
10946 extracted before the specific public key;
10947 - ECDSA engine support has been added.
10949 *Nils Larsch <nla@trustcenter.de>*
10951 * Include some named elliptic curves, and add OIDs from X9.62,
10952 SECG, and WAP/WTLS. Each curve can be obtained from the new
10954 EC_GROUP_new_by_curve_name(),
10955 and the list of available named curves can be obtained with
10956 EC_get_builtin_curves().
10957 Also add a 'curve_name' member to EC_GROUP objects, which can be
10959 EC_GROUP_set_curve_name()
10960 EC_GROUP_get_curve_name()
10962 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10964 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10965 was actually never needed) and in BN_mul(). The removal in BN_mul()
10966 required a small change in bn_mul_part_recursive() and the addition
10967 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10968 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10969 bn_sub_words() and bn_add_words() except they take arrays with
10974 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10976 * Cleanse PEM buffers before freeing them since they may contain
10979 *Benjamin Bennett <ben@psc.edu>*
10981 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10982 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10983 authentication-only ciphersuites.
10987 * Since AES128 and AES256 share a single mask bit in the logic of
10988 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10989 kludge to work properly if AES128 is available and AES256 isn't.
10993 * Expand security boundary to match 1.1.1 module.
10997 * Remove redundant features: hash file source, editing of test vectors
10998 modify fipsld to use external fips_premain.c signature.
11002 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11003 run algorithm test programs.
11007 * Make algorithm test programs more tolerant of whitespace.
11011 * Have SSL/TLS server implementation tolerate "mismatched" record
11012 protocol version while receiving ClientHello even if the
11013 ClientHello is fragmented. (The server can't insist on the
11014 particular protocol version it has chosen before the ServerHello
11015 message has informed the client about his choice.)
11019 * Load error codes if they are not already present instead of using a
11020 static variable. This allows them to be cleanly unloaded and reloaded.
11024 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11026 * Introduce limits to prevent malicious keys being able to
11027 cause a denial of service. ([CVE-2006-2940])
11029 *Steve Henson, Bodo Moeller*
11031 * Fix ASN.1 parsing of certain invalid structures that can result
11032 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11034 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11035 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11037 * Fix SSL client code which could crash if connecting to a
11038 malicious SSLv2 server. ([CVE-2006-4343])
11040 *Tavis Ormandy and Will Drewry, Google Security Team*
11042 * Change ciphersuite string processing so that an explicit
11043 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11044 will no longer include "AES128-SHA"), and any other similar
11045 ciphersuite (same bitmap) from *other* protocol versions (so that
11046 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11047 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11048 changes from 0.9.8b and 0.9.8d.
11052 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11054 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11055 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11057 * Change the Unix randomness entropy gathering to use poll() when
11058 possible instead of select(), since the latter has some
11059 undesirable limitations.
11061 *Darryl Miles via Richard Levitte and Bodo Moeller*
11063 * Disable rogue ciphersuites:
11065 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11066 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11067 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11069 The latter two were purportedly from
11070 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11073 Also deactivate the remaining ciphersuites from
11074 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11075 unofficial, and the ID has long expired.
11079 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11080 dual-core machines) and other potential thread-safety issues.
11084 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11086 * Adapt fipsld and the build system to link against the validated FIPS
11087 module in FIPS mode.
11091 * Fixes for VC++ 2005 build under Windows.
11095 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11096 from a Windows bash shell such as MSYS. It is autodetected from the
11097 "config" script when run from a VC++ environment. Modify standard VC++
11098 build to use fipscanister.o from the GNU make build.
11102 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11104 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11105 The value now differs depending on if you build for FIPS or not.
11106 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11107 safely run with a non-FIPSed libcrypto, as it may crash because of
11108 the difference induced by this change.
11112 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11114 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11115 (part of SSL_OP_ALL). This option used to disable the
11116 countermeasure against man-in-the-middle protocol-version
11117 rollback in the SSL 2.0 server implementation, which is a bad
11118 idea. ([CVE-2005-2969])
11120 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11121 for Information Security, National Institute of Advanced Industrial
11122 Science and Technology [AIST, Japan)]*
11124 * Minimal support for X9.31 signatures and PSS padding modes. This is
11125 mainly for FIPS compliance and not fully integrated at this stage.
11129 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11130 the exponentiation using a fixed-length exponent. (Otherwise,
11131 the information leaked through timing could expose the secret key
11132 after many signatures; cf. Bleichenbacher's attack on DSA with
11137 * Make a new fixed-window mod_exp implementation the default for
11138 RSA, DSA, and DH private-key operations so that the sequence of
11139 squares and multiplies and the memory access pattern are
11140 independent of the particular secret key. This will mitigate
11141 cache-timing and potential related attacks.
11143 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11144 and this is automatically used by BN_mod_exp_mont() if the new flag
11145 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11146 will use this BN flag for private exponents unless the flag
11147 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11148 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11150 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11152 * Change the client implementation for SSLv23_method() and
11153 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11154 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11155 (Previously, the SSL 2.0 backwards compatible Client Hello
11156 message format would be used even with SSL_OP_NO_SSLv2.)
11160 * Add support for smime-type MIME parameter in S/MIME messages which some
11165 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11166 a threadsafe manner. Modify rsa code to use new function and add calls
11167 to dsa and dh code (which had race conditions before).
11171 * Include the fixed error library code in the C error file definitions
11172 instead of fixing them up at runtime. This keeps the error code
11173 structures constant.
11177 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11179 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11182 * Fixes for newer kerberos headers. NB: the casts are needed because
11183 the 'length' field is signed on one version and unsigned on another
11184 with no (?) obvious way to tell the difference, without these VC++
11185 complains. Also the "definition" of FAR (blank) is no longer included
11186 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11187 some needed definitions.
11191 * Undo Cygwin change.
11195 * Added support for proxy certificates according to RFC 3820.
11196 Because they may be a security thread to unaware applications,
11197 they must be explicitly allowed in run-time. See
11198 docs/HOWTO/proxy_certificates.txt for further information.
11202 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11204 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11205 server and client random values. Previously
11206 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11207 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11209 This change has negligible security impact because:
11211 1. Server and client random values still have 24 bytes of pseudo random
11214 2. Server and client random values are sent in the clear in the initial
11217 3. The master secret is derived using the premaster secret (48 bytes in
11218 size for static RSA ciphersuites) as well as client server and random
11221 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11224 *Stephen Henson, reported by UK NISCC*
11226 * Use Windows randomness collection on Cygwin.
11230 * Fix hang in EGD/PRNGD query when communication socket is closed
11231 prematurely by EGD/PRNGD.
11233 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11235 * Prompt for pass phrases when appropriate for PKCS12 input format.
11239 * Back-port of selected performance improvements from development
11240 branch, as well as improved support for PowerPC platforms.
11244 * Add lots of checks for memory allocation failure, error codes to indicate
11245 failure and freeing up memory if a failure occurs.
11247 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11249 * Add new -passin argument to dgst.
11253 * Perform some character comparisons of different types in X509_NAME_cmp:
11254 this is needed for some certificates that re-encode DNs into UTF8Strings
11255 (in violation of RFC3280) and can't or won't issue name rollover
11260 * Make an explicit check during certificate validation to see that
11261 the CA setting in each certificate on the chain is correct. As a
11262 side effect always do the following basic checks on extensions,
11263 not just when there's an associated purpose to the check:
11265 - if there is an unhandled critical extension (unless the user
11266 has chosen to ignore this fault)
11267 - if the path length has been exceeded (if one is set at all)
11268 - that certain extensions fit the associated purpose (if one has
11273 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11275 * Avoid a race condition when CRLs are checked in a multi threaded
11276 environment. This would happen due to the reordering of the revoked
11277 entries during signature checking and serial number lookup. Now the
11278 encoding is cached and the serial number sort performed under a lock.
11279 Add new STACK function sk_is_sorted().
11283 * Add Delta CRL to the extension code.
11287 * Various fixes to s3_pkt.c so alerts are sent properly.
11289 *David Holmes <d.holmes@f5.com>*
11291 * Reduce the chances of duplicate issuer name and serial numbers (in
11292 violation of RFC3280) using the OpenSSL certificate creation utilities.
11293 This is done by creating a random 64 bit value for the initial serial
11294 number when a serial number file is created or when a self signed
11295 certificate is created using 'openssl req -x509'. The initial serial
11296 number file is created using 'openssl x509 -next_serial' in CA.pl
11297 rather than being initialized to 1.
11301 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11303 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11304 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11306 *Joe Orton, Steve Henson*
11308 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11311 *Joe Orton, Steve Henson*
11313 * Make it possible to have multiple active certificates with the same
11314 subject in the CA index file. This is done only if the keyword
11315 'unique_subject' is set to 'no' in the main CA section (default
11316 if 'CA_default') of the configuration file. The value is saved
11317 with the database itself in a separate index attribute file,
11318 named like the index file with '.attr' appended to the name.
11322 * X509 verify fixes. Disable broken certificate workarounds when
11323 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11324 keyUsage extension present. Don't accept CRLs with unhandled critical
11325 extensions: since verify currently doesn't process CRL extensions this
11326 rejects a CRL with *any* critical extensions. Add new verify error codes
11331 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11332 A clarification of RFC2560 will require the use of OCTET STRINGs and
11333 some implementations cannot handle the current raw format. Since OpenSSL
11334 copies and compares OCSP nonces as opaque blobs without any attempt at
11335 parsing them this should not create any compatibility issues.
11339 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11340 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11341 this HMAC (and other) operations are several times slower than OpenSSL
11346 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11348 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11350 * Use the correct content when signing type "other".
11354 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11356 * Fix various bugs revealed by running the NISCC test suite:
11358 Stop out of bounds reads in the ASN1 code when presented with
11359 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11361 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11363 If verify callback ignores invalid public key errors don't try to check
11364 certificate signature with the NULL public key.
11368 * New -ignore_err option in ocsp application to stop the server
11369 exiting on the first error in a request.
11373 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11374 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11379 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11380 extra data after the compression methods not only for TLS 1.0
11381 but also for SSL 3.0 (as required by the specification).
11383 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11385 * Change X509_certificate_type() to mark the key as exported/exportable
11386 when it's 512 *bits* long, not 512 bytes.
11390 * Change AES_cbc_encrypt() so it outputs exact multiple of
11391 blocks during encryption.
11395 * Various fixes to base64 BIO and non blocking I/O. On write
11396 flushes were not handled properly if the BIO retried. On read
11397 data was not being buffered properly and had various logic bugs.
11398 This also affects blocking I/O when the data being decoded is a
11403 * Various S/MIME bugfixes and compatibility changes:
11404 output correct application/pkcs7 MIME type if
11405 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11406 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11407 of files as .eml work). Correctly handle very long lines in MIME
11412 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11414 * Countermeasure against the Klima-Pokorny-Rosa extension of
11415 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11416 a protocol version number mismatch like a decryption error
11417 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11421 * Turn on RSA blinding by default in the default implementation
11422 to avoid a timing attack. Applications that don't want it can call
11423 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11424 They would be ill-advised to do so in most cases.
11426 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11428 * Change RSA blinding code so that it works when the PRNG is not
11429 seeded (in this case, the secret RSA exponent is abused as
11430 an unpredictable seed -- if it is not unpredictable, there
11431 is no point in blinding anyway). Make RSA blinding thread-safe
11432 by remembering the creator's thread ID in rsa->blinding and
11433 having all other threads use local one-time blinding factors
11434 (this requires more computation than sharing rsa->blinding, but
11435 avoids excessive locking; and if an RSA object is not shared
11436 between threads, blinding will still be very fast).
11440 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11441 ENGINE as defaults for all supported algorithms irrespective of
11442 the 'flags' parameter. 'flags' is now honoured, so applications
11443 should make sure they are passing it correctly.
11447 * Target "mingw" now allows native Windows code to be generated in
11448 the Cygwin environment as well as with the MinGW compiler.
11452 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11454 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11455 via timing by performing a MAC computation even if incorrect
11456 block cipher padding has been found. This is a countermeasure
11457 against active attacks where the attacker has to distinguish
11458 between bad padding and a MAC verification error. ([CVE-2003-0078])
11460 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11461 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11462 Martin Vuagnoux (EPFL, Ilion)*
11464 * Make the no-err option work as intended. The intention with no-err
11465 is not to have the whole error stack handling routines removed from
11466 libcrypto, it's only intended to remove all the function name and
11467 reason texts, thereby removing some of the footprint that may not
11468 be interesting if those errors aren't displayed anyway.
11470 NOTE: it's still possible for any application or module to have its
11471 own set of error texts inserted. The routines are there, just not
11472 used by default when no-err is given.
11476 * Add support for FreeBSD on IA64.
11478 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11480 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11481 Kerberos function mit_des_cbc_cksum(). Before this change,
11482 the value returned by DES_cbc_cksum() was like the one from
11483 mit_des_cbc_cksum(), except the bytes were swapped.
11485 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11487 * Allow an application to disable the automatic SSL chain building.
11488 Before this a rather primitive chain build was always performed in
11489 ssl3_output_cert_chain(): an application had no way to send the
11490 correct chain if the automatic operation produced an incorrect result.
11492 Now the chain builder is disabled if either:
11494 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11496 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11498 The reasoning behind this is that an application would not want the
11499 auto chain building to take place if extra chain certificates are
11500 present and it might also want a means of sending no additional
11501 certificates (for example the chain has two certificates and the
11506 * Add the possibility to build without the ENGINE framework.
11508 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11510 * Under Win32 gmtime() can return NULL: check return value in
11511 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11515 * DSA routines: under certain error conditions uninitialized BN objects
11516 could be freed. Solution: make sure initialization is performed early
11517 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11518 Nils Larsch <nla@trustcenter.de> via PR#459)
11522 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11523 checked on reconnect on the client side, therefore session resumption
11524 could still fail with a "ssl session id is different" error. This
11525 behaviour is masked when SSL_OP_ALL is used due to
11526 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11527 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11528 followup to PR #377.
11532 * IA-32 assembler support enhancements: unified ELF targets, support
11533 for SCO/Caldera platforms, fix for Cygwin shared build.
11537 * Add support for FreeBSD on sparc64. As a consequence, support for
11538 FreeBSD on non-x86 processors is separate from x86 processors on
11539 the config script, much like the NetBSD support.
11541 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11543 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11545 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11548 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11549 code (06) was taken as the first octet of the session ID and the last
11550 octet was ignored consequently. As a result SSLv2 client side session
11551 caching could not have worked due to the session ID mismatch between
11553 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11558 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11559 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11564 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11565 seems that in spite of existing for more than a year, many application
11566 author have done nothing to provide the necessary callbacks, which
11567 means that this particular engine will not work properly anywhere.
11568 This is a very unfortunate situation which forces us, in the name
11569 of usability, to give the hw_ncipher.c a static lock, which is part
11571 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11572 appear in 0.9.8 or later. We EXPECT application authors to have
11573 dealt properly with this when 0.9.8 is released (unless we actually
11574 make such changes in the libcrypto locking code that changes will
11575 have to be made anyway).
11579 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11580 octets have been read, EOF or an error occurs. Without this change
11581 some truncated ASN1 structures will not produce an error.
11585 * Disable Heimdal support, since it hasn't been fully implemented.
11586 Still give the possibility to force the use of Heimdal, but with
11587 warnings and a request that patches get sent to openssl-dev.
11591 * Add the VC-CE target, introduce the WINCE sysname, and add
11592 INSTALL.WCE and appropriate conditionals to make it build.
11594 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11596 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11597 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11598 edit numbers of the version.
11600 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11602 * Introduce safe string copy and catenation functions
11603 (BUF_strlcpy() and BUF_strlcat()).
11605 *Ben Laurie (CHATS) and Richard Levitte*
11607 * Avoid using fixed-size buffers for one-line DNs.
11609 *Ben Laurie (CHATS)*
11611 * Add BUF_MEM_grow_clean() to avoid information leakage when
11612 resizing buffers containing secrets, and use where appropriate.
11614 *Ben Laurie (CHATS)*
11616 * Avoid using fixed size buffers for configuration file location.
11618 *Ben Laurie (CHATS)*
11620 * Avoid filename truncation for various CA files.
11622 *Ben Laurie (CHATS)*
11624 * Use sizeof in preference to magic numbers.
11626 *Ben Laurie (CHATS)*
11628 * Avoid filename truncation in cert requests.
11630 *Ben Laurie (CHATS)*
11632 * Add assertions to check for (supposedly impossible) buffer
11635 *Ben Laurie (CHATS)*
11637 * Don't cache truncated DNS entries in the local cache (this could
11638 potentially lead to a spoofing attack).
11640 *Ben Laurie (CHATS)*
11642 * Fix various buffers to be large enough for hex/decimal
11643 representations in a platform independent manner.
11645 *Ben Laurie (CHATS)*
11647 * Add CRYPTO_realloc_clean() to avoid information leakage when
11648 resizing buffers containing secrets, and use where appropriate.
11650 *Ben Laurie (CHATS)*
11652 * Add BIO_indent() to avoid much slightly worrying code to do
11655 *Ben Laurie (CHATS)*
11657 * Convert sprintf()/BIO_puts() to BIO_printf().
11659 *Ben Laurie (CHATS)*
11661 * buffer_gets() could terminate with the buffer only half
11664 *Ben Laurie (CHATS)*
11666 * Add assertions to prevent user-supplied crypto functions from
11667 overflowing internal buffers by having large block sizes, etc.
11669 *Ben Laurie (CHATS)*
11671 * New OPENSSL_assert() macro (similar to assert(), but enabled
11674 *Ben Laurie (CHATS)*
11676 * Eliminate unused copy of key in RC4.
11678 *Ben Laurie (CHATS)*
11680 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11682 *Ben Laurie (CHATS)*
11684 * Fix off-by-one error in EGD path.
11686 *Ben Laurie (CHATS)*
11688 * If RANDFILE path is too long, ignore instead of truncating.
11690 *Ben Laurie (CHATS)*
11692 * Eliminate unused and incorrectly sized X.509 structure
11695 *Ben Laurie (CHATS)*
11697 * Eliminate unused and dangerous function knumber().
11699 *Ben Laurie (CHATS)*
11701 * Eliminate unused and dangerous structure, KSSL_ERR.
11703 *Ben Laurie (CHATS)*
11705 * Protect against overlong session ID context length in an encoded
11706 session object. Since these are local, this does not appear to be
11709 *Ben Laurie (CHATS)*
11711 * Change from security patch (see 0.9.6e below) that did not affect
11712 the 0.9.6 release series:
11714 Remote buffer overflow in SSL3 protocol - an attacker could
11715 supply an oversized master key in Kerberos-enabled versions.
11718 *Ben Laurie (CHATS)*
11720 * Change the SSL kerb5 codes to match RFC 2712.
11724 * Make -nameopt work fully for req and add -reqopt switch.
11726 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11728 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11730 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11732 * Make sure tests can be performed even if the corresponding algorithms
11733 have been removed entirely. This was also the last step to make
11734 OpenSSL compilable with DJGPP under all reasonable conditions.
11736 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11738 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11739 to allow version independent disabling of normally unselected ciphers,
11740 which may be activated as a side-effect of selecting a single cipher.
11742 (E.g., cipher list string "RSA" enables ciphersuites that are left
11743 out of "ALL" because they do not provide symmetric encryption.
11744 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11746 *Lutz Jaenicke, Bodo Moeller*
11748 * Add appropriate support for separate platform-dependent build
11749 directories. The recommended way to make a platform-dependent
11750 build directory is the following (tested on Linux), maybe with
11753 # Place yourself outside of the OpenSSL source tree. In
11754 # this example, the environment variable OPENSSL_SOURCE
11755 # is assumed to contain the absolute OpenSSL source directory.
11756 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11757 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11758 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11759 mkdir -p `dirname $F`
11760 ln -s $OPENSSL_SOURCE/$F $F
11763 To be absolutely sure not to disturb the source tree, a "make clean"
11764 is a good thing. If it isn't successful, don't worry about it,
11765 it probably means the source directory is very clean.
11769 * Make sure any ENGINE control commands make local copies of string
11770 pointers passed to them whenever necessary. Otherwise it is possible
11771 the caller may have overwritten (or deallocated) the original string
11772 data when a later ENGINE operation tries to use the stored values.
11774 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11776 * Improve diagnostics in file reading and command-line digests.
11778 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11780 * Add AES modes CFB and OFB to the object database. Correct an
11781 error in AES-CFB decryption.
11785 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11786 allows existing EVP_CIPHER_CTX structures to be reused after
11787 calling `EVP_*Final()`. This behaviour is used by encryption
11788 BIOs and some applications. This has the side effect that
11789 applications must explicitly clean up cipher contexts with
11790 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11794 * Check the values of dna and dnb in bn_mul_recursive before calling
11795 bn_mul_comba (a non zero value means the a or b arrays do not contain
11796 n2 elements) and fallback to bn_mul_normal if either is not zero.
11800 * Fix escaping of non-ASCII characters when using the -subj option
11801 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11805 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11806 form for "surname", serialNumber has no short form.
11807 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11808 therefore remove "mail" short name for "internet 7".
11809 The OID for unique identifiers in X509 certificates is
11810 x500UniqueIdentifier, not uniqueIdentifier.
11811 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11815 * Add an "init" command to the ENGINE config module and auto initialize
11816 ENGINEs. Without any "init" command the ENGINE will be initialized
11817 after all ctrl commands have been executed on it. If init=1 the
11818 ENGINE is initialized at that point (ctrls before that point are run
11819 on the uninitialized ENGINE and after on the initialized one). If
11820 init=0 then the ENGINE will not be initialized at all.
11824 * Fix the 'app_verify_callback' interface so that the user-defined
11825 argument is actually passed to the callback: In the
11826 SSL_CTX_set_cert_verify_callback() prototype, the callback
11827 declaration has been changed from
11830 int (*cb)(X509_STORE_CTX *,void *);
11831 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11832 i=s->ctx->app_verify_callback(&ctx)
11833 has been changed into
11834 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11836 To update applications using SSL_CTX_set_cert_verify_callback(),
11837 a dummy argument can be added to their callback functions.
11839 *D. K. Smetters <smetters@parc.xerox.com>*
11841 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11843 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11845 * Add and OPENSSL_LOAD_CONF define which will cause
11846 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11847 This allows older applications to transparently support certain
11848 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11849 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11850 load the config file and OPENSSL_add_all_algorithms_conf() which will
11851 always load it have also been added.
11855 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11856 Adjust NIDs and EVP layer.
11858 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11860 * Config modules support in openssl utility.
11862 Most commands now load modules from the config file,
11863 though in a few (such as version) this isn't done
11864 because it couldn't be used for anything.
11866 In the case of ca and req the config file used is
11867 the same as the utility itself: that is the -config
11868 command line option can be used to specify an
11873 * Move default behaviour from OPENSSL_config(). If appname is NULL
11874 use "openssl_conf" if filename is NULL use default openssl config file.
11878 * Add an argument to OPENSSL_config() to allow the use of an alternative
11879 config section name. Add a new flag to tolerate a missing config file
11880 and move code to CONF_modules_load_file().
11884 * Support for crypto accelerator cards from Accelerated Encryption
11885 Processing, www.aep.ie. (Use engine 'aep')
11886 The support was copied from 0.9.6c [engine] and adapted/corrected
11887 to work with the new engine framework.
11889 *AEP Inc. and Richard Levitte*
11891 * Support for SureWare crypto accelerator cards from Baltimore
11892 Technologies. (Use engine 'sureware')
11893 The support was copied from 0.9.6c [engine] and adapted
11894 to work with the new engine framework.
11898 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11899 make the newer ENGINE framework commands for the CHIL engine work.
11901 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11903 * Make it possible to produce shared libraries on ReliantUNIX.
11905 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11907 * Add the configuration target debug-linux-ppro.
11908 Make 'openssl rsa' use the general key loading routines
11909 implemented in `apps.c`, and make those routines able to
11910 handle the key format FORMAT_NETSCAPE and the variant
11913 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11915 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11917 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11919 * Add -keyform to rsautl, and document -engine.
11921 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11923 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11924 BIO_R_NO_SUCH_FILE error code rather than the generic
11925 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11929 * Add new functions
11930 ERR_peek_last_error
11931 ERR_peek_last_error_line
11932 ERR_peek_last_error_line_data.
11933 These are similar to
11935 ERR_peek_error_line
11936 ERR_peek_error_line_data,
11937 but report on the latest error recorded rather than the first one
11938 still in the error queue.
11940 *Ben Laurie, Bodo Moeller*
11942 * default_algorithms option in ENGINE config module. This allows things
11944 default_algorithms = ALL
11945 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11949 * Preliminary ENGINE config module.
11953 * New experimental application configuration code.
11957 * Change the AES code to follow the same name structure as all other
11958 symmetric ciphers, and behave the same way. Move everything to
11959 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11961 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11963 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11965 *Ben Laurie and Theo de Raadt*
11967 * Add option to output public keys in req command.
11969 *Massimiliano Pala madwolf@openca.org*
11971 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11972 (up to about 10% better than before for P-192 and P-224).
11976 * New functions/macros
11978 SSL_CTX_set_msg_callback(ctx, cb)
11979 SSL_CTX_set_msg_callback_arg(ctx, arg)
11980 SSL_set_msg_callback(ssl, cb)
11981 SSL_set_msg_callback_arg(ssl, arg)
11983 to request calling a callback function
11985 void cb(int write_p, int version, int content_type,
11986 const void *buf, size_t len, SSL *ssl, void *arg)
11988 whenever a protocol message has been completely received
11989 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11990 protocol version according to which the SSL library interprets
11991 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11992 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11993 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11994 specification (change_cipher_spec(20), alert(21), handshake(22)).
11995 'buf' and 'len' point to the actual message, 'ssl' to the
11996 SSL object, and 'arg' is the application-defined value set by
11997 SSL[_CTX]_set_msg_callback_arg().
11999 'openssl s_client' and 'openssl s_server' have new '-msg' options
12000 to enable a callback that displays all protocol messages.
12004 * Change the shared library support so shared libraries are built as
12005 soon as the corresponding static library is finished, and thereby get
12006 openssl and the test programs linked against the shared library.
12007 This still only happens when the keyword "shard" has been given to
12008 the configuration scripts.
12010 NOTE: shared library support is still an experimental thing, and
12011 backward binary compatibility is still not guaranteed.
12013 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12015 * Add support for Subject Information Access extension.
12017 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12019 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12020 additional bytes when new memory had to be allocated, not just
12021 when reusing an existing buffer.
12025 * New command line and configuration option 'utf8' for the req command.
12026 This allows field values to be specified as UTF8 strings.
12030 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12031 runs for the former and machine-readable output for the latter.
12035 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12036 of the e-mail address in the DN (i.e., it will go into a certificate
12037 extension only). The new configuration file option 'email_in_dn = no'
12038 has the same effect.
12040 *Massimiliano Pala madwolf@openca.org*
12042 * Change all functions with names starting with `des_` to be starting
12043 with `DES_` instead. Add wrappers that are compatible with libdes,
12044 but are named `_ossl_old_des_*`. Finally, add macros that map the
12045 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12046 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12047 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12050 Since we provide two compatibility mappings, the user needs to
12051 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12052 compatibility is desired. The default (i.e., when that macro
12053 isn't defined) is OpenSSL 0.9.6c compatibility.
12055 There are also macros that enable and disable the support of old
12056 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12057 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12058 are defined, the default will apply: to support the old des routines.
12060 In either case, one must include openssl/des.h to get the correct
12061 definitions. Do not try to just include openssl/des_old.h, that
12064 NOTE: This is a major break of an old API into a new one. Software
12065 authors are encouraged to switch to the `DES_` style functions. Some
12066 time in the future, des_old.h and the libdes compatibility functions
12067 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12068 default), and then completely removed.
12072 * Test for certificates which contain unsupported critical extensions.
12073 If such a certificate is found during a verify operation it is
12074 rejected by default: this behaviour can be overridden by either
12075 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12076 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12077 X509_supported_extension() has also been added which returns 1 if a
12078 particular extension is supported.
12082 * Modify the behaviour of EVP cipher functions in similar way to digests
12083 to retain compatibility with existing code.
12087 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12088 compatibility with existing code. In particular the 'ctx' parameter does
12089 not have to be to be initialized before the call to EVP_DigestInit() and
12090 it is tidied up after a call to EVP_DigestFinal(). New function
12091 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12092 EVP_MD_CTX_copy() changed to not require the destination to be
12093 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12094 requires the destination to be valid.
12096 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12097 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12101 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12102 so that complete 'Handshake' protocol structures are kept in memory
12103 instead of overwriting 'msg_type' and 'length' with 'body' data.
12107 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12109 *Massimo Santin via Richard Levitte*
12111 * Major restructuring to the underlying ENGINE code. This includes
12112 reduction of linker bloat, separation of pure "ENGINE" manipulation
12113 (initialisation, etc) from functionality dealing with implementations
12114 of specific crypto interfaces. This change also introduces integrated
12115 support for symmetric ciphers and digest implementations - so ENGINEs
12116 can now accelerate these by providing EVP_CIPHER and EVP_MD
12117 implementations of their own. This is detailed in
12118 [crypto/engine/README.md](crypto/engine/README.md)
12119 as it couldn't be adequately described here. However, there are a few
12120 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12121 were changed in the original introduction of ENGINE code have now
12122 reverted back - the hooking from this code to ENGINE is now a good
12123 deal more passive and at run-time, operations deal directly with
12124 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12125 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12126 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12127 they were not being used by the framework as there is no concept of a
12128 BIGNUM_METHOD and they could not be generalised to the new
12129 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12130 ENGINE_cpy() has been removed as it cannot be consistently defined in
12135 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12139 * Change mkdef.pl to sort symbols that get the same entry number,
12140 and make sure the automatically generated functions `ERR_load_*`
12141 become part of libeay.num as well.
12145 * New function SSL_renegotiate_pending(). This returns true once
12146 renegotiation has been requested (either SSL_renegotiate() call
12147 or HelloRequest/ClientHello received from the peer) and becomes
12148 false once a handshake has been completed.
12149 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12150 sends a HelloRequest, but does not ensure that a handshake takes
12151 place. SSL_renegotiate_pending() is useful for checking if the
12152 client has followed the request.)
12156 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12157 By default, clients may request session resumption even during
12158 renegotiation (if session ID contexts permit); with this option,
12159 session resumption is possible only in the first handshake.
12161 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12162 more bits available for options that should not be part of
12163 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12167 * Add some demos for certificate and certificate request creation.
12171 * Make maximum certificate chain size accepted from the peer application
12172 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12173 "Douglas E. Engert" <deengert@anl.gov>.
12177 * Add support for shared libraries for Unixware-7
12178 (Boyd Lynn Gerber <gerberb@zenez.com>).
12182 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12183 be done prior to destruction. Use this to unload error strings from
12184 ENGINEs that load their own error strings. NB: This adds two new API
12185 functions to "get" and "set" this destroy handler in an ENGINE.
12189 * Alter all existing ENGINE implementations (except "openssl" and
12190 "openbsd") to dynamically instantiate their own error strings. This
12191 makes them more flexible to be built both as statically-linked ENGINEs
12192 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12193 Also, add stub code to each that makes building them as self-contained
12194 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12198 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12199 implementations into applications that are completely implemented in
12200 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12201 commands that can be used to configure what shared-library to load and
12202 to control aspects of the way it is handled. Also, made an update to
12203 the [README-Engine.md](README-Engine.md) file
12204 that brings its information up-to-date and
12205 provides some information and instructions on the "dynamic" ENGINE
12206 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12210 * Make it possible to unload ranges of ERR strings with a new
12211 "ERR_unload_strings" function.
12215 * Add a copy() function to EVP_MD.
12219 * Make EVP_MD routines take a context pointer instead of just the
12220 md_data void pointer.
12224 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12225 that the digest can only process a single chunk of data
12226 (typically because it is provided by a piece of
12227 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12228 is only going to provide a single chunk of data, and hence the
12229 framework needn't accumulate the data for oneshot drivers.
12233 * As with "ERR", make it possible to replace the underlying "ex_data"
12234 functions. This change also alters the storage and management of global
12235 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12236 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12237 index counters. The API functions that use this state have been changed
12238 to take a "class_index" rather than pointers to the class's local STACK
12239 and counter, and there is now an API function to dynamically create new
12240 classes. This centralisation allows us to (a) plug a lot of the
12241 thread-safety problems that existed, and (b) makes it possible to clean
12242 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12243 such data would previously have always leaked in application code and
12244 workarounds were in place to make the memory debugging turn a blind eye
12245 to it. Application code that doesn't use this new function will still
12246 leak as before, but their memory debugging output will announce it now
12247 rather than letting it slide.
12249 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12250 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12251 has a return value to indicate success or failure.
12255 * Make it possible to replace the underlying "ERR" functions such that the
12256 global state (2 LHASH tables and 2 locks) is only used by the "default"
12257 implementation. This change also adds two functions to "get" and "set"
12258 the implementation prior to it being automatically set the first time
12259 any other ERR function takes place. Ie. an application can call "get",
12260 pass the return value to a module it has just loaded, and that module
12261 can call its own "set" function using that value. This means the
12262 module's "ERR" operations will use (and modify) the error state in the
12263 application and not in its own statically linked copy of OpenSSL code.
12267 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12268 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12269 the operation, and provides a more encapsulated way for external code
12270 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12271 to use these functions rather than manually incrementing the counts.
12273 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12277 * Add EVP test program.
12281 * Add symmetric cipher support to ENGINE. Expect the API to change!
12285 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12286 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12287 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12288 These allow a CRL to be built without having to access X509_CRL fields
12289 directly. Modify 'ca' application to use new functions.
12293 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12294 bug workarounds. Rollback attack detection is a security feature.
12295 The problem will only arise on OpenSSL servers when TLSv1 is not
12296 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12297 Software authors not wanting to support TLSv1 will have special reasons
12298 for their choice and can explicitly enable this option.
12300 *Bodo Moeller, Lutz Jaenicke*
12302 * Rationalise EVP so it can be extended: don't include a union of
12303 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12304 (similar to those existing for EVP_CIPHER_CTX).
12309 EVP_MD_CTX_init(&md); /* new function call */
12310 EVP_DigestInit(&md, EVP_sha1());
12311 EVP_DigestUpdate(&md, in, len);
12312 EVP_DigestFinal(&md, out, NULL);
12313 EVP_MD_CTX_cleanup(&md); /* new function call */
12317 * Make DES key schedule conform to the usual scheme, as well as
12318 correcting its structure. This means that calls to DES functions
12319 now have to pass a pointer to a des_key_schedule instead of a
12320 plain des_key_schedule (which was actually always a pointer
12323 des_key_schedule ks;
12325 des_set_key_checked(..., &ks);
12326 des_ncbc_encrypt(..., &ks, ...);
12328 (Note that a later change renames 'des_...' into 'DES_...'.)
12332 * Initial reduction of linker bloat: the use of some functions, such as
12333 PEM causes large amounts of unused functions to be linked in due to
12334 poor organisation. For example pem_all.c contains every PEM function
12335 which has a knock on effect of linking in large amounts of (unused)
12336 ASN1 code. Grouping together similar functions and splitting unrelated
12337 functions prevents this.
12341 * Cleanup of EVP macros.
12345 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12346 correct `_ecb suffix`.
12350 * Add initial OCSP responder support to ocsp application. The
12351 revocation information is handled using the text based index
12352 use by the ca application. The responder can either handle
12353 requests generated internally, supplied in files (for example
12354 via a CGI script) or using an internal minimal server.
12358 * Add configuration choices to get zlib compression for TLS.
12362 * Changes to Kerberos SSL for RFC 2712 compliance:
12363 1. Implemented real KerberosWrapper, instead of just using
12364 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12365 2. Implemented optional authenticator field of KerberosWrapper.
12367 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12368 and authenticator structs; see crypto/krb5/.
12370 Generalized Kerberos calls to support multiple Kerberos libraries.
12371 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12372 via Richard Levitte*
12374 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12375 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12376 values for each of the key sizes rather than having just
12377 parameters (and 'speed' generating keys each time).
12381 * Speed up EVP routines.
12384 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12385 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12386 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12387 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12389 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12390 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12391 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12394 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12396 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12400 * Added the OS2-EMX target.
12402 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12404 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12405 New functions to support `NCONF `routines in extension code.
12406 New function `CONF_set_nconf()`
12407 to allow functions which take an `NCONF` to also handle the old `LHASH`
12408 structure: this means that the old `CONF` compatible routines can be
12409 retained (in particular w.rt. extensions) without having to duplicate the
12410 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12414 * Enhance the general user interface with mechanisms for inner control
12415 and with possibilities to have yes/no kind of prompts.
12419 * Change all calls to low level digest routines in the library and
12420 applications to use EVP. Add missing calls to HMAC_cleanup() and
12421 don't assume HMAC_CTX can be copied using memcpy().
12423 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12425 * Add the possibility to control engines through control names but with
12426 arbitrary arguments instead of just a string.
12427 Change the key loaders to take a UI_METHOD instead of a callback
12428 function pointer. NOTE: this breaks binary compatibility with earlier
12429 versions of OpenSSL [engine].
12430 Adapt the nCipher code for these new conditions and add a card insertion
12435 * Enhance the general user interface with mechanisms to better support
12436 dialog box interfaces, application-defined prompts, the possibility
12437 to use defaults (for example default passwords from somewhere else)
12438 and interrupts/cancellations.
12442 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12443 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12447 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12448 tidy up some unnecessarily weird code in 'sk_new()').
12450 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12452 * Change the key loading routines for ENGINEs to use the same kind
12453 callback (pem_password_cb) as all other routines that need this
12458 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12459 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12460 than this minimum value is recommended.
12464 * New random seeder for OpenVMS, using the system process statistics
12465 that are easily reachable.
12469 * Windows apparently can't transparently handle global
12470 variables defined in DLLs. Initialisations such as:
12472 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12474 won't compile. This is used by the any applications that need to
12475 declare their own ASN1 modules. This was fixed by adding the option
12476 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12477 needed for static libraries under Win32.
12481 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12482 setting of purpose and trust fields. New X509_STORE trust and
12483 purpose functions and tidy up setting in other SSL functions.
12487 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12488 structure. These are inherited by X509_STORE_CTX when it is
12489 initialised. This allows various defaults to be set in the
12490 X509_STORE structure (such as flags for CRL checking and custom
12491 purpose or trust settings) for functions which only use X509_STORE_CTX
12492 internally such as S/MIME.
12494 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12495 trust settings if they are not set in X509_STORE. This allows X509_STORE
12496 purposes and trust (in S/MIME for example) to override any set by default.
12498 Add command line options for CRL checking to smime, s_client and s_server
12503 * Initial CRL based revocation checking. If the CRL checking flag(s)
12504 are set then the CRL is looked up in the X509_STORE structure and
12505 its validity and signature checked, then if the certificate is found
12506 in the CRL the verify fails with a revoked error.
12508 Various new CRL related callbacks added to X509_STORE_CTX structure.
12510 Command line options added to 'verify' application to support this.
12512 This needs some additional work, such as being able to handle multiple
12513 CRLs with different times, extension based lookup (rather than just
12514 by subject name) and ultimately more complete V2 CRL extension
12519 * Add a general user interface API (crypto/ui/). This is designed
12520 to replace things like des_read_password and friends (backward
12521 compatibility functions using this new API are provided).
12522 The purpose is to remove prompting functions from the DES code
12523 section as well as provide for prompting through dialog boxes in
12524 a window system and the like.
12528 * Add "ex_data" support to ENGINE so implementations can add state at a
12529 per-structure level rather than having to store it globally.
12533 * Make it possible for ENGINE structures to be copied when retrieved by
12534 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12535 This causes the "original" ENGINE structure to act like a template,
12536 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12537 operational state can be localised to each ENGINE structure, despite the
12538 fact they all share the same "methods". New ENGINE structures returned in
12539 this case have no functional references and the return value is the single
12540 structural reference. This matches the single structural reference returned
12541 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12546 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12547 needs to match any other type at all we need to manually clear the
12552 * Changes to the "openssl engine" utility to include;
12553 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12554 about an ENGINE's available control commands.
12555 - executing control commands from command line arguments using the
12556 '-pre' and '-post' switches. '-post' is only used if '-t' is
12557 specified and the ENGINE is successfully initialised. The syntax for
12558 the individual commands are colon-separated, for example;
12559 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12563 * New dynamic control command support for ENGINEs. ENGINEs can now
12564 declare their own commands (numbers), names (strings), descriptions,
12565 and input types for run-time discovery by calling applications. A
12566 subset of these commands are implicitly classed as "executable"
12567 depending on their input type, and only these can be invoked through
12568 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12569 can be based on user input, config files, etc). The distinction is
12570 that "executable" commands cannot return anything other than a boolean
12571 result and can only support numeric or string input, whereas some
12572 discoverable commands may only be for direct use through
12573 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12574 pointers, or other custom uses. The "executable" commands are to
12575 support parameterisations of ENGINE behaviour that can be
12576 unambiguously defined by ENGINEs and used consistently across any
12577 OpenSSL-based application. Commands have been added to all the
12578 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12579 control over shared-library paths without source code alterations.
12583 * Changed all ENGINE implementations to dynamically allocate their
12584 ENGINEs rather than declaring them statically. Apart from this being
12585 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12586 this also allows the implementations to compile without using the
12587 internal engine_int.h header.
12591 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12592 'const' value. Any code that should be able to modify a RAND_METHOD
12593 should already have non-const pointers to it (ie. they should only
12594 modify their own ones).
12598 * Made a variety of little tweaks to the ENGINE code.
12599 - "atalla" and "ubsec" string definitions were moved from header files
12600 to C code. "nuron" string definitions were placed in variables
12601 rather than hard-coded - allowing parameterisation of these values
12602 later on via ctrl() commands.
12603 - Removed unused "#if 0"'d code.
12604 - Fixed engine list iteration code so it uses ENGINE_free() to release
12605 structural references.
12606 - Constified the RAND_METHOD element of ENGINE structures.
12607 - Constified various get/set functions as appropriate and added
12608 missing functions (including a catch-all ENGINE_cpy that duplicates
12609 all ENGINE values onto a new ENGINE except reference counts/state).
12610 - Removed NULL parameter checks in get/set functions. Setting a method
12611 or function to NULL is a way of cancelling out a previously set
12612 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12613 and doesn't justify the extra error symbols and code.
12614 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12615 flags from engine_int.h to engine.h.
12616 - Changed prototypes for ENGINE handler functions (init(), finish(),
12617 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12621 * Implement binary inversion algorithm for BN_mod_inverse in addition
12622 to the algorithm using long division. The binary algorithm can be
12623 used only if the modulus is odd. On 32-bit systems, it is faster
12624 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12625 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12626 up to 450 bits. In 64-bit environments, the binary algorithm
12627 appears to be advantageous for much longer moduli; here we use it
12628 for moduli up to 2048 bits.
12632 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12633 could not support the combine flag in choice fields.
12637 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12638 extensions from a certificate request to the certificate.
12642 * Allow multiple 'certopt' and 'nameopt' options to be separated
12643 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12644 file: this allows the display of the certificate about to be
12645 signed to be customised, to allow certain fields to be included
12646 or excluded and extension details. The old system didn't display
12647 multicharacter strings properly, omitted fields not in the policy
12648 and couldn't display additional details such as extensions.
12652 * Function EC_POINTs_mul for multiple scalar multiplication
12653 of an arbitrary number of elliptic curve points
12654 \sum scalars[i]*points[i],
12655 optionally including the generator defined for the EC_GROUP:
12656 scalar*generator + \sum scalars[i]*points[i].
12658 EC_POINT_mul is a simple wrapper function for the typical case
12659 that the point list has just one item (besides the optional
12664 * First EC_METHODs for curves over GF(p):
12666 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12667 operations and provides various method functions that can also
12668 operate with faster implementations of modular arithmetic.
12670 EC_GFp_mont_method() reuses most functions that are part of
12671 EC_GFp_simple_method, but uses Montgomery arithmetic.
12673 *Bodo Moeller; point addition and point doubling
12674 implementation directly derived from source code provided by
12675 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12677 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12678 crypto/ec/ec_lib.c):
12680 Curves are EC_GROUP objects (with an optional group generator)
12681 based on EC_METHODs that are built into the library.
12683 Points are EC_POINT objects based on EC_GROUP objects.
12685 Most of the framework would be able to handle curves over arbitrary
12686 finite fields, but as there are no obvious types for fields other
12687 than GF(p), some functions are limited to that for now.
12691 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12692 that the file contains a complete HTTP response.
12696 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12697 change the def and num file printf format specifier from "%-40sXXX"
12698 to "%-39s XXX". The latter will always guarantee a space after the
12699 field while the former will cause them to run together if the field
12700 is 40 of more characters long.
12704 * Constify the cipher and digest 'method' functions and structures
12705 and modify related functions to take constant EVP_MD and EVP_CIPHER
12710 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12711 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12715 * Modify `EVP_Digest*()` routines so they now return values. Although the
12716 internal software routines can never fail additional hardware versions
12721 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12723 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12724 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12727 ERR_R_NESTED_ASN1_ERROR
12729 ERR_R_MISSING_ASN1_EOS
12730 were 4 .. 9, conflicting with
12731 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12733 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12734 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12736 Add new error code 'ERR_R_INTERNAL_ERROR'.
12740 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12745 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12746 sets the subject name for a new request or supersedes the
12747 subject name in a given request. Formats that can be parsed are
12748 'CN=Some Name, OU=myOU, C=IT'
12750 'CN=Some Name/OU=myOU/C=IT'.
12752 Add options '-batch' and '-verbose' to 'openssl req'.
12754 *Massimiliano Pala <madwolf@hackmasters.net>*
12756 * Introduce the possibility to access global variables through
12757 functions on platform were that's the best way to handle exporting
12758 global variables in shared libraries. To enable this functionality,
12759 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12760 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12761 is normally done by Configure or something similar).
12763 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12764 in the source file (foo.c) like this:
12766 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12767 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12769 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12770 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12772 OPENSSL_DECLARE_GLOBAL(int,foo);
12773 #define foo OPENSSL_GLOBAL_REF(foo)
12774 OPENSSL_DECLARE_GLOBAL(double,bar);
12775 #define bar OPENSSL_GLOBAL_REF(bar)
12777 The #defines are very important, and therefore so is including the
12778 header file everywhere where the defined globals are used.
12780 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12781 of ASN.1 items, but that structure is a bit different.
12783 The largest change is in util/mkdef.pl which has been enhanced with
12784 better and easier to understand logic to choose which symbols should
12785 go into the Windows .def files as well as a number of fixes and code
12786 cleanup (among others, algorithm keywords are now sorted
12787 lexicographically to avoid constant rewrites).
12791 * In BN_div() keep a copy of the sign of 'num' before writing the
12792 result to 'rm' because if rm==num the value will be overwritten
12793 and produce the wrong result if 'num' is negative: this caused
12794 problems with BN_mod() and BN_nnmod().
12798 * Function OCSP_request_verify(). This checks the signature on an
12799 OCSP request and verifies the signer certificate. The signer
12800 certificate is just checked for a generic purpose and OCSP request
12805 * Add OCSP_check_validity() function to check the validity of OCSP
12806 responses. OCSP responses are prepared in real time and may only
12807 be a few seconds old. Simply checking that the current time lies
12808 between thisUpdate and nextUpdate max reject otherwise valid responses
12809 caused by either OCSP responder or client clock inaccuracy. Instead
12810 we allow thisUpdate and nextUpdate to fall within a certain period of
12811 the current time. The age of the response can also optionally be
12812 checked. Two new options -validity_period and -status_age added to
12817 * If signature or public key algorithm is unrecognized print out its
12818 OID rather that just UNKNOWN.
12822 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12823 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12824 ID to be generated from the issuer certificate alone which can then be
12825 passed to OCSP_id_issuer_cmp().
12829 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12830 ASN1 modules to export functions returning ASN1_ITEM pointers
12831 instead of the ASN1_ITEM structures themselves. This adds several
12832 new macros which allow the underlying ASN1 function/structure to
12833 be accessed transparently. As a result code should not use ASN1_ITEM
12834 references directly (such as &X509_it) but instead use the relevant
12835 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12836 use of the new ASN1 code on platforms where exporting structures
12837 is problematical (for example in shared libraries) but exporting
12838 functions returning pointers to structures is not.
12842 * Add support for overriding the generation of SSL/TLS session IDs.
12843 These callbacks can be registered either in an SSL_CTX or per SSL.
12844 The purpose of this is to allow applications to control, if they wish,
12845 the arbitrary values chosen for use as session IDs, particularly as it
12846 can be useful for session caching in multiple-server environments. A
12847 command-line switch for testing this (and any client code that wishes
12848 to use such a feature) has been added to "s_server".
12850 *Geoff Thorpe, Lutz Jaenicke*
12852 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12853 of the form `#if defined(...) || defined(...) || ...` and
12854 `#if !defined(...) && !defined(...) && ...`. This also avoids
12855 the growing number of special cases it was previously handling.
12859 * Make all configuration macros available for application by making
12860 sure they are available in opensslconf.h, by giving them names starting
12861 with `OPENSSL_` to avoid conflicts with other packages and by making
12862 sure e_os2.h will cover all platform-specific cases together with
12864 Additionally, it is now possible to define configuration/platform-
12865 specific names (called "system identities"). In the C code, these
12866 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12867 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12868 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12873 * New option -set_serial to 'req' and 'x509' this allows the serial
12874 number to use to be specified on the command line. Previously self
12875 signed certificates were hard coded with serial number 0 and the
12876 CA options of 'x509' had to use a serial number in a file which was
12881 * New options to 'ca' utility to support V2 CRL entry extensions.
12882 Currently CRL reason, invalidity date and hold instruction are
12883 supported. Add new CRL extensions to V3 code and some new objects.
12887 * New function EVP_CIPHER_CTX_set_padding() this is used to
12888 disable standard block padding (aka PKCS#5 padding) in the EVP
12889 API, which was previously mandatory. This means that the data is
12890 not padded in any way and so the total length much be a multiple
12891 of the block size, otherwise an error occurs.
12895 * Initial (incomplete) OCSP SSL support.
12899 * New function OCSP_parse_url(). This splits up a URL into its host,
12900 port and path components: primarily to parse OCSP URLs. New -url
12901 option to ocsp utility.
12905 * New nonce behavior. The return value of OCSP_check_nonce() now
12906 reflects the various checks performed. Applications can decide
12907 whether to tolerate certain situations such as an absent nonce
12908 in a response when one was present in a request: the ocsp application
12909 just prints out a warning. New function OCSP_add1_basic_nonce()
12910 this is to allow responders to include a nonce in a response even if
12911 the request is nonce-less.
12915 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12916 skipped when using openssl x509 multiple times on a single input file,
12917 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12921 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12922 set string type: to handle setting ASN1_TIME structures. Fix ca
12923 utility to correctly initialize revocation date of CRLs.
12927 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12928 the clients preferred ciphersuites and rather use its own preferences.
12929 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12930 Internet Explorer by ensuring unchanged hash method during stepup.
12931 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12935 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12936 to aes and add a new 'exist' option to print out symbols that don't
12941 * Additional options to ocsp utility to allow flags to be set and
12942 additional certificates supplied.
12946 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12947 OCSP client a number of certificate to only verify the response
12952 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12953 handle the new API. Currently only ECB, CBC modes supported. Add new
12956 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12957 Encryption Standard (AES) Ciphersuites for Transport Layer
12958 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12959 not enabled by default and were not part of the "ALL" ciphersuite
12960 alias because they were not yet official; they could be
12961 explicitly requested by specifying the "AESdraft" ciphersuite
12962 group alias. In the final release of OpenSSL 0.9.7, the group
12963 alias is called "AES" and is part of "ALL".)
12965 *Ben Laurie, Steve Henson, Bodo Moeller*
12967 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12968 request to response.
12972 * Functions for OCSP responders. OCSP_request_onereq_count(),
12973 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12974 extract information from a certificate request. OCSP_response_create()
12975 creates a response and optionally adds a basic response structure.
12976 OCSP_basic_add1_status() adds a complete single response to a basic
12977 response and returns the OCSP_SINGLERESP structure just added (to allow
12978 extensions to be included for example). OCSP_basic_add1_cert() adds a
12979 certificate to a basic response and OCSP_basic_sign() signs a basic
12980 response with various flags. New helper functions ASN1_TIME_check()
12981 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12982 (converts ASN1_TIME to GeneralizedTime).
12986 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12987 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12988 structure from a certificate. X509_pubkey_digest() digests the public_key
12989 contents: this is used in various key identifiers.
12993 * Make sk_sort() tolerate a NULL argument.
12995 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12997 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12998 passed by the function are trusted implicitly. If any of them signed the
12999 response then it is assumed to be valid and is not verified.
13003 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13004 to data. This was previously part of the PKCS7 ASN1 code. This
13005 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13006 *Steve Henson, reported by Kenneth R. Robinette
13007 <support@securenetterm.com>*
13009 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13010 routines: without these tracing memory leaks is very painful.
13011 Fix leaks in PKCS12 and PKCS7 routines.
13015 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13016 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13017 effectively meant GeneralizedTime would never be used. Now it
13018 is initialised to -1 but X509_time_adj() now has to check the value
13019 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13020 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13021 *Steve Henson, reported by Kenneth R. Robinette
13022 <support@securenetterm.com>*
13024 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13025 result in a zero length in the ASN1_INTEGER structure which was
13026 not consistent with the structure when d2i_ASN1_INTEGER() was used
13027 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13028 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13029 where it did not print out a minus for negative ASN1_INTEGER.
13033 * Add summary printout to ocsp utility. The various functions which
13034 convert status values to strings have been renamed to:
13035 OCSP_response_status_str(), OCSP_cert_status_str() and
13036 OCSP_crl_reason_str() and are no longer static. New options
13037 to verify nonce values and to disable verification. OCSP response
13038 printout format cleaned up.
13042 * Add additional OCSP certificate checks. These are those specified
13043 in RFC2560. This consists of two separate checks: the CA of the
13044 certificate being checked must either be the OCSP signer certificate
13045 or the issuer of the OCSP signer certificate. In the latter case the
13046 OCSP signer certificate must contain the OCSP signing extended key
13047 usage. This check is performed by attempting to match the OCSP
13048 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13049 in the OCSP_CERTID structures of the response.
13053 * Initial OCSP certificate verification added to OCSP_basic_verify()
13054 and related routines. This uses the standard OpenSSL certificate
13055 verify routines to perform initial checks (just CA validity) and
13056 to obtain the certificate chain. Then additional checks will be
13057 performed on the chain. Currently the root CA is checked to see
13058 if it is explicitly trusted for OCSP signing. This is used to set
13059 a root CA as a global signing root: that is any certificate that
13060 chains to that CA is an acceptable OCSP signing certificate.
13064 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13065 extensions from a separate configuration file.
13066 As when reading extensions from the main configuration file,
13067 the '-extensions ...' option may be used for specifying the
13070 *Massimiliano Pala <madwolf@comune.modena.it>*
13072 * New OCSP utility. Allows OCSP requests to be generated or
13073 read. The request can be sent to a responder and the output
13074 parsed, outputted or printed in text form. Not complete yet:
13075 still needs to check the OCSP response validity.
13079 * New subcommands for 'openssl ca':
13080 `openssl ca -status <serial>` prints the status of the cert with
13081 the given serial number (according to the index file).
13082 `openssl ca -updatedb` updates the expiry status of certificates
13085 *Massimiliano Pala <madwolf@comune.modena.it>*
13087 * New '-newreq-nodes' command option to CA.pl. This is like
13088 '-newreq', but calls 'openssl req' with the '-nodes' option
13089 so that the resulting key is not encrypted.
13091 *Damien Miller <djm@mindrot.org>*
13093 * New configuration for the GNU Hurd.
13095 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13097 * Initial code to implement OCSP basic response verify. This
13098 is currently incomplete. Currently just finds the signer's
13099 certificate and verifies the signature on the response.
13103 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13104 value of OPENSSLDIR. This is available via the new '-d' option
13105 to 'openssl version', and is also included in 'openssl version -a'.
13109 * Allowing defining memory allocation callbacks that will be given
13110 file name and line number information in additional arguments
13111 (a `const char*` and an int). The basic functionality remains, as
13112 well as the original possibility to just replace malloc(),
13113 realloc() and free() by functions that do not know about these
13114 additional arguments. To register and find out the current
13115 settings for extended allocation functions, the following
13116 functions are provided:
13118 CRYPTO_set_mem_ex_functions
13119 CRYPTO_set_locked_mem_ex_functions
13120 CRYPTO_get_mem_ex_functions
13121 CRYPTO_get_locked_mem_ex_functions
13123 These work the same way as CRYPTO_set_mem_functions and friends.
13124 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13125 extended allocation function is enabled.
13126 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13127 a conventional allocation function is enabled.
13129 *Richard Levitte, Bodo Moeller*
13131 * Finish off removing the remaining LHASH function pointer casts.
13132 There should no longer be any prototype-casting required when using
13133 the LHASH abstraction, and any casts that remain are "bugs". See
13134 the callback types and macros at the head of lhash.h for details
13135 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13139 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13140 If /dev/[u]random devices are not available or do not return enough
13141 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13143 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13144 /etc/entropy will be queried once each in this sequence, querying stops
13145 when enough entropy was collected without querying more sockets.
13149 * Change the Unix RAND_poll() variant to be able to poll several
13150 random devices, as specified by DEVRANDOM, until a sufficient amount
13151 of data has been collected. We spend at most 10 ms on each file
13152 (select timeout) and read in non-blocking mode. DEVRANDOM now
13153 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13154 (previously it was just the string "/dev/urandom"), so on typical
13155 platforms the 10 ms delay will never occur.
13156 Also separate out the Unix variant to its own file, rand_unix.c.
13157 For VMS, there's a currently-empty rand_vms.c.
13161 * Move OCSP client related routines to ocsp_cl.c. These
13162 provide utility functions which an application needing
13163 to issue a request to an OCSP responder and analyse the
13164 response will typically need: as opposed to those which an
13165 OCSP responder itself would need which will be added later.
13167 OCSP_request_sign() signs an OCSP request with an API similar
13168 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13169 response. OCSP_response_get1_basic() extracts basic response
13170 from response. OCSP_resp_find_status(): finds and extracts status
13171 information from an OCSP_CERTID structure (which will be created
13172 when the request structure is built). These are built from lower
13173 level functions which work on OCSP_SINGLERESP structures but
13174 won't normally be used unless the application wishes to examine
13175 extensions in the OCSP response for example.
13177 Replace nonce routines with a pair of functions.
13178 OCSP_request_add1_nonce() adds a nonce value and optionally
13179 generates a random value. OCSP_check_nonce() checks the
13180 validity of the nonce in an OCSP response.
13184 * Change function OCSP_request_add() to OCSP_request_add0_id().
13185 This doesn't copy the supplied OCSP_CERTID and avoids the
13186 need to free up the newly created id. Change return type
13187 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13188 This can then be used to add extensions to the request.
13189 Deleted OCSP_request_new(), since most of its functionality
13190 is now in OCSP_REQUEST_new() (and the case insensitive name
13191 clash) apart from the ability to set the request name which
13192 will be added elsewhere.
13196 * Update OCSP API. Remove obsolete extensions argument from
13197 various functions. Extensions are now handled using the new
13198 OCSP extension code. New simple OCSP HTTP function which
13199 can be used to send requests and parse the response.
13203 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13204 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13205 uses the special reorder version of SET OF to sort the attributes
13206 and reorder them to match the encoded order. This resolves a long
13207 standing problem: a verify on a PKCS7 structure just after signing
13208 it used to fail because the attribute order did not match the
13209 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13210 it uses the received order. This is necessary to tolerate some broken
13211 software that does not order SET OF. This is handled by encoding
13212 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13213 to produce the required SET OF.
13217 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13218 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13219 files to get correct declarations of the ASN.1 item variables.
13223 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13224 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13225 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13226 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13227 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13228 ASN1_ITEM and no wrapper functions.
13232 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13233 replace the old function pointer based I/O routines. Change most of
13234 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13238 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13239 lines, recognize more "algorithms" that can be deselected, and make
13240 it complain about algorithm deselection that isn't recognised.
13244 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13245 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13246 to use new functions. Add NO_ASN1_OLD which can be set to remove
13247 some old style ASN1 functions: this can be used to determine if old
13248 code will still work when these eventually go away.
13252 * New extension functions for OCSP structures, these follow the
13253 same conventions as certificates and CRLs.
13257 * New function X509V3_add1_i2d(). This automatically encodes and
13258 adds an extension. Its behaviour can be customised with various
13259 flags to append, replace or delete. Various wrappers added for
13260 certificates and CRLs.
13264 * Fix to avoid calling the underlying ASN1 print routine when
13265 an extension cannot be parsed. Correct a typo in the
13266 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13270 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13271 entries for variables.
13275 * Add functionality to `apps/openssl.c` for detecting locking
13276 problems: As the program is single-threaded, all we have
13277 to do is register a locking callback using an array for
13278 storing which locks are currently held by the program.
13282 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13283 SSL_get_ex_data_X509_STORE_idx(), which is used in
13284 ssl_verify_cert_chain() and thus can be called at any time
13285 during TLS/SSL handshakes so that thread-safety is essential.
13286 Unfortunately, the ex_data design is not at all suited
13287 for multi-threaded use, so it probably should be abolished.
13291 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13293 *Broadcom, tweaked and integrated by Geoff Thorpe*
13295 * Move common extension printing code to new function
13296 X509V3_print_extensions(). Reorganise OCSP print routines and
13297 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13301 * New function X509_signature_print() to remove duplication in some
13306 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13307 set (this was treated exactly the same as SET OF previously). This
13308 is used to reorder the STACK representing the structure to match the
13309 encoding. This will be used to get round a problem where a PKCS7
13310 structure which was signed could not be verified because the STACK
13311 order did not reflect the encoded order.
13315 * Reimplement the OCSP ASN1 module using the new code.
13319 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13320 for its ASN1 operations. The old style function pointers still exist
13321 for now but they will eventually go away.
13325 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13326 completely replaces the old ASN1 functionality with a table driven
13327 encoder and decoder which interprets an ASN1_ITEM structure describing
13328 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13329 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13330 has also been converted to the new form.
13334 * Change BN_mod_exp_recp so that negative moduli are tolerated
13335 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13336 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13337 for negative moduli.
13341 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13342 of not touching the result's sign bit.
13346 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13351 * Changed the LHASH code to use prototypes for callbacks, and created
13352 macros to declare and implement thin (optionally static) functions
13353 that provide type-safety and avoid function pointer casting for the
13354 type-specific callbacks.
13358 * Added Kerberos Cipher Suites to be used with TLS, as written in
13360 *Veers Staats <staatsvr@asc.hpc.mil>,
13361 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13363 * Reformat the FAQ so the different questions and answers can be divided
13364 in sections depending on the subject.
13368 * Have the zlib compression code load ZLIB.DLL dynamically under
13373 * New function BN_mod_sqrt for computing square roots modulo a prime
13374 (using the probabilistic Tonelli-Shanks algorithm unless
13375 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13376 be handled deterministically).
13378 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13380 * Make BN_mod_inverse faster by explicitly handling small quotients
13381 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13382 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13386 * New function BN_kronecker.
13390 * Fix BN_gcd so that it works on negative inputs; the result is
13391 positive unless both parameters are zero.
13392 Previously something reasonably close to an infinite loop was
13393 possible because numbers could be growing instead of shrinking
13394 in the implementation of Euclid's algorithm.
13398 * Fix BN_is_word() and BN_is_one() macros to take into account the
13399 sign of the number in question.
13401 Fix BN_is_word(a,w) to work correctly for w == 0.
13403 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13404 because its test if the absolute value of 'a' equals 'w'.
13405 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13406 it exists mostly for use in the implementations of BN_is_zero(),
13407 BN_is_one(), and BN_is_word().
13411 * New function BN_swap.
13415 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13416 the exponentiation functions are more likely to produce reasonable
13417 results on negative inputs.
13421 * Change BN_mod_mul so that the result is always non-negative.
13422 Previously, it could be negative if one of the factors was negative;
13423 I don't think anyone really wanted that behaviour.
13427 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13428 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13429 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13430 and add new functions:
13439 BN_mod_lshift1_quick
13441 BN_mod_lshift_quick
13443 These functions always generate non-negative results.
13445 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13446 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13448 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13449 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13450 be reduced modulo `m`.
13452 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13455 The following entry accidentally appeared in the CHANGES file
13456 distributed with OpenSSL 0.9.7. The modifications described in
13457 it do *not* apply to OpenSSL 0.9.7.
13459 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13460 was actually never needed) and in BN_mul(). The removal in BN_mul()
13461 required a small change in bn_mul_part_recursive() and the addition
13462 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13463 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13464 bn_sub_words() and bn_add_words() except they take arrays with
13470 * In 'openssl passwd', verify passwords read from the terminal
13471 unless the '-salt' option is used (which usually means that
13472 verification would just waste user's time since the resulting
13473 hash is going to be compared with some given password hash)
13474 or the new '-noverify' option is used.
13476 This is an incompatible change, but it does not affect
13477 non-interactive use of 'openssl passwd' (passwords on the command
13478 line, '-stdin' option, '-in ...' option) and thus should not
13479 cause any problems.
13483 * Remove all references to RSAref, since there's no more need for it.
13487 * Make DSO load along a path given through an environment variable
13488 (SHLIB_PATH) with shl_load().
13492 * Constify the ENGINE code as a result of BIGNUM constification.
13493 Also constify the RSA code and most things related to it. In a
13494 few places, most notable in the depth of the ASN.1 code, ugly
13495 casts back to non-const were required (to be solved at a later
13500 * Make it so the openssl application has all engines loaded by default.
13504 * Constify the BIGNUM routines a little more.
13508 * Add the following functions:
13510 ENGINE_load_cswift()
13512 ENGINE_load_atalla()
13513 ENGINE_load_nuron()
13514 ENGINE_load_builtin_engines()
13516 That way, an application can itself choose if external engines that
13517 are built-in in OpenSSL shall ever be used or not. The benefit is
13518 that applications won't have to be linked with libdl or other dso
13519 libraries unless it's really needed.
13521 Changed 'openssl engine' to load all engines on demand.
13522 Changed the engine header files to avoid the duplication of some
13523 declarations (they differed!).
13527 * 'openssl engine' can now list capabilities.
13531 * Better error reporting in 'openssl engine'.
13535 * Never call load_dh_param(NULL) in s_server.
13539 * Add engine application. It can currently list engines by name and
13540 identity, and test if they are actually available.
13544 * Improve RPM specification file by forcing symbolic linking and making
13545 sure the installed documentation is also owned by root.root.
13547 *Damien Miller <djm@mindrot.org>*
13549 * Give the OpenSSL applications more possibilities to make use of
13550 keys (public as well as private) handled by engines.
13554 * Add OCSP code that comes from CertCo.
13558 * Add VMS support for the Rijndael code.
13562 * Added untested support for Nuron crypto accelerator.
13566 * Add support for external cryptographic devices. This code was
13567 previously distributed separately as the "engine" branch.
13569 *Geoff Thorpe, Richard Levitte*
13571 * Rework the filename-translation in the DSO code. It is now possible to
13572 have far greater control over how a "name" is turned into a filename
13573 depending on the operating environment and any oddities about the
13574 different shared library filenames on each system.
13578 * Support threads on FreeBSD-elf in Configure.
13582 * Fix for SHA1 assembly problem with MASM: it produces
13583 warnings about corrupt line number information when assembling
13584 with debugging information. This is caused by the overlapping
13587 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13590 NCONF_get_number() has no error checking at all. As a replacement,
13591 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13592 promoted strongly. The old NCONF_get_number is kept around for
13593 binary backward compatibility.
13594 Make it possible for methods to load from something other than a BIO,
13595 by providing a function pointer that is given a name instead of a BIO.
13596 For example, this could be used to load configuration data from an
13601 * Fix for non blocking accept BIOs. Added new I/O special reason
13602 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13603 with non blocking I/O was not possible because no retry code was
13604 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13609 * Added the beginnings of Rijndael support.
13613 * Fix for bug in DirectoryString mask setting. Add support for
13614 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13615 to allow certificate printing to more controllable, additional
13616 'certopt' option to 'x509' to allow new printing options to be
13621 * Clean old EAY MD5 hack from e_os.h.
13625 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13627 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13628 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13630 *Joe Orton, Steve Henson*
13632 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13634 * Fix additional bug revealed by the NISCC test suite:
13636 Stop bug triggering large recursion when presented with
13637 certain ASN.1 tags ([CVE-2003-0851])
13641 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13643 * Fix various bugs revealed by running the NISCC test suite:
13645 Stop out of bounds reads in the ASN1 code when presented with
13646 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13648 If verify callback ignores invalid public key errors don't try to check
13649 certificate signature with the NULL public key.
13653 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13654 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13659 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13660 extra data after the compression methods not only for TLS 1.0
13661 but also for SSL 3.0 (as required by the specification).
13663 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13665 * Change X509_certificate_type() to mark the key as exported/exportable
13666 when it's 512 *bits* long, not 512 bytes.
13670 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13672 * Countermeasure against the Klima-Pokorny-Rosa extension of
13673 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13674 a protocol version number mismatch like a decryption error
13675 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13679 * Turn on RSA blinding by default in the default implementation
13680 to avoid a timing attack. Applications that don't want it can call
13681 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13682 They would be ill-advised to do so in most cases.
13684 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13686 * Change RSA blinding code so that it works when the PRNG is not
13687 seeded (in this case, the secret RSA exponent is abused as
13688 an unpredictable seed -- if it is not unpredictable, there
13689 is no point in blinding anyway). Make RSA blinding thread-safe
13690 by remembering the creator's thread ID in rsa->blinding and
13691 having all other threads use local one-time blinding factors
13692 (this requires more computation than sharing rsa->blinding, but
13693 avoids excessive locking; and if an RSA object is not shared
13694 between threads, blinding will still be very fast).
13698 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13700 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13701 via timing by performing a MAC computation even if incorrect
13702 block cipher padding has been found. This is a countermeasure
13703 against active attacks where the attacker has to distinguish
13704 between bad padding and a MAC verification error. ([CVE-2003-0078])
13706 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13707 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13708 Martin Vuagnoux (EPFL, Ilion)*
13710 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13712 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13713 memory from its contents. This is done with a counter that will
13714 place alternating values in each byte. This can be used to solve
13715 two issues: 1) the removal of calls to memset() by highly optimizing
13716 compilers, and 2) cleansing with other values than 0, since those can
13717 be read through on certain media, for example a swap space on disk.
13721 * Bugfix: client side session caching did not work with external caching,
13722 because the session->cipher setting was not restored when reloading
13723 from the external cache. This problem was masked, when
13724 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13725 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13729 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13730 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13732 *Zeev Lieber <zeev-l@yahoo.com>*
13734 * Undo an undocumented change introduced in 0.9.6e which caused
13735 repeated calls to OpenSSL_add_all_ciphers() and
13736 OpenSSL_add_all_digests() to be ignored, even after calling
13741 * Change the default configuration reader to deal with last line not
13742 being properly terminated.
13746 * Change X509_NAME_cmp() so it applies the special rules on handling
13747 DN values that are of type PrintableString, as well as RDNs of type
13748 emailAddress where the value has the type ia5String.
13750 *stefank@valicert.com via Richard Levitte*
13752 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13753 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13754 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13755 the bitwise-OR of the two for use by the majority of applications
13756 wanting this behaviour, and update the docs. The documented
13757 behaviour and actual behaviour were inconsistent and had been
13758 changing anyway, so this is more a bug-fix than a behavioural
13761 *Geoff Thorpe, diagnosed by Nadav Har'El*
13763 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13764 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13768 * Fix initialization code race conditions in
13769 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13770 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13771 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13772 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13773 ssl2_get_cipher_by_char(),
13774 ssl3_get_cipher_by_char().
13776 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13778 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13779 the cached sessions are flushed, as the remove_cb() might use ex_data
13780 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13781 (see [openssl.org #212]).
13783 *Geoff Thorpe, Lutz Jaenicke*
13785 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13786 length, instead of the encoding length to d2i_ASN1_OBJECT.
13790 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13792 * [In 0.9.6g-engine release:]
13793 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13795 *Lynn Gazis <lgazis@rainbow.com>*
13797 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13799 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13800 and get fix the header length calculation.
13801 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13802 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13804 * Use proper error handling instead of 'assertions' in buffer
13805 overflow checks added in 0.9.6e. This prevents DoS (the
13806 assertions could call abort()).
13808 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13810 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13812 * Add various sanity checks to asn1_get_length() to reject
13813 the ASN1 length bytes if they exceed sizeof(long), will appear
13814 negative or the content length exceeds the length of the
13817 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13819 * Fix cipher selection routines: ciphers without encryption had no flags
13820 for the cipher strength set and where therefore not handled correctly
13821 by the selection routines (PR #130).
13825 * Fix EVP_dsa_sha macro.
13830 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13831 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13832 that was added in OpenSSL 0.9.6d.
13834 As the countermeasure turned out to be incompatible with some
13835 broken SSL implementations, the new option is part of SSL_OP_ALL.
13836 SSL_OP_ALL is usually employed when compatibility with weird SSL
13837 implementations is desired (e.g. '-bugs' option to 's_client' and
13838 's_server'), so the new option is automatically set in many
13843 * Changes in security patch:
13845 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13846 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13847 Air Force Materiel Command, USAF, under agreement number
13850 * Add various sanity checks to asn1_get_length() to reject
13851 the ASN1 length bytes if they exceed sizeof(long), will appear
13852 negative or the content length exceeds the length of the
13853 supplied buffer. ([CVE-2002-0659])
13855 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13857 * Assertions for various potential buffer overflows, not known to
13858 happen in practice.
13860 *Ben Laurie (CHATS)*
13862 * Various temporary buffers to hold ASCII versions of integers were
13863 too small for 64 bit platforms. ([CVE-2002-0655])
13864 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13866 * Remote buffer overflow in SSL3 protocol - an attacker could
13867 supply an oversized session ID to a client. ([CVE-2002-0656])
13869 *Ben Laurie (CHATS)*
13871 * Remote buffer overflow in SSL2 protocol - an attacker could
13872 supply an oversized client master key. ([CVE-2002-0656])
13874 *Ben Laurie (CHATS)*
13876 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13878 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13879 encoded as NULL) with id-dsa-with-sha1.
13881 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13883 * Check various `X509_...()` return values in `apps/req.c`.
13885 *Nils Larsch <nla@trustcenter.de>*
13887 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13888 an end-of-file condition would erroneously be flagged, when the CRLF
13889 was just at the end of a processed block. The bug was discovered when
13890 processing data through a buffering memory BIO handing the data to a
13891 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13892 <ptsekov@syntrex.com> and Nedelcho Stanev.
13896 * Implement a countermeasure against a vulnerability recently found
13897 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13898 before application data chunks to avoid the use of known IVs
13899 with data potentially chosen by the attacker.
13903 * Fix length checks in ssl3_get_client_hello().
13907 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13908 to prevent ssl3_read_internal() from incorrectly assuming that
13909 ssl3_read_bytes() found application data while handshake
13910 processing was enabled when in fact s->s3->in_read_app_data was
13911 merely automatically cleared during the initial handshake.
13913 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13915 * Fix object definitions for Private and Enterprise: they were not
13916 recognized in their shortname (=lowercase) representation. Extend
13917 obj_dat.pl to issue an error when using undefined keywords instead
13918 of silently ignoring the problem (Svenning Sorensen
13919 <sss@sss.dnsalias.net>).
13923 * Fix DH_generate_parameters() so that it works for 'non-standard'
13924 generators, i.e. generators other than 2 and 5. (Previously, the
13925 code did not properly initialise the 'add' and 'rem' values to
13926 BN_generate_prime().)
13928 In the new general case, we do not insist that 'generator' is
13929 actually a primitive root: This requirement is rather pointless;
13930 a generator of the order-q subgroup is just as good, if not
13935 * Map new X509 verification errors to alerts. Discovered and submitted by
13936 Tom Wu <tom@arcot.com>.
13940 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13941 returning non-zero before the data has been completely received
13942 when using non-blocking I/O.
13944 *Bodo Moeller; problem pointed out by John Hughes*
13946 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13948 *Ben Laurie, Lutz Jaenicke*
13950 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13951 Yoram Zahavi <YoramZ@gilian.com>).
13955 * Add information about CygWin 1.3 and on, and preserve proper
13956 configuration for the versions before that.
13958 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13960 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13961 check whether we deal with a copy of a session and do not delete from
13962 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13963 <izhar@checkpoint.com>.
13967 * Do not store session data into the internal session cache, if it
13968 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13969 flag is set). Proposed by Aslam <aslam@funk.com>.
13973 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13978 * [In 0.9.6d-engine release:]
13979 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13981 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13983 * Add the configuration target linux-s390x.
13985 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13987 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13988 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13989 variable as an indication that a ClientHello message has been
13990 received. As the flag value will be lost between multiple
13991 invocations of ssl3_accept when using non-blocking I/O, the
13992 function may not be aware that a handshake has actually taken
13993 place, thus preventing a new session from being added to the
13996 To avoid this problem, we now set s->new_session to 2 instead of
13997 using a local variable.
13999 *Lutz Jaenicke, Bodo Moeller*
14001 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14002 if the SSL_R_LENGTH_MISMATCH error is detected.
14004 *Geoff Thorpe, Bodo Moeller*
14006 * New 'shared_ldflag' column in Configure platform table.
14010 * Fix EVP_CIPHER_mode macro.
14012 *"Dan S. Camper" <dan@bti.net>*
14014 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14015 type, we must throw them away by setting rr->length to 0.
14017 *D P Chang <dpc@qualys.com>*
14019 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14021 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14022 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14023 worked incorrectly for those cases where range = `10..._2` and
14024 `3*range` is two bits longer than range.)
14028 * Only add signing time to PKCS7 structures if it is not already
14033 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14034 OBJ_ld_ce should be OBJ_id_ce.
14035 Also some ip-pda OIDs in crypto/objects/objects.txt were
14036 incorrect (cf. RFC 3039).
14038 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14040 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14041 returns early because it has nothing to do.
14043 *Andy Schneider <andy.schneider@bjss.co.uk>*
14045 * [In 0.9.6c-engine release:]
14046 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14048 *Andy Schneider <andy.schneider@bjss.co.uk>*
14050 * [In 0.9.6c-engine release:]
14051 Add support for Cryptographic Appliance's keyserver technology.
14052 (Use engine 'keyclient')
14054 *Cryptographic Appliances and Geoff Thorpe*
14056 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14057 is called via tools/c89.sh because arguments have to be
14058 rearranged (all '-L' options must appear before the first object
14061 *Richard Shapiro <rshapiro@abinitio.com>*
14063 * [In 0.9.6c-engine release:]
14064 Add support for Broadcom crypto accelerator cards, backported
14067 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14069 * [In 0.9.6c-engine release:]
14070 Add support for SureWare crypto accelerator cards from
14071 Baltimore Technologies. (Use engine 'sureware')
14073 *Baltimore Technologies and Mark Cox*
14075 * [In 0.9.6c-engine release:]
14076 Add support for crypto accelerator cards from Accelerated
14077 Encryption Processing, www.aep.ie. (Use engine 'aep')
14079 *AEP Inc. and Mark Cox*
14081 * Add a configuration entry for gcc on UnixWare.
14083 *Gary Benson <gbenson@redhat.com>*
14085 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14086 messages are stored in a single piece (fixed-length part and
14087 variable-length part combined) and fix various bugs found on the way.
14091 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14092 instead. BIO_gethostbyname() does not know what timeouts are
14093 appropriate, so entries would stay in cache even when they have
14095 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14097 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14098 faced with a pathologically small ClientHello fragment that does
14099 not contain client_version: Instead of aborting with an error,
14100 simply choose the highest available protocol version (i.e.,
14101 TLS 1.0 unless it is disabled). In practice, ClientHello
14102 messages are never sent like this, but this change gives us
14103 strictly correct behaviour at least for TLS.
14107 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14108 never resets s->method to s->ctx->method when called from within
14109 one of the SSL handshake functions.
14111 *Bodo Moeller; problem pointed out by Niko Baric*
14113 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14114 (sent using the client's version number) if client_version is
14115 smaller than the protocol version in use. Also change
14116 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14117 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14118 the client will at least see that alert.
14122 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14127 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14128 client receives HelloRequest while in a handshake.
14130 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14132 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14133 should end in 'break', not 'goto end' which circumvents various
14134 cleanups done in state SSL_ST_OK. But session related stuff
14135 must be disabled for SSL_ST_OK in the case that we just sent a
14138 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14139 before just sending a HelloRequest.
14141 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14143 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14144 reveal whether illegal block cipher padding was found or a MAC
14145 verification error occurred. (Neither SSLerr() codes nor alerts
14146 are directly visible to potential attackers, but the information
14147 may leak via logfiles.)
14149 Similar changes are not required for the SSL 2.0 implementation
14150 because the number of padding bytes is sent in clear for SSL 2.0,
14151 and the extra bytes are just ignored. However ssl/s2_pkt.c
14152 failed to verify that the purported number of padding bytes is in
14157 * Add OpenUNIX-8 support including shared libraries
14158 (Boyd Lynn Gerber <gerberb@zenez.com>).
14162 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14163 'wristwatch attack' using huge encoding parameters (cf.
14164 James H. Manger's CRYPTO 2001 paper). Note that the
14165 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14166 encoding parameters and hence was not vulnerable.
14170 * BN_sqr() bug fix.
14172 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14174 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14175 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14176 followed by modular reduction.
14178 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14180 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14181 equivalent based on BN_pseudo_rand() instead of BN_rand().
14185 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14186 This function was broken, as the check for a new client hello message
14187 to handle SGC did not allow these large messages.
14188 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14192 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14196 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14197 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14201 * Rework the configuration and shared library support for Tru64 Unix.
14202 The configuration part makes use of modern compiler features and
14203 still retains old compiler behavior for those that run older versions
14204 of the OS. The shared library support part includes a variant that
14205 uses the RPATH feature, and is available through the special
14206 configuration target "alpha-cc-rpath", which will never be selected
14209 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14211 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14212 with the same message size as in ssl3_get_certificate_request().
14213 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14214 messages might inadvertently be reject as too long.
14216 *Petr Lampa <lampa@fee.vutbr.cz>*
14218 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14222 * Modified SSL library such that the verify_callback that has been set
14223 specifically for an SSL object with SSL_set_verify() is actually being
14224 used. Before the change, a verify_callback set with this function was
14225 ignored and the verify_callback() set in the SSL_CTX at the time of
14226 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14227 to allow the necessary settings.
14231 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14232 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14233 done automatically (in contradiction to the requirements of the C
14234 standard). This made problems when used from OpenSSH.
14238 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14239 dh->length and always used
14241 BN_rand_range(priv_key, dh->p).
14243 BN_rand_range() is not necessary for Diffie-Hellman, and this
14244 specific range makes Diffie-Hellman unnecessarily inefficient if
14245 dh->length (recommended exponent length) is much smaller than the
14246 length of dh->p. We could use BN_rand_range() if the order of
14247 the subgroup was stored in the DH structure, but we only have
14252 BN_rand(priv_key, l, ...)
14254 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14261 RSA_eay_public_encrypt
14262 RSA_eay_private_decrypt
14263 RSA_eay_private_encrypt (signing)
14264 RSA_eay_public_decrypt (signature verification)
14266 (default implementations for RSA_public_encrypt,
14267 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14268 always reject numbers >= n.
14272 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14273 to synchronize access to 'locking_thread'. This is necessary on
14274 systems where access to 'locking_thread' (an 'unsigned long'
14275 variable) is not atomic.
14279 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14280 *before* setting the 'crypto_lock_rand' flag. The previous code had
14281 a race condition if 0 is a valid thread ID.
14283 *Travis Vitek <vitek@roguewave.com>*
14285 * Add support for shared libraries under Irix.
14287 *Albert Chin-A-Young <china@thewrittenword.com>*
14289 * Add configuration option to build on Linux on both big-endian and
14290 little-endian MIPS.
14292 *Ralf Baechle <ralf@uni-koblenz.de>*
14294 * Add the possibility to create shared libraries on HP-UX.
14298 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14300 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14301 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14302 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14303 PRNG state recovery was possible based on the output of
14304 one PRNG request appropriately sized to gain knowledge on
14305 'md' followed by enough consecutive 1-byte PRNG requests
14306 to traverse all of 'state'.
14308 1. When updating 'md_local' (the current thread's copy of 'md')
14309 during PRNG output generation, hash all of the previous
14310 'md_local' value, not just the half used for PRNG output.
14312 2. Make the number of bytes from 'state' included into the hash
14313 independent from the number of PRNG bytes requested.
14315 The first measure alone would be sufficient to avoid
14316 Markku-Juhani's attack. (Actually it had never occurred
14317 to me that the half of 'md_local' used for chaining was the
14318 half from which PRNG output bytes were taken -- I had always
14319 assumed that the secret half would be used.) The second
14320 measure makes sure that additional data from 'state' is never
14321 mixed into 'md_local' in small portions; this heuristically
14322 further strengthens the PRNG.
14326 * Fix crypto/bn/asm/mips3.s.
14330 * When only the key is given to "enc", the IV is undefined. Print out
14331 an error message in this case.
14335 * Handle special case when X509_NAME is empty in X509 printing routines.
14339 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14340 positive and less than q.
14344 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14345 used: it isn't thread safe and the add_lock_callback should handle
14348 *Paul Rose <Paul.Rose@bridge.com>*
14350 * Verify that incoming data obeys the block size in
14351 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14357 *Ulf Möller, Bodo Möller*
14359 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14360 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14361 when fixing the server behaviour for backwards-compatible 'client
14362 hello' messages. (Note that the attack is impractical against
14363 SSL 3.0 and TLS 1.0 anyway because length and version checking
14364 means that the probability of guessing a valid ciphertext is
14365 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14368 Before 0.9.5, the countermeasure (hide the error by generating a
14369 random 'decryption result') did not work properly because
14370 ERR_clear_error() was missing, meaning that SSL_get_error() would
14371 detect the supposedly ignored error.
14373 Both problems are now fixed.
14377 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14378 (previously it was 1024).
14382 * Fix for compatibility mode trust settings: ignore trust settings
14383 unless some valid trust or reject settings are present.
14387 * Fix for blowfish EVP: its a variable length cipher.
14391 * Fix various bugs related to DSA S/MIME verification. Handle missing
14392 parameters in DSA public key structures and return an error in the
14393 DSA routines if parameters are absent.
14397 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14398 in the current directory if neither $RANDFILE nor $HOME was set.
14399 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14400 caused some confusion to Windows users who haven't defined $HOME.
14401 Thus RAND_file_name() is changed again: e_os.h can define a
14402 DEFAULT_HOME, which will be used if $HOME is not set.
14403 For Windows, we use "C:"; on other platforms, we still require
14404 environment variables.
14406 * Move 'if (!initialized) RAND_poll()' into regions protected by
14407 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14408 having multiple threads call RAND_poll() concurrently.
14412 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14413 combination of a flag and a thread ID variable.
14414 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14415 flag), *other* threads can enter ssleay_add_bytes without obeying
14416 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14417 that they do not hold after the first thread unsets add_do_not_lock).
14421 * Change bctest again: '-x' expressions are not available in all
14422 versions of 'test'.
14426 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14428 * Fix a couple of memory leaks in PKCS7_dataDecode()
14430 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14432 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14433 the default extension for executables, if any. Also, make the perl
14434 scripts that use symlink() to test if it really exists and use "cp"
14435 if it doesn't. All this made OpenSSL compilable and installable in
14440 * Fix for asn1_GetSequence() for indefinite length constructed data.
14441 If SEQUENCE is length is indefinite just set c->slen to the total
14442 amount of data available.
14444 *Steve Henson, reported by shige@FreeBSD.org*
14446 *This change does not apply to 0.9.7.*
14448 * Change bctest to avoid here-documents inside command substitution
14449 (workaround for FreeBSD /bin/sh bug).
14450 For compatibility with Ultrix, avoid shell functions (introduced
14451 in the bctest version that searches along $PATH).
14455 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14456 with des_encrypt() defined on some operating systems, like Solaris
14461 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14462 On the Importance of Eliminating Errors in Cryptographic
14463 Computations, J. Cryptology 14 (2001) 2, 101-119,
14464 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14468 * MIPS assembler BIGNUM division bug fix.
14472 * Disabled incorrect Alpha assembler code.
14476 * Fix PKCS#7 decode routines so they correctly update the length
14477 after reading an EOC for the EXPLICIT tag.
14481 *This change does not apply to 0.9.7.*
14483 * Fix bug in PKCS#12 key generation routines. This was triggered
14484 if a 3DES key was generated with a 0 initial byte. Include
14485 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14486 (but broken) behaviour.
14490 * Enhance bctest to search for a working bc along $PATH and print
14493 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14495 * Fix memory leaks in err.c: free err_data string if necessary;
14496 don't write to the wrong index in ERR_set_error_data.
14500 * Implement ssl23_peek (analogous to ssl23_read), which previously
14505 * Replace rdtsc with `_emit` statements for VC++ version 5.
14507 *Jeremy Cooper <jeremy@baymoo.org>*
14509 * Make it possible to reuse SSLv2 sessions.
14513 * In copy_email() check for >= 0 as a return value for
14514 X509_NAME_get_index_by_NID() since 0 is a valid index.
14516 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14518 * Avoid coredump with unsupported or invalid public keys by checking if
14519 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14520 PKCS7_verify() fails with non detached data.
14524 * Don't use getenv in library functions when run as setuid/setgid.
14525 New function OPENSSL_issetugid().
14529 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14530 due to incorrect handling of multi-threading:
14532 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14534 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14536 3. Count how many times MemCheck_off() has been called so that
14537 nested use can be treated correctly. This also avoids
14538 inband-signalling in the previous code (which relied on the
14539 assumption that thread ID 0 is impossible).
14543 * Add "-rand" option also to s_client and s_server.
14547 * Fix CPU detection on Irix 6.x.
14548 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14549 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14551 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14556 *This change does not apply to 0.9.7.*
14558 * Use the cached encoding of an X509_NAME structure rather than
14559 copying it. This is apparently the reason for the libsafe "errors"
14560 but the code is actually correct.
14564 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14565 Bleichenbacher's DSA attack.
14566 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14567 to be set and top=0 forces the highest bit to be set; top=-1 is new
14568 and leaves the highest bit random.
14570 *Ulf Moeller, Bodo Moeller*
14572 * In the `NCONF_...`-based implementations for `CONF_...` queries
14573 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14574 a temporary CONF structure with the data component set to NULL
14575 (which gives segmentation faults in lh_retrieve).
14576 Instead, use NULL for the CONF pointer in CONF_get_string and
14577 CONF_get_number (which may use environment variables) and directly
14578 return NULL from CONF_get_section.
14582 * Fix potential buffer overrun for EBCDIC.
14586 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14587 keyUsage if basicConstraints absent for a CA.
14591 * Make SMIME_write_PKCS7() write mail header values with a format that
14592 is more generally accepted (no spaces before the semicolon), since
14593 some programs can't parse those values properly otherwise. Also make
14594 sure BIO's that break lines after each write do not create invalid
14599 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14600 macros previously used would not encode an empty SEQUENCE OF
14601 and break the signature.
14605 *This change does not apply to 0.9.7.*
14607 * Zero the premaster secret after deriving the master secret in
14612 * Add some EVP_add_digest_alias registrations (as found in
14613 OpenSSL_add_all_digests()) to SSL_library_init()
14614 aka OpenSSL_add_ssl_algorithms(). This provides improved
14615 compatibility with peers using X.509 certificates
14616 with unconventional AlgorithmIdentifier OIDs.
14620 * Fix for Irix with NO_ASM.
14622 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14624 * ./config script fixes.
14626 *Ulf Moeller, Richard Levitte*
14628 * Fix 'openssl passwd -1'.
14632 * Change PKCS12_key_gen_asc() so it can cope with non null
14633 terminated strings whose length is passed in the passlen
14634 parameter, for example from PEM callbacks. This was done
14635 by adding an extra length parameter to asc2uni().
14637 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14639 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14640 call failed, free the DSA structure.
14644 * Fix to uni2asc() to cope with zero length Unicode strings.
14645 These are present in some PKCS#12 files.
14649 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14650 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14651 when writing a 32767 byte record.
14653 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14655 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14656 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14658 (RSA objects have a reference count access to which is protected
14659 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14660 so they are meant to be shared between threads.)
14661 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14662 "Reddie, Steven" <Steven.Reddie@ca.com>*
14664 * Fix a deadlock in CRYPTO_mem_leaks().
14668 * Use better test patterns in bntest.
14672 * rand_win.c fix for Borland C.
14676 * BN_rshift bugfix for n == 0.
14680 * Add a 'bctest' script that checks for some known 'bc' bugs
14681 so that 'make test' does not abort just because 'bc' is broken.
14685 * Store verify_result within SSL_SESSION also for client side to
14686 avoid potential security hole. (Re-used sessions on the client side
14687 always resulted in verify_result==X509_V_OK, not using the original
14688 result of the server certificate verification.)
14692 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14693 SSL3_RT_APPLICATION_DATA, return 0.
14694 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14699 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14700 releases, have been re-implemented by renaming the previous
14701 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14702 and ssl3_read_internal, respectively, and adding 'peek' parameters
14703 to them. The new ssl[23]_{read,peek} functions are calls to
14704 ssl[23]_read_internal with the 'peek' flag set appropriately.
14705 A 'peek' parameter has also been added to ssl3_read_bytes, which
14706 does the actual work for ssl3_read_internal.
14710 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14711 the method-specific "init()" handler. Also clean up ex_data after
14712 calling the method-specific "finish()" handler. Previously, this was
14713 happening the other way round.
14717 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14718 The previous value, 12, was not always sufficient for BN_mod_exp().
14722 * Make sure that shared libraries get the internal name engine with
14723 the full version number and not just 0. This should mark the
14724 shared libraries as not backward compatible. Of course, this should
14725 be changed again when we can guarantee backward binary compatibility.
14729 * Fix typo in get_cert_by_subject() in by_dir.c
14731 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14733 * Rework the system to generate shared libraries:
14735 - Make note of the expected extension for the shared libraries and
14736 if there is a need for symbolic links from for example libcrypto.so.0
14737 to libcrypto.so.0.9.7. There is extended info in Configure for
14740 - Make as few rebuilds of the shared libraries as possible.
14742 - Still avoid linking the OpenSSL programs with the shared libraries.
14744 - When installing, install the shared libraries separately from the
14749 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14751 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14752 and not in SSL_clear because the latter is also used by the
14753 accept/connect functions; previously, the settings made by
14754 SSL_set_read_ahead would be lost during the handshake.
14756 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14758 * Correct util/mkdef.pl to be selective about disabled algorithms.
14759 Previously, it would create entries for disabled algorithms no
14764 * Added several new manual pages for SSL_* function.
14768 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14770 * In ssl23_get_client_hello, generate an error message when faced
14771 with an initial SSL 3.0/TLS record that is too small to contain the
14772 first two bytes of the ClientHello message, i.e. client_version.
14773 (Note that this is a pathologic case that probably has never happened
14774 in real life.) The previous approach was to use the version number
14775 from the record header as a substitute; but our protocol choice
14776 should not depend on that one because it is not authenticated
14777 by the Finished messages.
14781 * More robust randomness gathering functions for Windows.
14783 *Jeffrey Altman <jaltman@columbia.edu>*
14785 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14786 not set then we don't setup the error code for issuer check errors
14787 to avoid possibly overwriting other errors which the callback does
14788 handle. If an application does set the flag then we assume it knows
14789 what it is doing and can handle the new informational codes
14794 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14795 a general "ANY" type, as such it should be able to decode anything
14796 including tagged types. However it didn't check the class so it would
14797 wrongly interpret tagged types in the same way as their universal
14798 counterpart and unknown types were just rejected. Changed so that the
14799 tagged and unknown types are handled in the same way as a SEQUENCE:
14800 that is the encoding is stored intact. There is also a new type
14801 "V_ASN1_OTHER" which is used when the class is not universal, in this
14802 case we have no idea what the actual type is so we just lump them all
14807 * On VMS, stdout may very well lead to a file that is written to
14808 in a record-oriented fashion. That means that every write() will
14809 write a separate record, which will be read separately by the
14810 programs trying to read from it. This can be very confusing.
14812 The solution is to put a BIO filter in the way that will buffer
14813 text until a linefeed is reached, and then write everything a
14814 line at a time, so every record written will be an actual line,
14815 not chunks of lines and not (usually doesn't happen, but I've
14816 seen it once) several lines in one record. BIO_f_linebuffer() is
14819 Currently, it's a VMS-only method, because that's where it has
14820 been tested well enough.
14824 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14825 it can return incorrect results.
14826 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14827 but it was in 0.9.6-beta[12].)
14831 * Disable the check for content being present when verifying detached
14832 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14833 include zero length content when signing messages.
14837 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14838 BIO_ctrl (for BIO pairs).
14842 * Add DSO method for VMS.
14846 * Bug fix: Montgomery multiplication could produce results with the
14851 * Add RPM specification openssl.spec and modify it to build three
14852 packages. The default package contains applications, application
14853 documentation and run-time libraries. The devel package contains
14854 include files, static libraries and function documentation. The
14855 doc package contains the contents of the doc directory. The original
14856 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14860 * Add a large number of documentation files for many SSL routines.
14862 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14864 * Add a configuration entry for Sony News 4.
14866 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14868 * Don't set the two most significant bits to one when generating a
14869 random number < q in the DSA library.
14873 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14874 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14875 the underlying transport is blocking) if a handshake took place.
14876 (The default behaviour is needed by applications such as s_client
14877 and s_server that use select() to determine when to use SSL_read;
14878 but for applications that know in advance when to expect data, it
14879 just makes things more complicated.)
14883 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14888 * Add a few more EBCDIC conditionals that make `req` and `x509`
14889 work better on such systems.
14891 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14893 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14894 Update PKCS12_parse() so it copies the friendlyName and the
14895 keyid to the certificates aux info.
14899 * Fix bug in PKCS7_verify() which caused an infinite loop
14900 if there was more than one signature.
14902 *Sven Uszpelkat <su@celocom.de>*
14904 * Major change in util/mkdef.pl to include extra information
14905 about each symbol, as well as presenting variables as well
14906 as functions. This change means that there's n more need
14907 to rebuild the .num files when some algorithms are excluded.
14911 * Allow the verify time to be set by an application,
14912 rather than always using the current time.
14916 * Phase 2 verify code reorganisation. The certificate
14917 verify code now looks up an issuer certificate by a
14918 number of criteria: subject name, authority key id
14919 and key usage. It also verifies self signed certificates
14920 by the same criteria. The main comparison function is
14921 X509_check_issued() which performs these checks.
14923 Lot of changes were necessary in order to support this
14924 without completely rewriting the lookup code.
14926 Authority and subject key identifier are now cached.
14928 The LHASH 'certs' is X509_STORE has now been replaced
14929 by a STACK_OF(X509_OBJECT). This is mainly because an
14930 LHASH can't store or retrieve multiple objects with
14931 the same hash value.
14933 As a result various functions (which were all internal
14934 use only) have changed to handle the new X509_STORE
14935 structure. This will break anything that messed round
14936 with X509_STORE internally.
14938 The functions X509_STORE_add_cert() now checks for an
14939 exact match, rather than just subject name.
14941 The X509_STORE API doesn't directly support the retrieval
14942 of multiple certificates matching a given criteria, however
14943 this can be worked round by performing a lookup first
14944 (which will fill the cache with candidate certificates)
14945 and then examining the cache for matches. This is probably
14946 the best we can do without throwing out X509_LOOKUP
14947 entirely (maybe later...).
14949 The X509_VERIFY_CTX structure has been enhanced considerably.
14951 All certificate lookup operations now go via a get_issuer()
14952 callback. Although this currently uses an X509_STORE it
14953 can be replaced by custom lookups. This is a simple way
14954 to bypass the X509_STORE hackery necessary to make this
14955 work and makes it possible to use more efficient techniques
14956 in future. A very simple version which uses a simple
14957 STACK for its trusted certificate store is also provided
14958 using X509_STORE_CTX_trusted_stack().
14960 The verify_cb() and verify() callbacks now have equivalents
14961 in the X509_STORE_CTX structure.
14963 X509_STORE_CTX also has a 'flags' field which can be used
14964 to customise the verify behaviour.
14968 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14969 excludes S/MIME capabilities.
14973 * When a certificate request is read in keep a copy of the
14974 original encoding of the signed data and use it when outputting
14975 again. Signatures then use the original encoding rather than
14976 a decoded, encoded version which may cause problems if the
14977 request is improperly encoded.
14981 * For consistency with other BIO_puts implementations, call
14982 buffer_write(b, ...) directly in buffer_puts instead of calling
14985 In BIO_puts, increment b->num_write as in BIO_write.
14987 *Peter.Sylvester@EdelWeb.fr*
14989 * Fix BN_mul_word for the case where the word is 0. (We have to use
14990 BN_zero, we may not return a BIGNUM with an array consisting of
14991 words set to zero.)
14995 * Avoid calling abort() from within the library when problems are
14996 detected, except if preprocessor symbols have been defined
14997 (such as REF_CHECK, BN_DEBUG etc.).
15001 * New openssl application 'rsautl'. This utility can be
15002 used for low level RSA operations. DER public key
15003 BIO/fp routines also added.
15007 * New Configure entry and patches for compiling on QNX 4.
15009 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15011 * A demo state-machine implementation was sponsored by
15012 Nuron (<http://www.nuron.com/>) and is now available in
15013 demos/state_machine.
15017 * New options added to the 'dgst' utility for signature
15018 generation and verification.
15022 * Unrecognized PKCS#7 content types are now handled via a
15023 catch all ASN1_TYPE structure. This allows unsupported
15024 types to be stored as a "blob" and an application can
15025 encode and decode it manually.
15029 * Fix various signed/unsigned issues to make a_strex.c
15030 compile under VC++.
15032 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15034 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15035 length if passed a buffer. ASN1_INTEGER_to_BN failed
15036 if passed a NULL BN and its argument was negative.
15038 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15040 * Modification to PKCS#7 encoding routines to output definite
15041 length encoding. Since currently the whole structures are in
15042 memory there's not real point in using indefinite length
15043 constructed encoding. However if OpenSSL is compiled with
15044 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15048 * Added BIO_vprintf() and BIO_vsnprintf().
15052 * Added more prefixes to parse for in the strings written
15053 through a logging bio, to cover all the levels that are available
15054 through syslog. The prefixes are now:
15056 PANIC, EMERG, EMR => LOG_EMERG
15057 ALERT, ALR => LOG_ALERT
15058 CRIT, CRI => LOG_CRIT
15059 ERROR, ERR => LOG_ERR
15060 WARNING, WARN, WAR => LOG_WARNING
15061 NOTICE, NOTE, NOT => LOG_NOTICE
15062 INFO, INF => LOG_INFO
15063 DEBUG, DBG => LOG_DEBUG
15065 and as before, if none of those prefixes are present at the
15066 beginning of the string, LOG_ERR is chosen.
15068 On Win32, the `LOG_*` levels are mapped according to this:
15070 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15071 LOG_WARNING => EVENTLOG_WARNING_TYPE
15072 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15076 * Made it possible to reconfigure with just the configuration
15077 argument "reconf" or "reconfigure". The command line arguments
15078 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15079 and are retrieved from there when reconfiguring.
15085 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15087 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15091 * The obj_dat.pl script was messing up the sorting of object
15092 names. The reason was that it compared the quoted version
15093 of strings as a result "OCSP" > "OCSP Signing" because
15094 " > SPACE. Changed script to store unquoted versions of
15095 names and add quotes on output. It was also omitting some
15096 names from the lookup table if they were given a default
15097 value (that is if SN is missing it is given the same
15098 value as LN and vice versa), these are now added on the
15099 grounds that if an object has a name we should be able to
15100 look it up. Finally added warning output when duplicate
15101 short or long names are found.
15105 * Changes needed for Tandem NSK.
15107 *Scott Uroff <scott@xypro.com>*
15109 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15110 RSA_padding_check_SSLv23(), special padding was never detected
15111 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15112 version rollback attacks was not effective.
15114 In s23_clnt.c, don't use special rollback-attack detection padding
15115 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15116 client; similarly, in s23_srvr.c, don't do the rollback check if
15117 SSL 2.0 is the only protocol enabled in the server.
15121 * Make it possible to get hexdumps of unprintable data with 'openssl
15122 asn1parse'. By implication, the functions ASN1_parse_dump() and
15123 BIO_dump_indent() are added.
15127 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15128 these print out strings and name structures based on various
15129 flags including RFC2253 support and proper handling of
15130 multibyte characters. Added options to the 'x509' utility
15131 to allow the various flags to be set.
15135 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15136 Also change the functions X509_cmp_current_time() and
15137 X509_gmtime_adj() work with an ASN1_TIME structure,
15138 this will enable certificates using GeneralizedTime in validity
15139 dates to be checked.
15143 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15144 negative public key encodings) on by default,
15145 NO_NEG_PUBKEY_BUG can be set to disable it.
15149 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15150 content octets. An i2c_ASN1_OBJECT is unnecessary because
15151 the encoding can be trivially obtained from the structure.
15155 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15156 not read locks (`CRYPTO_r_[un]lock`).
15160 * A first attempt at creating official support for shared
15161 libraries through configuration. I've kept it so the
15162 default is static libraries only, and the OpenSSL programs
15163 are always statically linked for now, but there are
15164 preparations for dynamic linking in place.
15165 This has been tested on Linux and Tru64.
15169 * Randomness polling function for Win9x, as described in:
15170 Peter Gutmann, Software Generation of Practically Strong
15175 * Fix so PRNG is seeded in req if using an already existing
15180 * New options to smime application. -inform and -outform
15181 allow alternative formats for the S/MIME message including
15182 PEM and DER. The -content option allows the content to be
15183 specified separately. This should allow things like Netscape
15184 form signing output easier to verify.
15188 * Fix the ASN1 encoding of tags using the 'long form'.
15192 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15193 STRING types. These convert content octets to and from the
15194 underlying type. The actual tag and length octets are
15195 already assumed to have been read in and checked. These
15196 are needed because all other string types have virtually
15197 identical handling apart from the tag. By having versions
15198 of the ASN1 functions that just operate on content octets
15199 IMPLICIT tagging can be handled properly. It also allows
15200 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15201 and ASN1_INTEGER are identical apart from the tag.
15205 * Change the handling of OID objects as follows:
15207 - New object identifiers are inserted in objects.txt, following
15208 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15209 - objects.pl is used to process obj_mac.num and create a new
15211 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15214 This is currently kind of a hack, and the perl code in objects.pl
15215 isn't very elegant, but it works as I intended. The simplest way
15216 to check that it worked correctly is to look in obj_dat.h and
15217 check the array nid_objs and make sure the objects haven't moved
15218 around (this is important!). Additions are OK, as well as
15219 consistent name changes.
15223 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15227 * Addition of the command line parameter '-rand file' to 'openssl req'.
15228 The given file adds to whatever has already been seeded into the
15229 random pool through the RANDFILE configuration file option or
15230 environment variable, or the default random state file.
15234 * mkstack.pl now sorts each macro group into lexical order.
15235 Previously the output order depended on the order the files
15236 appeared in the directory, resulting in needless rewriting
15241 * Patches to make OpenSSL compile under Win32 again. Mostly
15242 work arounds for the VC++ problem that it treats func() as
15243 func(void). Also stripped out the parts of mkdef.pl that
15244 added extra typesafe functions: these no longer exist.
15248 * Reorganisation of the stack code. The macros are now all
15249 collected in safestack.h . Each macro is defined in terms of
15250 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15251 DEBUG_SAFESTACK is now handled in terms of function casts,
15252 this has the advantage of retaining type safety without the
15253 use of additional functions. If DEBUG_SAFESTACK is not defined
15254 then the non typesafe macros are used instead. Also modified the
15255 mkstack.pl script to handle the new form. Needs testing to see
15256 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15257 the default if no major problems. Similar behaviour for ASN1_SET_OF
15258 and PKCS12_STACK_OF.
15262 * When some versions of IIS use the 'NET' form of private key the
15263 key derivation algorithm is different. Normally MD5(password) is
15264 used as a 128 bit RC4 key. In the modified case
15265 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15266 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15267 as the old Netscape_RSA functions except they have an additional
15268 'sgckey' parameter which uses the modified algorithm. Also added
15269 an -sgckey command line option to the rsa utility. Thanks to
15270 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15271 algorithm to openssl-dev.
15275 * The evp_local.h macros were using 'c.##kname' which resulted in
15276 invalid expansion on some systems (SCO 5.0.5 for example).
15277 Corrected to 'c.kname'.
15279 *Phillip Porch <root@theporch.com>*
15281 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15282 a STACK of email addresses from a certificate or request, these look
15283 in the subject name and the subject alternative name extensions and
15284 omit any duplicate addresses.
15288 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15289 This makes DSA verification about 2 % faster.
15293 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15294 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15295 plus overhead for 1024 bit moduli).
15296 This makes exponentiations about 0.5 % faster for 1024 bit
15297 exponents (as measured by "openssl speed rsa2048").
15301 * Rename memory handling macros to avoid conflicts with other
15303 Malloc => OPENSSL_malloc
15304 Malloc_locked => OPENSSL_malloc_locked
15305 Realloc => OPENSSL_realloc
15306 Free => OPENSSL_free
15310 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15311 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15315 * CygWin32 support.
15317 *John Jarvie <jjarvie@newsguy.com>*
15319 * The type-safe stack code has been rejigged. It is now only compiled
15320 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15321 by default all type-specific stack functions are "#define"d back to
15322 standard stack functions. This results in more streamlined output
15323 but retains the type-safety checking possibilities of the original
15328 * The STACK code has been cleaned up, and certain type declarations
15329 that didn't make a lot of sense have been brought in line. This has
15330 also involved a cleanup of sorts in safestack.h to more correctly
15331 map type-safe stack functions onto their plain stack counterparts.
15332 This work has also resulted in a variety of "const"ifications of
15333 lots of the code, especially `_cmp` operations which should normally
15334 be prototyped with "const" parameters anyway.
15338 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15339 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15340 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15341 where all of 'md' is used each time the PRNG is used, but 'state'
15342 is used only indexed by a cyclic counter. As entropy may not be
15343 well distributed from the beginning, 'md' is important as a
15344 chaining variable. However, the output function chains only half
15345 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15346 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15347 in all of 'state' being rewritten, with the new values depending
15348 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15352 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15353 the handshake is continued after ssl_verify_cert_chain();
15354 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15355 can lead to 'unexplainable' connection aborts later.
15357 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15359 * Major EVP API cipher revision.
15360 Add hooks for extra EVP features. This allows various cipher
15361 parameters to be set in the EVP interface. Support added for variable
15362 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15363 setting of RC2 and RC5 parameters.
15365 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15368 Remove lots of duplicated code from the EVP library. For example *every*
15369 cipher init() function handles the 'iv' in the same way according to the
15370 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15371 for CFB and OFB modes they zero ctx->num.
15373 New functionality allows removal of S/MIME code RC2 hack.
15375 Most of the routines have the same form and so can be declared in terms
15378 By shifting this to the top level EVP_CipherInit() it can be removed from
15379 all individual ciphers. If the cipher wants to handle IVs or keys
15380 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15383 Change lots of functions like EVP_EncryptUpdate() to now return a
15384 value: although software versions of the algorithms cannot fail
15385 any installed hardware versions can.
15389 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15390 this option is set, tolerate broken clients that send the negotiated
15391 protocol version number instead of the requested protocol version
15396 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15397 i.e. non-zero for export ciphersuites, zero otherwise.
15398 Previous versions had this flag inverted, inconsistent with
15399 rsa_tmp_cb (..._TMP_RSA_CB).
15401 *Bodo Moeller; problem reported by Amit Chopra*
15403 * Add missing DSA library text string. Work around for some IIS
15404 key files with invalid SEQUENCE encoding.
15408 * Add a document (doc/standards.txt) that list all kinds of standards
15409 and so on that are implemented in OpenSSL.
15413 * Enhance c_rehash script. Old version would mishandle certificates
15414 with the same subject name hash and wouldn't handle CRLs at all.
15415 Added -fingerprint option to crl utility, to support new c_rehash
15420 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15424 * Fix for SSL server purpose checking. Server checking was
15425 rejecting certificates which had extended key usage present
15426 but no ssl client purpose.
15428 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15430 * Make PKCS#12 code work with no password. The PKCS#12 spec
15431 is a little unclear about how a blank password is handled.
15432 Since the password in encoded as a BMPString with terminating
15433 double NULL a zero length password would end up as just the
15434 double NULL. However no password at all is different and is
15435 handled differently in the PKCS#12 key generation code. NS
15436 treats a blank password as zero length. MSIE treats it as no
15437 password on export: but it will try both on import. We now do
15438 the same: PKCS12_parse() tries zero length and no password if
15439 the password is set to "" or NULL (NULL is now a valid password:
15440 it wasn't before) as does the pkcs12 application.
15444 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15445 perror when PEM_read_bio_X509_REQ fails, the error message must
15446 be obtained from the error queue.
15450 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15451 it in ERR_remove_state if appropriate, and change ERR_get_state
15452 accordingly to avoid race conditions (this is necessary because
15453 thread_hash is no longer constant once set).
15457 * Bugfix for linux-elf makefile.one.
15461 * RSA_get_default_method() will now cause a default
15462 RSA_METHOD to be chosen if one doesn't exist already.
15463 Previously this was only set during a call to RSA_new()
15464 or RSA_new_method(NULL) meaning it was possible for
15465 RSA_get_default_method() to return NULL.
15469 * Added native name translation to the existing DSO code
15470 that will convert (if the flag to do so is set) filenames
15471 that are sufficiently small and have no path information
15472 into a canonical native form. Eg. "blah" converted to
15473 "libblah.so" or "blah.dll" etc.
15477 * New function ERR_error_string_n(e, buf, len) which is like
15478 ERR_error_string(e, buf), but writes at most 'len' bytes
15479 including the 0 terminator. For ERR_error_string_n, 'buf'
15482 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15484 * CONF library reworked to become more general. A new CONF
15485 configuration file reader "class" is implemented as well as a
15486 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15487 old `CONF_*` functions are still there, but are reimplemented to
15488 work in terms of the new functions. Also, a set of functions
15489 to handle the internal storage of the configuration data is
15490 provided to make it easier to write new configuration file
15491 reader "classes" (I can definitely see something reading a
15492 configuration file in XML format, for example), called `_CONF_*`,
15493 or "the configuration storage API"...
15495 The new configuration file reading functions are:
15497 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15498 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15500 NCONF_default, NCONF_WIN32
15502 NCONF_dump_fp, NCONF_dump_bio
15504 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15505 NCONF_new creates a new CONF object. This works in the same way
15506 as other interfaces in OpenSSL, like the BIO interface.
15507 `NCONF_dump_*` dump the internal storage of the configuration file,
15508 which is useful for debugging. All other functions take the same
15509 arguments as the old `CONF_*` functions with the exception of the
15510 first that must be a `CONF *` instead of a `LHASH *`.
15512 To make it easier to use the new classes with the old `CONF_*` functions,
15513 the function CONF_set_default_method is provided.
15517 * Add '-tls1' option to 'openssl ciphers', which was already
15518 mentioned in the documentation but had not been implemented.
15519 (This option is not yet really useful because even the additional
15520 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15524 * Initial DSO code added into libcrypto for letting OpenSSL (and
15525 OpenSSL-based applications) load shared libraries and bind to
15526 them in a portable way.
15528 *Geoff Thorpe, with contributions from Richard Levitte*
15530 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15532 * Make sure _lrotl and _lrotr are only used with MSVC.
15534 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15535 (the default implementation of RAND_status).
15537 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15538 to '-clrext' (= clear extensions), as intended and documented.
15539 *Bodo Moeller; inconsistency pointed out by Michael Attili
15540 <attili@amaxo.com>*
15542 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15543 was larger than the MD block size.
15545 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15547 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15548 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15549 using the passed key: if the passed key was a private key the result
15550 of X509_print(), for example, would be to print out all the private key
15555 * des_quad_cksum() byte order bug fix.
15556 *Ulf Möller, using the problem description in krb4-0.9.7, where
15557 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15559 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15562 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15564 * For easily testing in shell scripts whether some command
15565 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15566 returns with exit code 0 iff no command of the given name is available.
15567 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15568 the output goes to stdout and nothing is printed to stderr.
15569 Additional arguments are always ignored.
15571 Since for each cipher there is a command of the same name,
15572 the 'no-cipher' compilation switches can be tested this way.
15574 ('openssl no-XXX' is not able to detect pseudo-commands such
15575 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15579 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15583 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15584 is set; it will be thrown away anyway because each handshake creates
15586 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15587 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15588 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15589 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15593 * New s_client option -ign_eof: EOF at stdin is ignored, and
15594 'Q' and 'R' lose their special meanings (quit/renegotiate).
15595 This is part of what -quiet does; unlike -quiet, -ign_eof
15596 does not suppress any output.
15600 * Add compatibility options to the purpose and trust code. The
15601 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15602 accepts a certificate or CA, this was the previous behaviour,
15603 with all the associated security issues.
15605 X509_TRUST_COMPAT is the old trust behaviour: only and
15606 automatically trust self signed roots in certificate store. A
15607 new trust setting X509_TRUST_DEFAULT is used to specify that
15608 a purpose has no associated trust setting and it should instead
15609 use the value in the default purpose.
15613 * Fix the PKCS#8 DSA private key code so it decodes keys again
15614 and fix a memory leak.
15618 * In util/mkerr.pl (which implements 'make errors'), preserve
15619 reason strings from the previous version of the .c file, as
15620 the default to have only downcase letters (and digits) in
15621 automatically generated reasons codes is not always appropriate.
15625 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15626 using strerror. Previously, ERR_reason_error_string() returned
15627 library names as reason strings for SYSerr; but SYSerr is a special
15628 case where small numbers are errno values, not library numbers.
15632 * Add '-dsaparam' option to 'openssl dhparam' application. This
15633 converts DSA parameters into DH parameters. (When creating parameters,
15634 DSA_generate_parameters is used.)
15638 * Include 'length' (recommended exponent length) in C code generated
15639 by 'openssl dhparam -C'.
15643 * The second argument to set_label in perlasm was already being used
15644 so couldn't be used as a "file scope" flag. Moved to third argument
15649 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15650 instead of RAND_bytes for encryption IVs and salts.
15654 * Include RAND_status() into RAND_METHOD instead of implementing
15655 it only for md_rand.c Otherwise replacing the PRNG by calling
15656 RAND_set_rand_method would be impossible.
15660 * Don't let DSA_generate_key() enter an infinite loop if the random
15661 number generation fails.
15665 * New 'rand' application for creating pseudo-random output.
15669 * Added configuration support for Linux/IA64
15671 *Rolf Haberrecker <rolf@suse.de>*
15673 * Assembler module support for Mingw32.
15677 * Shared library support for HPUX (in shlib/).
15679 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15681 * Shared library support for Solaris gcc.
15683 *Lutz Behnke <behnke@trustcenter.de>*
15685 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15687 * PKCS7_encrypt() was adding text MIME headers twice because they
15688 were added manually and by SMIME_crlf_copy().
15692 * In bntest.c don't call BN_rand with zero bits argument.
15694 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15696 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15697 case was implemented. This caused BN_div_recp() to fail occasionally.
15701 * Add an optional second argument to the set_label() in the perl
15702 assembly language builder. If this argument exists and is set
15703 to 1 it signals that the assembler should use a symbol whose
15704 scope is the entire file, not just the current function. This
15705 is needed with MASM which uses the format label:: for this scope.
15707 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15709 * Change the ASN1 types so they are typedefs by default. Before
15710 almost all types were #define'd to ASN1_STRING which was causing
15711 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15716 * Change names of new functions to the new get1/get0 naming
15717 convention: After 'get1', the caller owns a reference count
15718 and has to call `..._free`; 'get0' returns a pointer to some
15719 data structure without incrementing reference counters.
15720 (Some of the existing 'get' functions increment a reference
15721 counter, some don't.)
15722 Similarly, 'set1' and 'add1' functions increase reference
15723 counters or duplicate objects.
15727 * Allow for the possibility of temp RSA key generation failure:
15728 the code used to assume it always worked and crashed on failure.
15732 * Fix potential buffer overrun problem in BIO_printf().
15733 *Ulf Möller, using public domain code by Patrick Powell; problem
15734 pointed out by David Sacerdote <das33@cornell.edu>*
15736 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15737 RAND_egd() and RAND_status(). In the command line application,
15738 the EGD socket can be specified like a seed file using RANDFILE
15743 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15744 Some CAs (e.g. Verisign) distribute certificates in this form.
15748 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15749 list to exclude them. This means that no special compilation option
15750 is needed to use anonymous DH: it just needs to be included in the
15755 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15756 EVP_MD_type. The old functionality is available in a new macro called
15757 EVP_MD_md(). Change code that uses it and update docs.
15761 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15762 where the `void *` argument is replaced by a function pointer argument.
15763 Previously `void *` was abused to point to functions, which works on
15764 many platforms, but is not correct. As these functions are usually
15765 called by macros defined in OpenSSL header files, most source code
15766 should work without changes.
15770 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15771 sections with information on -D... compiler switches used for
15772 compiling the library so that applications can see them. To enable
15773 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15774 must be defined. E.g.,
15775 #define OPENSSL_ALGORITHM_DEFINES
15776 #include <openssl/opensslconf.h>
15777 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15779 *Richard Levitte, Ulf and Bodo Möller*
15781 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15786 * Change the 'other' type in certificate aux info to a STACK_OF
15787 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15788 the required ASN1 format: arbitrary types determined by an OID.
15792 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15793 argument to 'req'. This is not because the function is newer or
15794 better than others it just uses the work 'NEW' in the certificate
15795 request header lines. Some software needs this.
15799 * Reorganise password command line arguments: now passwords can be
15800 obtained from various sources. Delete the PEM_cb function and make
15801 it the default behaviour: i.e. if the callback is NULL and the
15802 usrdata argument is not NULL interpret it as a null terminated pass
15803 phrase. If usrdata and the callback are NULL then the pass phrase
15804 is prompted for as usual.
15808 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15809 the support is automatically enabled. The resulting binaries will
15810 autodetect the card and use it if present.
15812 *Ben Laurie and Compaq Inc.*
15814 * Work around for Netscape hang bug. This sends certificate request
15815 and server done in one record. Since this is perfectly legal in the
15816 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15817 the bugs/SSLv3 entry for more info.
15821 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15825 * Add -rand argument to smime and pkcs12 applications and read/write
15830 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15834 * Add command line password options to the remaining applications.
15838 * Bug fix for BN_div_recp() for numerators with an even number of
15843 * More tests in bntest.c, and changed test_bn output.
15847 * ./config recognizes MacOS X now.
15851 * Bug fix for BN_div() when the first words of num and divisor are
15852 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15856 * Add support for various broken PKCS#8 formats, and command line
15857 options to produce them.
15861 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15862 get temporary BIGNUMs from a BN_CTX.
15866 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15871 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15872 include a #define from the old name to the new. The original intent
15873 was that statically linked binaries could for example just call
15874 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15875 link with digests. This never worked because SSLeay_add_all_digests()
15876 and SSLeay_add_all_ciphers() were in the same source file so calling
15877 one would link with the other. They are now in separate source files.
15881 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15885 * Use a less unusual form of the Miller-Rabin primality test (it used
15886 a binary algorithm for exponentiation integrated into the Miller-Rabin
15887 loop, our standard modexp algorithms are faster).
15891 * Support for the EBCDIC character set completed.
15893 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15895 * Source code cleanups: use const where appropriate, eliminate casts,
15896 use `void *` instead of `char *` in lhash.
15900 * Bugfix: ssl3_send_server_key_exchange was not restartable
15901 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15902 this the server could overwrite ephemeral keys that the client
15907 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15908 using 50 iterations of the Rabin-Miller test.
15910 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15911 iterations of the Rabin-Miller test as required by the appendix
15912 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15913 As BN_is_prime_fasttest includes trial division, DSA parameter
15914 generation becomes much faster.
15916 This implies a change for the callback functions in DSA_is_prime
15917 and DSA_generate_parameters: The callback function is called once
15918 for each positive witness in the Rabin-Miller test, not just
15919 occasionally in the inner loop; and the parameters to the
15920 callback function now provide an iteration count for the outer
15921 loop rather than for the current invocation of the inner loop.
15922 DSA_generate_parameters additionally can call the callback
15923 function with an 'iteration count' of -1, meaning that a
15924 candidate has passed the trial division test (when q is generated
15925 from an application-provided seed, trial division is skipped).
15929 * New function BN_is_prime_fasttest that optionally does trial
15930 division before starting the Rabin-Miller test and has
15931 an additional BN_CTX * argument (whereas BN_is_prime always
15932 has to allocate at least one BN_CTX).
15933 'callback(1, -1, cb_arg)' is called when a number has passed the
15934 trial division stage.
15938 * Fix for bug in CRL encoding. The validity dates weren't being handled
15943 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15947 * New function BN_pseudo_rand().
15951 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15952 bignum version of BN_from_montgomery() with the working code from
15953 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15958 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15959 made it impossible to use the same SSL_SESSION data structure in
15960 SSL2 clients in multiple threads.
15964 * The return value of RAND_load_file() no longer counts bytes obtained
15965 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15966 to seed the PRNG (previously an explicit byte count was required).
15968 *Ulf Möller, Bodo Möller*
15970 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15971 used `char *` instead of `void *` and had casts all over the place.
15975 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15979 * Retain source code compatibility for BN_prime_checks macro:
15980 BN_is_prime(..., BN_prime_checks, ...) now uses
15981 BN_prime_checks_for_size to determine the appropriate number of
15982 Rabin-Miller iterations.
15986 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15987 DH_CHECK_P_NOT_SAFE_PRIME.
15988 (Check if this is true? OpenPGP calls them "strong".)
15992 * Merge the functionality of "dh" and "gendh" programs into a new program
15993 "dhparam". The old programs are retained for now but will handle DH keys
15994 (instead of parameters) in future.
15998 * Make the ciphers, s_server and s_client programs check the return values
15999 when a new cipher list is set.
16003 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16004 ciphers. Before when the 56bit ciphers were enabled the sorting was
16007 The syntax for the cipher sorting has been extended to support sorting by
16008 cipher-strength (using the strength_bits hard coded in the tables).
16009 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16011 Fix a bug in the cipher-command parser: when supplying a cipher command
16012 string with an "undefined" symbol (neither command nor alphanumeric
16013 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16014 an error is flagged.
16016 Due to the strength-sorting extension, the code of the
16017 ssl_create_cipher_list() function was completely rearranged. I hope that
16018 the readability was also increased :-)
16020 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16022 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16023 for the first serial number and places 2 in the serial number file. This
16024 avoids problems when the root CA is created with serial number zero and
16025 the first user certificate has the same issuer name and serial number
16030 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16031 the new code. Add documentation for this stuff.
16035 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16036 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16037 structures and behave in an analogous way to the X509v3 functions:
16038 they shouldn't be called directly but wrapper functions should be used
16041 So we also now have some wrapper functions that call the X509at functions
16042 when passed certificate requests. (TO DO: similar things can be done with
16043 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16044 things. Some of these need some d2i or i2d and print functionality
16045 because they handle more complex structures.)
16049 * Add missing #ifndefs that caused missing symbols when building libssl
16050 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16051 NO_RSA in `ssl/s2*.c`.
16053 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16055 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16056 has a return value which indicates the quality of the random data
16057 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16058 error queue. New function RAND_pseudo_bytes() generates output that is
16059 guaranteed to be unique but not unpredictable. RAND_add is like
16060 RAND_seed, but takes an extra argument for an entropy estimate
16061 (RAND_seed always assumes full entropy).
16065 * Do more iterations of Rabin-Miller probable prime test (specifically,
16066 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16067 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16068 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16069 false-positive rate of at most 2^-80 for random input.
16073 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16077 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16078 in the 0.9.5 release), this returns the chain
16079 from an X509_CTX structure with a dup of the stack and all
16080 the X509 reference counts upped: so the stack will exist
16081 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16084 Also make SSL_SESSION_print() print out the verify return
16089 * Add manpage for the pkcs12 command. Also change the default
16090 behaviour so MAC iteration counts are used unless the new
16091 -nomaciter option is used. This improves file security and
16092 only older versions of MSIE (4.0 for example) need it.
16096 * Honor the no-xxx Configure options when creating .DEF files.
16100 * Add PKCS#10 attributes to field table: challengePassword,
16101 unstructuredName and unstructuredAddress. These are taken from
16102 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16103 international characters are used.
16105 More changes to X509_ATTRIBUTE code: allow the setting of types
16106 based on strings. Remove the 'loc' parameter when adding
16107 attributes because these will be a SET OF encoding which is sorted
16112 * Initial changes to the 'req' utility to allow request generation
16113 automation. This will allow an application to just generate a template
16114 file containing all the field values and have req construct the
16117 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16118 used all over the place including certificate requests and PKCS#7
16119 structures. They are currently handled manually where necessary with
16120 some primitive wrappers for PKCS#7. The new functions behave in a
16121 manner analogous to the X509 extension functions: they allow
16122 attributes to be looked up by NID and added.
16124 Later something similar to the X509V3 code would be desirable to
16125 automatically handle the encoding, decoding and printing of the
16126 more complex types. The string types like challengePassword can
16127 be handled by the string table functions.
16129 Also modified the multi byte string table handling. Now there is
16130 a 'global mask' which masks out certain types. The table itself
16131 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16132 is useful when for example there is only one permissible type
16133 (as in countryName) and using the mask might result in no valid
16138 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16139 SSL_get_peer_finished to allow applications to obtain the latest
16140 Finished messages sent to the peer or expected from the peer,
16141 respectively. (SSL_get_peer_finished is usually the Finished message
16142 actually received from the peer, otherwise the protocol will be aborted.)
16144 As the Finished message are message digests of the complete handshake
16145 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16146 be used for external authentication procedures when the authentication
16147 provided by SSL/TLS is not desired or is not enough.
16151 * Enhanced support for Alpha Linux is added. Now ./config checks if
16152 the host supports BWX extension and if Compaq C is present on the
16153 $PATH. Just exploiting of the BWX extension results in 20-30%
16154 performance kick for some algorithms, e.g. DES and RC4 to mention
16155 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16160 * Add support for MS "fast SGC". This is arguably a violation of the
16161 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16162 weak crypto and after checking the certificate is SGC a second one
16163 with strong crypto. MS SGC stops the first handshake after receiving
16164 the server certificate message and sends a second client hello. Since
16165 a server will typically do all the time consuming operations before
16166 expecting any further messages from the client (server key exchange
16167 is the most expensive) there is little difference between the two.
16169 To get OpenSSL to support MS SGC we have to permit a second client
16170 hello message after we have sent server done. In addition we have to
16171 reset the MAC if we do get this second client hello.
16175 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16176 if a DER encoded private key is RSA or DSA traditional format. Changed
16177 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16178 format DER encoded private key. Newer code should use PKCS#8 format which
16179 has the key type encoded in the ASN1 structure. Added DER private key
16180 support to pkcs8 application.
16184 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16185 ciphersuites has been selected (as required by the SSL 3/TLS 1
16186 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16187 is set, we interpret this as a request to violate the specification
16188 (the worst that can happen is a handshake failure, and 'correct'
16189 behaviour would result in a handshake failure anyway).
16193 * In SSL_CTX_add_session, take into account that there might be multiple
16194 SSL_SESSION structures with the same session ID (e.g. when two threads
16195 concurrently obtain them from an external cache).
16196 The internal cache can handle only one SSL_SESSION with a given ID,
16197 so if there's a conflict, we now throw out the old one to achieve
16202 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16203 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16204 some routines that use cipher OIDs: some ciphers do not have OIDs
16205 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16210 * Simplify the trust setting structure and code. Now we just have
16211 two sequences of OIDs for trusted and rejected settings. These will
16212 typically have values the same as the extended key usage extension
16213 and any application specific purposes.
16215 The trust checking code now has a default behaviour: it will just
16216 check for an object with the same NID as the passed id. Functions can
16217 be provided to override either the default behaviour or the behaviour
16218 for a given id. SSL client, server and email already have functions
16219 in place for compatibility: they check the NID and also return "trusted"
16220 if the certificate is self signed.
16224 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16225 traditional format into an EVP_PKEY structure.
16229 * Add a password callback function PEM_cb() which either prompts for
16230 a password if usr_data is NULL or otherwise assumes it is a null
16231 terminated password. Allow passwords to be passed on command line
16232 environment or config files in a few more utilities.
16236 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16237 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16238 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16239 Update documentation.
16243 * Support for ASN1 "NULL" type. This could be handled before by using
16244 ASN1_TYPE but there wasn't any function that would try to read a NULL
16245 and produce an error if it couldn't. For compatibility we also have
16246 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16247 don't allocate anything because they don't need to.
16251 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16254 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16256 * Rebuild of the memory allocation routines used by OpenSSL code and
16257 possibly others as well. The purpose is to make an interface that
16258 provide hooks so anyone can build a separate set of allocation and
16259 deallocation routines to be used by OpenSSL, for example memory
16260 pool implementations, or something else, which was previously hard
16261 since Malloc(), Realloc() and Free() were defined as macros having
16262 the values malloc, realloc and free, respectively (except for Win32
16263 compilations). The same is provided for memory debugging code.
16264 OpenSSL already comes with functionality to find memory leaks, but
16265 this gives people a chance to debug other memory problems.
16267 With these changes, a new set of functions and macros have appeared:
16269 CRYPTO_set_mem_debug_functions() [F]
16270 CRYPTO_get_mem_debug_functions() [F]
16271 CRYPTO_dbg_set_options() [F]
16272 CRYPTO_dbg_get_options() [F]
16273 CRYPTO_malloc_debug_init() [M]
16275 The memory debug functions are NULL by default, unless the library
16276 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16277 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16278 gives the standard debugging functions that come with OpenSSL) or
16279 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16280 provided by the library user) must be used. When the standard
16281 debugging functions are used, CRYPTO_dbg_set_options can be used to
16282 request additional information:
16283 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16284 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16286 Also, things like CRYPTO_set_mem_functions will always give the
16287 expected result (the new set of functions is used for allocation
16288 and deallocation) at all times, regardless of platform and compiler
16291 To finish it up, some functions that were never use in any other
16292 way than through macros have a new API and new semantic:
16294 CRYPTO_dbg_malloc()
16295 CRYPTO_dbg_realloc()
16298 All macros of value have retained their old syntax.
16300 *Richard Levitte and Bodo Moeller*
16302 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16303 ordering of SMIMECapabilities wasn't in "strength order" and there
16304 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16309 * Some ASN1 types with illegal zero length encoding (INTEGER,
16310 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16312 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16314 * Merge in my S/MIME library for OpenSSL. This provides a simple
16315 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16316 functionality to handle multipart/signed properly) and a utility
16317 called 'smime' to call all this stuff. This is based on code I
16318 originally wrote for Celo who have kindly allowed it to be
16319 included in OpenSSL.
16323 * Add variants des_set_key_checked and des_set_key_unchecked of
16324 des_set_key (aka des_key_sched). Global variable des_check_key
16325 decides which of these is called by des_set_key; this way
16326 des_check_key behaves as it always did, but applications and
16327 the library itself, which was buggy for des_check_key == 1,
16328 have a cleaner way to pick the version they need.
16332 * New function PKCS12_newpass() which changes the password of a
16337 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16338 dynamic mix. In both cases the ids can be used as an index into the
16339 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16340 functions so they accept a list of the field values and the
16341 application doesn't need to directly manipulate the X509_TRUST
16346 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16351 * Modify the way the V3 extension code looks up extensions. This now
16352 works in a similar way to the object code: we have some "standard"
16353 extensions in a static table which is searched with OBJ_bsearch()
16354 and the application can add dynamic ones if needed. The file
16355 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16356 updated whenever a new extension is added to the core code and kept
16357 in ext_nid order. There is a simple program 'tabtest.c' which checks
16358 this. New extensions are not added too often so this file can readily
16359 be maintained manually.
16361 There are two big advantages in doing things this way. The extensions
16362 can be looked up immediately and no longer need to be "added" using
16363 X509V3_add_standard_extensions(): this function now does nothing.
16364 Side note: I get *lots* of email saying the extension code doesn't
16365 work because people forget to call this function.
16366 Also no dynamic allocation is done unless new extensions are added:
16367 so if we don't add custom extensions there is no need to call
16368 X509V3_EXT_cleanup().
16372 * Modify enc utility's salting as follows: make salting the default. Add a
16373 magic header, so unsalted files fail gracefully instead of just decrypting
16374 to garbage. This is because not salting is a big security hole, so people
16375 should be discouraged from doing it.
16379 * Fixes and enhancements to the 'x509' utility. It allowed a message
16380 digest to be passed on the command line but it only used this
16381 parameter when signing a certificate. Modified so all relevant
16382 operations are affected by the digest parameter including the
16383 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16384 DSA key was used because it didn't fix the digest.
16388 * Initial certificate chain verify code. Currently tests the untrusted
16389 certificates for consistency with the verify purpose (which is set
16390 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16392 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16393 this is because it will reject chains with invalid extensions whereas
16394 every previous version of OpenSSL and SSLeay made no checks at all.
16396 Trust code: checks the root CA for the relevant trust settings. Trust
16397 settings have an initial value consistent with the verify purpose: e.g.
16398 if the verify purpose is for SSL client use it expects the CA to be
16399 trusted for SSL client use. However the default value can be changed to
16400 permit custom trust settings: one example of this would be to only trust
16401 certificates from a specific "secure" set of CAs.
16403 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16404 which should be used for version portability: especially since the
16405 verify structure is likely to change more often now.
16407 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16408 to set them. If not set then assume SSL clients will verify SSL servers
16411 Two new options to the verify program: -untrusted allows a set of
16412 untrusted certificates to be passed in and -purpose which sets the
16413 intended purpose of the certificate. If a purpose is set then the
16414 new chain verify code is used to check extension consistency.
16418 * Support for the authority information access extension.
16422 * Modify RSA and DSA PEM read routines to transparently handle
16423 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16424 public keys in a format compatible with certificate
16425 SubjectPublicKeyInfo structures. Unfortunately there were already
16426 functions called *_PublicKey_* which used various odd formats so
16427 these are retained for compatibility: however the DSA variants were
16428 never in a public release so they have been deleted. Changed dsa/rsa
16429 utilities to handle the new format: note no releases ever handled public
16430 keys so we should be OK.
16432 The primary motivation for this change is to avoid the same fiasco
16433 that dogs private keys: there are several incompatible private key
16434 formats some of which are standard and some OpenSSL specific and
16435 require various evil hacks to allow partial transparent handling and
16436 even then it doesn't work with DER formats. Given the option anything
16437 other than PKCS#8 should be dumped: but the other formats have to
16438 stay in the name of compatibility.
16440 With public keys and the benefit of hindsight one standard format
16441 is used which works with EVP_PKEY, RSA or DSA structures: though
16442 it clearly returns an error if you try to read the wrong kind of key.
16444 Added a -pubkey option to the 'x509' utility to output the public key.
16445 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16446 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16447 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16448 that do the same as the `EVP_PKEY_assign_*()` except they up the
16449 reference count of the added key (they don't "swallow" the
16454 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16455 CRLs would fail if the file contained no certificates or no CRLs:
16456 added a new function to read in both types and return the number
16457 read: this means that if none are read it will be an error. The
16458 DER versions of the certificate and CRL reader would always fail
16459 because it isn't possible to mix certificates and CRLs in DER format
16460 without choking one or the other routine. Changed this to just read
16461 a certificate: this is the best we can do. Also modified the code
16462 in `apps/verify.c` to take notice of return codes: it was previously
16463 attempting to read in certificates from NULL pointers and ignoring
16464 any errors: this is one reason why the cert and CRL reader seemed
16465 to work. It doesn't check return codes from the default certificate
16466 routines: these may well fail if the certificates aren't installed.
16470 * Code to support otherName option in GeneralName.
16474 * First update to verify code. Change the verify utility
16475 so it warns if it is passed a self signed certificate:
16476 for consistency with the normal behaviour. X509_verify
16477 has been modified to it will now verify a self signed
16478 certificate if *exactly* the same certificate appears
16479 in the store: it was previously impossible to trust a
16480 single self signed certificate. This means that:
16481 openssl verify ss.pem
16482 now gives a warning about a self signed certificate but
16483 openssl verify -CAfile ss.pem ss.pem
16488 * For servers, store verify_result in SSL_SESSION data structure
16489 (and add it to external session representation).
16490 This is needed when client certificate verifications fails,
16491 but an application-provided verification callback (set by
16492 SSL_CTX_set_cert_verify_callback) allows accepting the session
16493 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16494 but returns 1): When the session is reused, we have to set
16495 ssl->verify_result to the appropriate error code to avoid
16498 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16500 * Fix a bug in the new PKCS#7 code: it didn't consider the
16501 case in PKCS7_dataInit() where the signed PKCS7 structure
16502 didn't contain any existing data because it was being created.
16504 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16506 * Add a salt to the key derivation routines in enc.c. This
16507 forms the first 8 bytes of the encrypted file. Also add a
16508 -S option to allow a salt to be input on the command line.
16512 * New function X509_cmp(). Oddly enough there wasn't a function
16513 to compare two certificates. We do this by working out the SHA1
16514 hash and comparing that. X509_cmp() will be needed by the trust
16519 * SSL_get1_session() is like SSL_get_session(), but increments
16520 the reference count in the SSL_SESSION returned.
16522 *Geoff Thorpe <geoff@eu.c2.net>*
16524 * Fix for 'req': it was adding a null to request attributes.
16525 Also change the X509_LOOKUP and X509_INFO code to handle
16526 certificate auxiliary information.
16530 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16535 * Add the possibility to add extra information to the memory leak
16536 detecting output, to form tracebacks, showing from where each
16537 allocation was originated: CRYPTO_push_info("constant string") adds
16538 the string plus current file name and line number to a per-thread
16539 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16540 is like calling CYRPTO_pop_info() until the stack is empty.
16541 Also updated memory leak detection code to be multi-thread-safe.
16545 * Add options -text and -noout to pkcs7 utility and delete the
16546 encryption options which never did anything. Update docs.
16550 * Add options to some of the utilities to allow the pass phrase
16551 to be included on either the command line (not recommended on
16552 OSes like Unix) or read from the environment. Update the
16553 manpages and fix a few bugs.
16557 * Add a few manpages for some of the openssl commands.
16561 * Fix the -revoke option in ca. It was freeing up memory twice,
16562 leaking and not finding already revoked certificates.
16566 * Extensive changes to support certificate auxiliary information.
16567 This involves the use of X509_CERT_AUX structure and X509_AUX
16568 functions. An X509_AUX function such as PEM_read_X509_AUX()
16569 can still read in a certificate file in the usual way but it
16570 will also read in any additional "auxiliary information". By
16571 doing things this way a fair degree of compatibility can be
16572 retained: existing certificates can have this information added
16573 using the new 'x509' options.
16575 Current auxiliary information includes an "alias" and some trust
16576 settings. The trust settings will ultimately be used in enhanced
16577 certificate chain verification routines: currently a certificate
16578 can only be trusted if it is self signed and then it is trusted
16583 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16584 The problem was that one of the replacement routines had not been working
16585 since SSLeay releases. For now the offending routine has been replaced
16586 with non-optimised assembler. Even so, this now gives around 95%
16587 performance improvement for 1024 bit RSA signs.
16591 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16592 handling. Most clients have the effective key size in bits equal to
16593 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16594 A few however don't do this and instead use the size of the decrypted key
16595 to determine the RC2 key length and the AlgorithmIdentifier to determine
16596 the effective key length. In this case the effective key length can still
16597 be 40 bits but the key length can be 168 bits for example. This is fixed
16598 by manually forcing an RC2 key into the EVP_PKEY structure because the
16599 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16600 the key length and effective key length are equal.
16604 * Add a bunch of functions that should simplify the creation of
16605 X509_NAME structures. Now you should be able to do:
16606 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16607 and have it automatically work out the correct field type and fill in
16608 the structures. The more adventurous can try:
16609 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16610 and it will (hopefully) work out the correct multibyte encoding.
16614 * Change the 'req' utility to use the new field handling and multibyte
16615 copy routines. Before the DN field creation was handled in an ad hoc
16616 way in req, ca, and x509 which was rather broken and didn't support
16617 BMPStrings or UTF8Strings. Since some software doesn't implement
16618 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16619 using the dirstring_type option. See the new comment in the default
16620 openssl.cnf for more info.
16624 * Make crypto/rand/md_rand.c more robust:
16625 - Assure unique random numbers after fork().
16626 - Make sure that concurrent threads access the global counter and
16627 md serializably so that we never lose entropy in them
16628 or use exactly the same state in multiple threads.
16629 Access to the large state is not always serializable because
16630 the additional locking could be a performance killer, and
16631 md should be large enough anyway.
16635 * New file `apps/app_rand.c` with commonly needed functionality
16636 for handling the random seed file.
16638 Use the random seed file in some applications that previously did not:
16640 dsaparam -genkey (which also ignored its '-rand' option),
16643 x509 (when signing).
16644 Except on systems with /dev/urandom, it is crucial to have a random
16645 seed file at least for key creation, DSA signing, and for DH exchanges;
16646 for RSA signatures we could do without one.
16648 gendh and gendsa (unlike genrsa) used to read only the first byte
16649 of each file listed in the '-rand' option. The function as previously
16650 found in genrsa is now in app_rand.c and is used by all programs
16651 that support '-rand'.
16655 * In RAND_write_file, use mode 0600 for creating files;
16656 don't just chmod when it may be too late.
16660 * Report an error from X509_STORE_load_locations
16661 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16665 * New function ASN1_mbstring_copy() this copies a string in either
16666 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16667 into an ASN1_STRING type. A mask of permissible types is passed
16668 and it chooses the "minimal" type to use or an error if not type
16673 * Add function equivalents to the various macros in asn1.h. The old
16674 macros are retained with an `M_` prefix. Code inside the library can
16675 use the `M_` macros. External code (including the openssl utility)
16676 should *NOT* in order to be "shared library friendly".
16680 * Add various functions that can check a certificate's extensions
16681 to see if it usable for various purposes such as SSL client,
16682 server or S/MIME and CAs of these types. This is currently
16683 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16684 verification. Also added a -purpose flag to x509 utility to
16685 print out all the purposes.
16689 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16694 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16695 for, obtain and decode and extension and obtain its critical flag.
16696 This allows all the necessary extension code to be handled in a
16697 single function call.
16701 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16702 platforms. See crypto/rc4/rc4_enc.c for further details.
16706 * New -noout option to asn1parse. This causes no output to be produced
16707 its main use is when combined with -strparse and -out to extract data
16708 from a file (which may not be in ASN.1 format).
16712 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16713 when producing the local key id.
16715 *Richard Levitte <levitte@stacken.kth.se>*
16717 * New option -dhparam in s_server. This allows a DH parameter file to be
16718 stated explicitly. If it is not stated then it tries the first server
16719 certificate file. The previous behaviour hard coded the filename
16724 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16725 a public key to be input or output. For example:
16726 openssl rsa -in key.pem -pubout -out pubkey.pem
16727 Also added necessary DSA public key functions to handle this.
16731 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16732 in the message. This was handled by allowing
16733 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16735 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16737 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16738 to the end of the strings whereas this didn't. This would cause problems
16739 if strings read with d2i_ASN1_bytes() were later modified.
16741 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16743 * Fix for base64 decode bug. When a base64 bio reads only one line of
16744 data and it contains EOF it will end up returning an error. This is
16745 caused by input 46 bytes long. The cause is due to the way base64
16746 BIOs find the start of base64 encoded data. They do this by trying a
16747 trial decode on each line until they find one that works. When they
16748 do a flag is set and it starts again knowing it can pass all the
16749 data directly through the decoder. Unfortunately it doesn't reset
16750 the context it uses. This means that if EOF is reached an attempt
16751 is made to pass two EOFs through the context and this causes the
16752 resulting error. This can also cause other problems as well. As is
16753 usual with these problems it takes *ages* to find and the fix is
16754 trivial: move one line.
16756 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16758 * Ugly workaround to get s_client and s_server working under Windows. The
16759 old code wouldn't work because it needed to select() on sockets and the
16760 tty (for keypresses and to see if data could be written). Win32 only
16761 supports select() on sockets so we select() with a 1s timeout on the
16762 sockets and then see if any characters are waiting to be read, if none
16763 are present then we retry, we also assume we can always write data to
16764 the tty. This isn't nice because the code then blocks until we've
16765 received a complete line of data and it is effectively polling the
16766 keyboard at 1s intervals: however it's quite a bit better than not
16767 working at all :-) A dedicated Windows application might handle this
16768 with an event loop for example.
16772 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16773 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16774 will be called when RSA_sign() and RSA_verify() are used. This is useful
16775 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16776 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16777 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16778 This necessitated the support of an extra signature type NID_md5_sha1
16779 for SSL signatures and modifications to the SSL library to use it instead
16780 of calling RSA_public_decrypt() and RSA_private_encrypt().
16784 * Add new -verify -CAfile and -CApath options to the crl program, these
16785 will lookup a CRL issuers certificate and verify the signature in a
16786 similar way to the verify program. Tidy up the crl program so it
16787 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16788 less strict. It will now permit CRL extensions even if it is not
16789 a V2 CRL: this will allow it to tolerate some broken CRLs.
16793 * Initialize all non-automatic variables each time one of the openssl
16794 sub-programs is started (this is necessary as they may be started
16795 multiple times from the "OpenSSL>" prompt).
16797 *Lennart Bang, Bodo Moeller*
16799 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16800 removing all other RSA functionality (this is what NO_RSA does). This
16801 is so (for example) those in the US can disable those operations covered
16802 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16807 * Non-copying interface to BIO pairs.
16808 (still largely untested)
16812 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16813 ASCII string. This was handled independently in various places before.
16817 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16818 UTF8 strings a character at a time.
16822 * Use client_version from client hello to select the protocol
16823 (s23_srvr.c) and for RSA client key exchange verification
16824 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16828 * Add various utility functions to handle SPKACs, these were previously
16829 handled by poking round in the structure internals. Added new function
16830 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16831 print, verify and generate SPKACs. Based on an original idea from
16832 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16836 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16840 * Allow the config file extension section to be overwritten on the
16841 command line. Based on an original idea from Massimiliano Pala
16842 <madwolf@comune.modena.it>. The new option is called -extensions
16843 and can be applied to ca, req and x509. Also -reqexts to override
16844 the request extensions in req and -crlexts to override the crl extensions
16849 * Add new feature to the SPKAC handling in ca. Now you can include
16850 the same field multiple times by preceding it by "XXXX." for example:
16853 this is the same syntax as used in the req config file.
16857 * Allow certificate extensions to be added to certificate requests. These
16858 are specified in a 'req_extensions' option of the req section of the
16859 config file. They can be printed out with the -text option to req but
16860 are otherwise ignored at present.
16864 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16865 data read consists of only the final block it would not decrypted because
16866 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16867 A misplaced 'break' also meant the decrypted final block might not be
16868 copied until the next read.
16872 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16873 a few extra parameters to the DH structure: these will be useful if
16874 for example we want the value of 'q' or implement X9.42 DH.
16878 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16879 provides hooks that allow the default DSA functions or functions on a
16880 "per key" basis to be replaced. This allows hardware acceleration and
16881 hardware key storage to be handled without major modification to the
16882 library. Also added low level modexp hooks and CRYPTO_EX structure and
16883 associated functions.
16887 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16888 as "read only": it can't be written to and the buffer it points to will
16889 not be freed. Reading from a read only BIO is much more efficient than
16890 a normal memory BIO. This was added because there are several times when
16891 an area of memory needs to be read from a BIO. The previous method was
16892 to create a memory BIO and write the data to it, this results in two
16893 copies of the data and an O(n^2) reading algorithm. There is a new
16894 function BIO_new_mem_buf() which creates a read only memory BIO from
16895 an area of memory. Also modified the PKCS#7 routines to use read only
16900 * Bugfix: ssl23_get_client_hello did not work properly when called in
16901 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16902 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16903 but a retry condition occurred while trying to read the rest.
16907 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16908 NID_pkcs7_encrypted by default: this was wrong since this should almost
16909 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16910 the encrypted data type: this is a more sensible place to put it and it
16911 allows the PKCS#12 code to be tidied up that duplicated this
16916 * Changed obj_dat.pl script so it takes its input and output files on
16917 the command line. This should avoid shell escape redirection problems
16922 * Initial support for certificate extension requests, these are included
16923 in things like Xenroll certificate requests. Included functions to allow
16924 extensions to be obtained and added.
16928 * -crlf option to s_client and s_server for sending newlines as
16929 CRLF (as required by many protocols).
16933 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16935 * Install libRSAglue.a when OpenSSL is built with RSAref.
16937 *Ralf S. Engelschall*
16939 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16941 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16943 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16948 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16949 DH parameters/keys (q is lost during that conversion, but the resulting
16950 DH parameters contain its length).
16952 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16953 much faster than DH_generate_parameters (which creates parameters
16954 where `p = 2*q + 1`), and also the smaller q makes DH computations
16955 much more efficient (160-bit exponentiation instead of 1024-bit
16956 exponentiation); so this provides a convenient way to support DHE
16957 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16958 utter importance to use
16959 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16961 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16962 when such DH parameters are used, because otherwise small subgroup
16963 attacks may become possible!
16967 * Avoid memory leak in i2d_DHparams.
16971 * Allow the -k option to be used more than once in the enc program:
16972 this allows the same encrypted message to be read by multiple recipients.
16976 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16977 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16978 it will always use the numerical form of the OID, even if it has a short
16983 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16984 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16985 otherwise bn_mod_exp was called. In the case of hardware keys for example
16986 no private key components need be present and it might store extra data
16987 in the RSA structure, which cannot be accessed from bn_mod_exp.
16988 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16989 private key operations.
16993 * Added support for SPARC Linux.
16997 * pem_password_cb function type incompatibly changed from
16998 typedef int pem_password_cb(char *buf, int size, int rwflag);
17000 ....(char *buf, int size, int rwflag, void *userdata);
17001 so that applications can pass data to their callbacks:
17002 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17003 additional void * argument, which is just handed through whenever
17004 the password callback is called.
17006 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17008 New function SSL_CTX_set_default_passwd_cb_userdata.
17010 Compatibility note: As many C implementations push function arguments
17011 onto the stack in reverse order, the new library version is likely to
17012 interoperate with programs that have been compiled with the old
17013 pem_password_cb definition (PEM_whatever takes some data that
17014 happens to be on the stack as its last argument, and the callback
17015 just ignores this garbage); but there is no guarantee whatsoever that
17018 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17019 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17020 problems not only on Windows, but also on some Unix platforms.
17021 To avoid problematic command lines, these definitions are now in an
17022 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17023 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17027 * MIPS III/IV assembler module is reimplemented.
17031 * More DES library cleanups: remove references to srand/rand and
17032 delete an unused file.
17036 * Add support for the free Netwide assembler (NASM) under Win32,
17037 since not many people have MASM (ml) and it can be hard to obtain.
17038 This is currently experimental but it seems to work OK and pass all
17039 the tests. Check out INSTALL.W32 for info.
17043 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17044 without temporary keys kept an extra copy of the server key,
17045 and connections with temporary keys did not free everything in case
17050 * New function RSA_check_key and new openssl rsa option -check
17051 for verifying the consistency of RSA keys.
17053 *Ulf Moeller, Bodo Moeller*
17055 * Various changes to make Win32 compile work:
17056 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17057 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17058 comparison" warnings.
17059 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17063 * Add a debugging option to PKCS#5 v2 key generation function: when
17064 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17065 derived keys are printed to stderr.
17069 * Copy the flags in ASN1_STRING_dup().
17071 *Roman E. Pavlov <pre@mo.msk.ru>*
17073 * The x509 application mishandled signing requests containing DSA
17074 keys when the signing key was also DSA and the parameters didn't match.
17076 It was supposed to omit the parameters when they matched the signing key:
17077 the verifying software was then supposed to automatically use the CA's
17078 parameters if they were absent from the end user certificate.
17080 Omitting parameters is no longer recommended. The test was also
17081 the wrong way round! This was probably due to unusual behaviour in
17082 EVP_cmp_parameters() which returns 1 if the parameters match.
17083 This meant that parameters were omitted when they *didn't* match and
17084 the certificate was useless. Certificates signed with 'ca' didn't have
17087 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17089 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17090 The interface is as follows:
17091 Applications can use
17092 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17093 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17094 "off" is now the default.
17095 The library internally uses
17096 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17097 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17098 to disable memory-checking temporarily.
17100 Some inconsistent states that previously were possible (and were
17101 even the default) are now avoided.
17103 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17104 with each memory chunk allocated; this is occasionally more helpful
17105 than just having a counter.
17107 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17109 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17114 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17115 which largely parallels "options", but is for changing API behaviour,
17116 whereas "options" are about protocol behaviour.
17117 Initial "mode" flags are:
17119 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17120 a single record has been written.
17121 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17122 retries use the same buffer location.
17123 (But all of the contents must be
17128 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17131 * Fix problems with no-hmac etc.
17133 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17135 * New functions RSA_get_default_method(), RSA_set_method() and
17136 RSA_get_method(). These allows replacement of RSA_METHODs without having
17137 to mess around with the internals of an RSA structure.
17141 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17142 Also really enable memory leak checks in openssl.c and in some
17145 *Chad C. Mulligan, Bodo Moeller*
17147 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17148 up the length of negative integers. This has now been simplified to just
17149 store the length when it is first determined and use it later, rather
17150 than trying to keep track of where data is copied and updating it to
17152 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17154 * Add a new function PKCS7_signatureVerify. This allows the verification
17155 of a PKCS#7 signature but with the signing certificate passed to the
17156 function itself. This contrasts with PKCS7_dataVerify which assumes the
17157 certificate is present in the PKCS#7 structure. This isn't always the
17158 case: certificates can be omitted from a PKCS#7 structure and be
17159 distributed by "out of band" means (such as a certificate database).
17163 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17164 function prototypes in pem.h, also change util/mkdef.pl to add the
17165 necessary function names.
17169 * mk1mf.pl (used by Windows builds) did not properly read the
17170 options set by Configure in the top level Makefile, and Configure
17171 was not even able to write more than one option correctly.
17172 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17176 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17177 file to be loaded from a BIO or FILE pointer. The BIO version will
17178 for example allow memory BIOs to contain config info.
17182 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17183 Whoever hopes to achieve shared-library compatibility across versions
17184 must use this, not the compile-time macro.
17185 (Exercise 0.9.4: Which is the minimum library version required by
17187 Note: All this applies only to multi-threaded programs, others don't
17192 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17193 through a BIO pair triggered the default case, i.e.
17194 SSLerr(...,SSL_R_UNKNOWN_STATE).
17198 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17199 can use the SSL library even if none of the specific BIOs is
17204 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17205 for the encoded length.
17207 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17209 * Add initial documentation of the X509V3 functions.
17213 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17214 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17215 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17216 secure PKCS#8 private key format with a high iteration count.
17220 * Fix determination of Perl interpreter: A perl or perl5
17221 *directory* in $PATH was also accepted as the interpreter.
17223 *Ralf S. Engelschall*
17225 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17226 wrong with it but it was very old and did things like calling
17227 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17228 unusual formatting.
17232 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17233 to use the new extension code.
17237 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17238 with macros. This should make it easier to change their form, add extra
17239 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17244 * Add to configuration table a new entry that can specify an alternative
17245 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17246 according to Mark Crispin <MRC@Panda.COM>.
17250 * DES CBC did not update the IV. Weird.
17254 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17255 Changing the behaviour of the former might break existing programs --
17256 where IV updating is needed, des_ncbc_encrypt can be used.
17259 * When bntest is run from "make test" it drives bc to check its
17260 calculations, as well as internally checking them. If an internal check
17261 fails, it needs to cause bc to give a non-zero result or make test carries
17262 on without noticing the failure. Fixed.
17266 * DES library cleanups.
17270 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17271 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17272 ciphers. NOTE: although the key derivation function has been verified
17273 against some published test vectors it has not been extensively tested
17274 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17279 * Instead of "mkdir -p", which is not fully portable, use new
17280 Perl script "util/mkdir-p.pl".
17284 * Rewrite the way password based encryption (PBE) is handled. It used to
17285 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17286 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17287 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17288 the 'parameter' field of the AlgorithmIdentifier is passed to the
17289 underlying key generation function so it must do its own ASN1 parsing.
17290 This has also changed the EVP_PBE_CipherInit() function which now has a
17291 'parameter' argument instead of literal salt and iteration count values
17292 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17296 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17297 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17298 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17299 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17300 value was just used as a "magic string" and not used directly its
17301 value doesn't matter.
17305 * Introduce some semblance of const correctness to BN. Shame C doesn't
17310 * "linux-sparc64" configuration (ultrapenguin).
17312 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17313 "linux-sparc" configuration.
17315 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17317 * config now generates no-xxx options for missing ciphers.
17321 * Support the EBCDIC character set (work in progress).
17322 File ebcdic.c not yet included because it has a different license.
17324 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17326 * Support BS2000/OSD-POSIX.
17328 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17330 * Make callbacks for key generation use `void *` instead of `char *`.
17334 * Make S/MIME samples compile (not yet tested).
17338 * Additional typesafe stacks.
17342 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17346 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17348 * New configuration variant "sco5-gcc".
17350 * Updated some demos.
17352 *Sean O Riordain, Wade Scholine*
17354 * Add missing BIO_free at exit of pkcs12 application.
17358 * Fix memory leak in conf.c.
17362 * Updates for Win32 to assembler version of MD5.
17366 * Set #! path to perl in `apps/der_chop` to where we found it
17367 instead of using a fixed path.
17371 * SHA library changes for irix64-mips4-cc.
17375 * Improvements for VMS support.
17379 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17381 * Bignum library bug fix. IRIX 6 passes "make test" now!
17382 This also avoids the problems with SC4.2 and unpatched SC5.
17384 *Andy Polyakov <appro@fy.chalmers.se>*
17386 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17387 These are required because of the typesafe stack would otherwise break
17388 existing code. If old code used a structure member which used to be STACK
17389 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17390 sk_num or sk_value it would produce an error because the num, data members
17391 are not present in STACK_OF. Now it just produces a warning. sk_set
17392 replaces the old method of assigning a value to sk_value
17393 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17394 that does this will no longer work (and should use sk_set instead) but
17395 this could be regarded as a "questionable" behaviour anyway.
17399 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17400 correctly handle encrypted S/MIME data.
17404 * Change type of various DES function arguments from des_cblock
17405 (which means, in function argument declarations, pointer to char)
17406 to des_cblock * (meaning pointer to array with 8 char elements),
17407 which allows the compiler to do more typechecking; it was like
17408 that back in SSLeay, but with lots of ugly casts.
17410 Introduce new type const_des_cblock.
17414 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17415 problems: find RecipientInfo structure that matches recipient certificate
17416 and initialise the ASN1 structures properly based on passed cipher.
17420 * Belatedly make the BN tests actually check the results.
17424 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17425 to and from BNs: it was completely broken. New compilation option
17426 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17427 key elements as negative integers.
17431 * Reorganize and speed up MD5.
17433 *Andy Polyakov <appro@fy.chalmers.se>*
17437 *Richard Levitte <richard@levitte.org>*
17439 * New option -out to asn1parse to allow the parsed structure to be
17440 output to a file. This is most useful when combined with the -strparse
17441 option to examine the output of things like OCTET STRINGS.
17445 * Make SSL library a little more fool-proof by not requiring any longer
17446 that `SSL_set_{accept,connect}_state` be called before
17447 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17448 in many applications because usually everything *appeared* to work as
17449 intended anyway -- now it really works as intended).
17453 * Move openssl.cnf out of lib/.
17457 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17458 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17459 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17461 *Ralf S. Engelschall*
17463 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17464 handle PKCS#7 enveloped data properly.
17466 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17468 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17469 copying pointers. The cert_st handling is changed by this in
17470 various ways (and thus what used to be known as ctx->default_cert
17471 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17472 any longer when s->cert does not give us what we need).
17473 ssl_cert_instantiate becomes obsolete by this change.
17474 As soon as we've got the new code right (possibly it already is?),
17475 we have solved a couple of bugs of the earlier code where s->cert
17476 was used as if it could not have been shared with other SSL structures.
17478 Note that using the SSL API in certain dirty ways now will result
17479 in different behaviour than observed with earlier library versions:
17480 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17481 does not influence s as it used to.
17483 In order to clean up things more thoroughly, inside SSL_SESSION
17484 we don't use CERT any longer, but a new structure SESS_CERT
17485 that holds per-session data (if available); currently, this is
17486 the peer's certificate chain and, for clients, the server's certificate
17487 and temporary key. CERT holds only those values that can have
17488 meaningful defaults in an SSL_CTX.
17492 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17493 from the internal representation. Various PKCS#7 fixes: remove some
17494 evil casts and set the enc_dig_alg field properly based on the signing
17499 * Allow PKCS#12 password to be set from the command line or the
17500 environment. Let 'ca' get its config file name from the environment
17501 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17506 * Allow certificate policies extension to use an IA5STRING for the
17507 organization field. This is contrary to the PKIX definition but
17508 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17513 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17514 without disallowing inline assembler and the like for non-pedantic builds.
17518 * Support Borland C++ builder.
17520 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17526 * SHA-1 cleanups and performance enhancements.
17528 *Andy Polyakov <appro@fy.chalmers.se>*
17530 * Sparc v8plus assembler for the bignum library.
17532 *Andy Polyakov <appro@fy.chalmers.se>*
17534 * Accept any -xxx and +xxx compiler options in Configure.
17538 * Update HPUX configuration.
17542 * Add missing `sk_<type>_unshift()` function to safestack.h
17544 *Ralf S. Engelschall*
17546 * New function SSL_CTX_use_certificate_chain_file that sets the
17547 "extra_cert"s in addition to the certificate. (This makes sense
17548 only for "PEM" format files, as chains as a whole are not
17553 * Support verify_depth from the SSL API.
17554 x509_vfy.c had what can be considered an off-by-one-error:
17555 Its depth (which was not part of the external interface)
17556 was actually counting the number of certificates in a chain;
17557 now it really counts the depth.
17561 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17562 instead of X509err, which often resulted in confusing error
17563 messages since the error codes are not globally unique
17564 (e.g. an alleged error in ssl3_accept when a certificate
17565 didn't match the private key).
17567 * New function SSL_CTX_set_session_id_context that allows to set a default
17568 value (so that you don't need SSL_set_session_id_context for each
17569 connection using the SSL_CTX).
17573 * OAEP decoding bug fix.
17577 * Support INSTALL_PREFIX for package builders, as proposed by
17582 * New Configure options "threads" and "no-threads". For systems
17583 where the proper compiler options are known (currently Solaris
17584 and Linux), "threads" is the default.
17588 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17592 * Install various scripts to $(OPENSSLDIR)/misc, not to
17593 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17594 such as /usr/local/bin.
17598 * "make linux-shared" to build shared libraries.
17600 *Niels Poppe <niels@netbox.org>*
17602 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17606 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17607 extension adding in x509 utility.
17611 * Remove NOPROTO sections and error code comments.
17615 * Partial rewrite of the DEF file generator to now parse the ANSI
17620 * New Configure options --prefix=DIR and --openssldir=DIR.
17624 * Complete rewrite of the error code script(s). It is all now handled
17625 by one script at the top level which handles error code gathering,
17626 header rewriting and C source file generation. It should be much better
17627 than the old method: it now uses a modified version of Ulf's parser to
17628 read the ANSI prototypes in all header files (thus the old K&R definitions
17629 aren't needed for error creation any more) and do a better job of
17630 translating function codes into names. The old 'ASN1 error code embedded
17631 in a comment' is no longer necessary and it doesn't use .err files which
17632 have now been deleted. Also the error code call doesn't have to appear all
17633 on one line (which resulted in some large lines...).
17637 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17641 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17642 0 (which usually indicates a closed connection), but continue reading.
17646 * Fix some race conditions.
17650 * Add support for CRL distribution points extension. Add Certificate
17651 Policies and CRL distribution points documentation.
17655 * Move the autogenerated header file parts to crypto/opensslconf.h.
17659 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17660 8 of keying material. Merlin has also confirmed interop with this fix
17661 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17663 *Merlin Hughes <merlin@baltimore.ie>*
17665 * Fix lots of warnings.
17667 *Richard Levitte <levitte@stacken.kth.se>*
17669 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17670 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17672 *Richard Levitte <levitte@stacken.kth.se>*
17674 * Fix problems with sizeof(long) == 8.
17676 *Andy Polyakov <appro@fy.chalmers.se>*
17678 * Change functions to ANSI C.
17682 * Fix typos in error codes.
17684 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17686 * Remove defunct assembler files from Configure.
17690 * SPARC v8 assembler BIGNUM implementation.
17692 *Andy Polyakov <appro@fy.chalmers.se>*
17694 * Support for Certificate Policies extension: both print and set.
17695 Various additions to support the r2i method this uses.
17699 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17700 return a const string when you are expecting an allocated buffer.
17704 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17705 types DirectoryString and DisplayText.
17709 * Add code to allow r2i extensions to access the configuration database,
17710 add an LHASH database driver and add several ctx helper functions.
17714 * Fix an evil bug in bn_expand2() which caused various BN functions to
17715 fail when they extended the size of a BIGNUM.
17719 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17720 support typesafe stack.
17724 * Fix typo in SSL_[gs]et_options().
17726 *Nils Frostberg <nils@medcom.se>*
17728 * Delete various functions and files that belonged to the (now obsolete)
17729 old X509V3 handling code.
17733 * New Configure option "rsaref".
17737 * Don't auto-generate pem.h.
17741 * Introduce type-safe ASN.1 SETs.
17745 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17747 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17749 * Introduce type-safe STACKs. This will almost certainly break lots of code
17750 that links with OpenSSL (well at least cause lots of warnings), but fear
17751 not: the conversion is trivial, and it eliminates loads of evil casts. A
17752 few STACKed things have been converted already. Feel free to convert more.
17753 In the fullness of time, I'll do away with the STACK type altogether.
17757 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17758 specified in `<certfile>` by updating the entry in the index.txt file.
17759 This way one no longer has to edit the index.txt file manually for
17760 revoking a certificate. The -revoke option does the gory details now.
17762 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17764 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17765 `-text` option at all and this way the `-noout -text` combination was
17766 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17768 *Ralf S. Engelschall*
17770 * Make sure a corresponding plain text error message exists for the
17771 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17772 verify callback function determined that a certificate was revoked.
17774 *Ralf S. Engelschall*
17776 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17777 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17778 all available ciphers including rc5, which was forgotten until now.
17779 In order to let the testing shell script know which algorithms
17780 are available, a new (up to now undocumented) command
17781 `openssl list-cipher-commands` is used.
17785 * Bugfix: s_client occasionally would sleep in select() when
17786 it should have checked SSL_pending() first.
17790 * New functions DSA_do_sign and DSA_do_verify to provide access to
17791 the raw DSA values prior to ASN.1 encoding.
17795 * Tweaks to Configure
17797 *Niels Poppe <niels@netbox.org>*
17799 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17804 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17808 * New config option to avoid instructions that are illegal on the 80386.
17809 The default code is faster, but requires at least a 486.
17813 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17814 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17815 same as SSL2_VERSION anyway.
17819 * New "-showcerts" option for s_client.
17823 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17824 application. Various cleanups and fixes.
17828 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17829 modify error routines to work internally. Add error codes and PBE init
17830 to library startup routines.
17834 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17835 packing functions to asn1 and evp. Changed function names and error
17836 codes along the way.
17840 * PKCS12 integration: and so it begins... First of several patches to
17841 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17842 objects to objects.h
17846 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17847 and display support for Thawte strong extranet extension.
17851 * Add LinuxPPC support.
17853 *Jeff Dubrule <igor@pobox.org>*
17855 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17856 bn_div_words in alpha.s.
17858 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17860 * Make sure the RSA OAEP test is skipped under -DRSAref because
17861 OAEP isn't supported when OpenSSL is built with RSAref.
17863 *Ulf Moeller <ulf@fitug.de>*
17865 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17866 so they no longer are missing under -DNOPROTO.
17868 *Soren S. Jorvang <soren@t.dk>*
17870 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17872 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17873 doesn't work when the session is reused. Coming soon!
17877 * Fix a security hole, that allows sessions to be reused in the wrong
17878 context thus bypassing client cert protection! All software that uses
17879 client certs and session caches in multiple contexts NEEDS PATCHING to
17880 allow session reuse! A fuller solution is in the works.
17882 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17884 * Some more source tree cleanups (removed obsolete files
17885 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17886 permission on "config" script to be executable) and a fix for the INSTALL
17889 *Ulf Moeller <ulf@fitug.de>*
17891 * Remove some legacy and erroneous uses of malloc, free instead of
17894 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17896 * Make rsa_oaep_test return non-zero on error.
17898 *Ulf Moeller <ulf@fitug.de>*
17900 * Add support for native Solaris shared libraries. Configure
17901 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17902 if someone would make that last step automatic.
17904 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17906 * ctx_size was not built with the right compiler during "make links". Fixed.
17910 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17911 except NULL ciphers". This means the default cipher list will no longer
17912 enable NULL ciphers. They need to be specifically enabled e.g. with
17913 the string "DEFAULT:eNULL".
17917 * Fix to RSA private encryption routines: if p < q then it would
17918 occasionally produce an invalid result. This will only happen with
17919 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17923 * Be less restrictive and allow also `perl util/perlpath.pl
17924 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17925 because this way one can also use an interpreter named `perl5` (which is
17926 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17927 installed as `perl`).
17929 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17931 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17933 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17935 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17936 advapi32.lib to Win32 build and change the pem test comparison
17937 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17938 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17939 and crypto/des/ede_cbcm_enc.c.
17943 * DES quad checksum was broken on big-endian architectures. Fixed.
17947 * Comment out two functions in bio.h that aren't implemented. Fix up the
17948 Win32 test batch file so it (might) work again. The Win32 test batch file
17949 is horrible: I feel ill....
17953 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17954 in e_os.h. Audit of header files to check ANSI and non ANSI
17955 sections: 10 functions were absent from non ANSI section and not exported
17956 from Windows DLLs. Fixed up libeay.num for new functions.
17960 * Make `openssl version` output lines consistent.
17962 *Ralf S. Engelschall*
17964 * Fix Win32 symbol export lists for BIO functions: Added
17965 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17966 to ms/libeay{16,32}.def.
17968 *Ralf S. Engelschall*
17970 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17971 fine under Unix and passes some trivial tests I've now added. But the
17972 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17973 added to make sure no one expects that this stuff really works in the
17974 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17975 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17978 *Ralf S. Engelschall*
17980 * Fix the generation of two part addresses in perl.
17982 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17984 * Add config entry for Linux on MIPS.
17986 *John Tobey <jtobey@channel1.com>*
17988 * Make links whenever Configure is run, unless we are on Windoze.
17992 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17993 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17998 * Add a useful kludge to allow package maintainers to specify compiler and
17999 other platforms details on the command line without having to patch the
18000 Configure script every time: One now can use
18001 `perl Configure <id>:<details>`,
18002 i.e. platform ids are allowed to have details appended
18003 to them (separated by colons). This is treated as there would be a static
18004 pre-configured entry in Configure's %table under key `<id>` with value
18005 `<details>` and `perl Configure <id>` is called. So, when you want to
18006 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18007 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18008 now, which overrides the FreeBSD-elf entry on-the-fly.
18010 *Ralf S. Engelschall*
18012 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18016 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18017 on the `perl Configure ...` command line. This way one can compile
18018 OpenSSL libraries with Position Independent Code (PIC) which is needed
18019 for linking it into DSOs.
18021 *Ralf S. Engelschall*
18023 * Remarkably, export ciphers were totally broken and no-one had noticed!
18028 * Cleaned up the LICENSE document: The official contact for any license
18029 questions now is the OpenSSL core team under openssl-core@openssl.org.
18030 And add a paragraph about the dual-license situation to make sure people
18031 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18032 to the OpenSSL toolkit.
18034 *Ralf S. Engelschall*
18036 * General source tree makefile cleanups: Made `making xxx in yyy...`
18037 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18038 Additionally cleaned up the `make links` target: Remove unnecessary
18039 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18040 to speed processing and no longer clutter the display with confusing
18041 stuff. Instead only the actually done links are displayed.
18043 *Ralf S. Engelschall*
18045 * Permit null encryption ciphersuites, used for authentication only. It used
18046 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18047 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18052 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18053 signed attributes when verifying signatures (this would break them),
18054 the detached data encoding was wrong and public keys obtained using
18055 X509_get_pubkey() weren't freed.
18059 * Add text documentation for the BUFFER functions. Also added a work around
18060 to a Win95 console bug. This was triggered by the password read stuff: the
18061 last character typed gets carried over to the next fread(). If you were
18062 generating a new cert request using 'req' for example then the last
18063 character of the passphrase would be CR which would then enter the first
18068 * Added the new 'Includes OpenSSL Cryptography Software' button as
18069 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18070 button and can be used by applications based on OpenSSL to show the
18071 relationship to the OpenSSL project.
18073 *Ralf S. Engelschall*
18075 * Remove confusing variables in function signatures in files
18076 ssl/ssl_lib.c and ssl/ssl.h.
18078 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18080 * Don't install bss_file.c under PREFIX/include/
18082 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18084 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18085 functions that return function pointers and has support for NT specific
18086 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18087 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18088 unsigned to signed types: this was killing the Win32 compile.
18092 * Add new certificate file to stack functions,
18093 SSL_add_dir_cert_subjects_to_stack() and
18094 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18095 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18096 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18097 This means that Apache-SSL and similar packages don't have to mess around
18098 to add as many CAs as they want to the preferred list.
18102 * Experiment with doxygen documentation. Currently only partially applied to
18104 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18105 openssl.doxy as the configuration file.
18109 * Get rid of remaining C++-style comments which strict C compilers hate.
18111 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18113 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18114 compiled in by default: it has problems with large keys.
18118 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18119 DH private keys and/or callback functions which directly correspond to
18120 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18121 is needed for applications which have to configure certificates on a
18122 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18124 For the RSA certificate situation is makes no difference, but
18125 for the DSA certificate situation this fixes the "no shared cipher"
18126 problem where the OpenSSL cipher selection procedure failed because the
18127 temporary keys were not overtaken from the context and the API provided
18128 no way to reconfigure them.
18129 The new functions now let applications reconfigure the stuff and they
18130 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18131 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18132 non-public-API function ssl_cert_instantiate() is used as a helper
18133 function and also to reduce code redundancy inside ssl_rsa.c.
18135 *Ralf S. Engelschall*
18137 * Move s_server -dcert and -dkey options out of the undocumented feature
18138 area because they are useful for the DSA situation and should be
18139 recognized by the users.
18141 *Ralf S. Engelschall*
18143 * Fix the cipher decision scheme for export ciphers: the export bits are
18144 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18145 SSL_EXP_MASK. So, the original variable has to be used instead of the
18146 already masked variable.
18148 *Richard Levitte <levitte@stacken.kth.se>*
18150 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18152 *Richard Levitte <levitte@stacken.kth.se>*
18154 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18155 from `int` to `unsigned int` because it is a length and initialized by
18156 EVP_DigestFinal() which expects an `unsigned int *`.
18158 *Richard Levitte <levitte@stacken.kth.se>*
18160 * Don't hard-code path to Perl interpreter on shebang line of Configure
18161 script. Instead use the usual Shell->Perl transition trick.
18163 *Ralf S. Engelschall*
18165 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18166 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18167 -noout -modulus` as it's already the case for `openssl rsa -noout
18168 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18169 currently the public key is printed (a decision which was already done by
18170 `openssl dsa -modulus` in the past) which serves a similar purpose.
18171 Additionally the NO_RSA no longer completely removes the whole -modulus
18172 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18175 *Ralf S. Engelschall*
18177 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18178 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18180 *Arne Ansper <arne@ats.cyber.ee>*
18182 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18183 to be added. Now both 'req' and 'ca' can use new objects defined in the
18188 * Add cool BIO that does syslog (or event log on NT).
18190 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18192 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18193 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18194 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18195 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18199 * Add preliminary config info for new extension code.
18203 * Make RSA_NO_PADDING really use no padding.
18205 *Ulf Moeller <ulf@fitug.de>*
18207 * Generate errors when private/public key check is done.
18211 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18212 for some CRL extensions and new objects added.
18216 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18217 key usage extension and fuller support for authority key id.
18221 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18222 padding method for RSA, which is recommended for new applications in PKCS
18223 #1 v2.0 (RFC 2437, October 1998).
18224 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18225 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18226 against Bleichbacher's attack on RSA.
18227 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18230 * Updates to the new SSL compression code
18232 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18234 * Fix so that the version number in the master secret, when passed
18235 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18236 (because the server will not accept higher), that the version number
18237 is 0x03,0x01, not 0x03,0x00
18239 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18241 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18242 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18243 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18247 * Support for RAW extensions where an arbitrary extension can be
18248 created by including its DER encoding. See `apps/openssl.cnf` for
18253 * Make sure latest Perl versions don't interpret some generated C array
18254 code as Perl array code in the crypto/err/err_genc.pl script.
18256 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18258 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18259 not many people have the assembler. Various Win32 compilation fixes and
18260 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18261 build instructions.
18265 * Modify configure script 'Configure' to automatically create crypto/date.h
18266 file under Win32 and also build pem.h from pem.org. New script
18267 util/mkfiles.pl to create the MINFO file on environments that can't do a
18268 'make files': perl util/mkfiles.pl >MINFO should work.
18272 * Major rework of DES function declarations, in the pursuit of correctness
18273 and purity. As a result, many evil casts evaporated, and some weirdness,
18274 too. You may find this causes warnings in your code. Zapping your evil
18275 casts will probably fix them. Mostly.
18279 * Fix for a typo in asn1.h. Bug fix to object creation script
18280 obj_dat.pl. It considered a zero in an object definition to mean
18281 "end of object": none of the objects in objects.h have any zeros
18282 so it wasn't spotted.
18284 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18286 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18287 Masking (CBCM). In the absence of test vectors, the best I have been able
18288 to do is check that the decrypt undoes the encrypt, so far. Send me test
18289 vectors if you have them.
18293 * Correct calculation of key length for export ciphers (too much space was
18294 allocated for null ciphers). This has not been tested!
18298 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18299 message is now correct (it understands "crypto" and "ssl" on its
18300 command line). There is also now an "update" option. This will update
18301 the util/ssleay.num and util/libeay.num files with any new functions.
18303 perl util/mkdef.pl crypto ssl update
18304 it will update them.
18308 * Overhauled the Perl interface:
18309 - ported BN stuff to OpenSSL's different BN library
18310 - made the perl/ source tree CVS-aware
18311 - renamed the package from SSLeay to OpenSSL (the files still contain
18312 their history because I've copied them in the repository)
18313 - removed obsolete files (the test scripts will be replaced
18314 by better Test::Harness variants in the future)
18316 *Ralf S. Engelschall*
18318 * First cut for a very conservative source tree cleanup:
18319 1. merge various obsolete readme texts into doc/ssleay.txt
18320 where we collect the old documents and readme texts.
18321 2. remove the first part of files where I'm already sure that we no
18322 longer need them because of three reasons: either they are just temporary
18323 files which were left by Eric or they are preserved original files where
18324 I've verified that the diff is also available in the CVS via "cvs diff
18325 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18326 the crypto/md/ stuff).
18328 *Ralf S. Engelschall*
18330 * More extension code. Incomplete support for subject and issuer alt
18331 name, issuer and authority key id. Change the i2v function parameters
18332 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18333 what that's for :-) Fix to ASN1 macro which messed up
18334 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18338 * Preliminary support for ENUMERATED type. This is largely copied from the
18343 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18345 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18347 * Make sure `make rehash` target really finds the `openssl` program.
18349 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18351 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18352 like to hear about it if this slows down other processors.
18356 * Add CygWin32 platform information to Configure script.
18358 *Alan Batie <batie@aahz.jf.intel.com>*
18360 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18362 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18364 * New program nseq to manipulate netscape certificate sequences
18368 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18373 * Fixes to BN code. Previously the default was to define BN_RECURSION
18374 but the BN code had some problems that would cause failures when
18375 doing certificate verification and some other functions.
18377 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18379 * Add ASN1 and PEM code to support netscape certificate sequences.
18383 * Add ASN1 and PEM code to support netscape certificate sequences.
18387 * Add several PKIX and private extended key usage OIDs.
18391 * Modify the 'ca' program to handle the new extension code. Modify
18392 openssl.cnf for new extension format, add comments.
18396 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18397 and add a sample to openssl.cnf so req -x509 now adds appropriate
18402 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18403 error code, add initial support to X509_print() and x509 application.
18407 * Takes a deep breath and start adding X509 V3 extension support code. Add
18408 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18409 stuff is currently isolated and isn't even compiled yet.
18413 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18414 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18415 Removed the versions check from X509 routines when loading extensions:
18416 this allows certain broken certificates that don't set the version
18417 properly to be processed.
18421 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18422 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18423 can still be regenerated with "make depend".
18427 * Spelling mistake in C version of CAST-128.
18429 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18431 * Changes to the error generation code. The perl script err-code.pl
18432 now reads in the old error codes and retains the old numbers, only
18433 adding new ones if necessary. It also only changes the .err files if new
18434 codes are added. The makefiles have been modified to only insert errors
18435 when needed (to avoid needlessly modifying header files). This is done
18436 by only inserting errors if the .err file is newer than the auto generated
18437 C file. To rebuild all the error codes from scratch (the old behaviour)
18438 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18439 or delete all the .err files.
18443 * CAST-128 was incorrectly implemented for short keys. The C version has
18444 been fixed, but is untested. The assembler versions are also fixed, but
18445 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18446 to regenerate it if needed.
18447 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18448 Hagino <itojun@kame.net>*
18450 * File was opened incorrectly in randfile.c.
18452 *Ulf Möller <ulf@fitug.de>*
18454 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18455 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18456 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18457 al: it's just almost always a UTCTime. Note this patch adds new error
18458 codes so do a "make errors" if there are problems.
18462 * Correct Linux 1 recognition in config.
18464 *Ulf Möller <ulf@fitug.de>*
18466 * Remove pointless MD5 hash when using DSA keys in ca.
18468 *Anonymous <nobody@replay.com>*
18470 * Generate an error if given an empty string as a cert directory. Also
18471 generate an error if handed NULL (previously returned 0 to indicate an
18472 error, but didn't set one).
18474 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18476 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18480 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18481 parameters. This was causing a warning which killed off the Win32 compile.
18485 * Remove C++ style comments from crypto/bn/bn_local.h.
18487 *Neil Costigan <neil.costigan@celocom.com>*
18489 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18490 based on a text string, looking up short and long names and finally
18491 "dot" format. The "dot" format stuff didn't work. Added new function
18492 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18493 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18494 OID is not part of the table.
18498 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18499 X509_LOOKUP_by_alias().
18503 * Sort openssl functions by name.
18507 * Get the `gendsa` command working and add it to the `list` command. Remove
18508 encryption from sample DSA keys (in case anyone is interested the password
18513 * Make *all* `*_free` functions accept a NULL pointer.
18515 *Frans Heymans <fheymans@isaserver.be>*
18517 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18520 *Anonymous <nobody@replay.com>*
18522 * s_server should send the CAfile as acceptable CAs, not its own cert.
18524 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18526 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18528 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18530 * Temp key "for export" tests were wrong in s3_srvr.c.
18532 *Anonymous <nobody@replay.com>*
18534 * Add prototype for temp key callback functions
18535 SSL_CTX_set_tmp_{rsa,dh}_callback().
18539 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18540 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18544 * X509_name_add_entry() freed the wrong thing after an error.
18546 *Arne Ansper <arne@ats.cyber.ee>*
18548 * rsa_eay.c would attempt to free a NULL context.
18550 *Arne Ansper <arne@ats.cyber.ee>*
18552 * BIO_s_socket() had a broken should_retry() on Windoze.
18554 *Arne Ansper <arne@ats.cyber.ee>*
18556 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18558 *Arne Ansper <arne@ats.cyber.ee>*
18560 * Make sure the already existing X509_STORE->depth variable is initialized
18561 in X509_STORE_new(), but document the fact that this variable is still
18562 unused in the certificate verification process.
18564 *Ralf S. Engelschall*
18566 * Fix the various library and `apps/` files to free up pkeys obtained from
18567 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18571 * Fix reference counting in X509_PUBKEY_get(). This makes
18572 demos/maurice/example2.c work, amongst others, probably.
18574 *Steve Henson and Ben Laurie*
18576 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18577 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18578 are no longer created. This way we have a single and consistent command
18579 line interface `openssl <command>`, similar to `cvs <command>`.
18581 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18583 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18584 BIT STRING wrapper always have zero unused bits.
18588 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18592 * Make the top-level INSTALL documentation easier to understand.
18596 * Makefiles updated to exit if an error occurs in a sub-directory
18597 make (including if user presses ^C) [Paul Sutton]
18599 * Make Montgomery context stuff explicit in RSA data structure.
18603 * Fix build order of pem and err to allow for generated pem.h.
18607 * Fix renumbering bug in X509_NAME_delete_entry().
18611 * Enhanced the err-ins.pl script so it makes the error library number
18612 global and can add a library name. This is needed for external ASN1 and
18613 other error libraries.
18617 * Fixed sk_insert which never worked properly.
18621 * Fix ASN1 macros so they can handle indefinite length constructed
18622 EXPLICIT tags. Some non standard certificates use these: they can now
18627 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18628 into a single doc/ssleay.txt bundle. This way the information is still
18629 preserved but no longer messes up this directory. Now it's new room for
18630 the new set of documentation files.
18632 *Ralf S. Engelschall*
18634 * SETs were incorrectly DER encoded. This was a major pain, because they
18635 shared code with SEQUENCEs, which aren't coded the same. This means that
18636 almost everything to do with SETs or SEQUENCEs has either changed name or
18637 number of arguments.
18639 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18641 * Fix test data to work with the above.
18645 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18646 was already fixed by Eric for 0.9.1 it seems.
18648 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18650 * Autodetect FreeBSD3.
18654 * Fix various bugs in Configure. This affects the following platforms:
18658 unixware-2.0-pentium
18663 * Eliminate generated files from CVS. Reorder tests to regenerate files
18664 before they are needed.
18668 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18672 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18674 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18675 changed SSLeay to OpenSSL in version strings.
18677 *Ralf S. Engelschall*
18679 * Some fixups to the top-level documents.
18683 * Fixed the nasty bug where rsaref.h was not found under compile-time
18684 because the symlink to include/ was missing.
18686 *Ralf S. Engelschall*
18688 * Incorporated the popular no-RSA/DSA-only patches
18689 which allow to compile a RSA-free SSLeay.
18691 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18693 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18694 when "ssleay" is still not found.
18696 *Ralf S. Engelschall*
18698 * Added more platforms to Configure: Cray T3E, HPUX 11,
18700 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18702 * Updated the README file.
18704 *Ralf S. Engelschall*
18706 * Added various .cvsignore files in the CVS repository subdirs
18707 to make a "cvs update" really silent.
18709 *Ralf S. Engelschall*
18711 * Recompiled the error-definition header files and added
18712 missing symbols to the Win32 linker tables.
18714 *Ralf S. Engelschall*
18716 * Cleaned up the top-level documents;
18717 o new files: CHANGES and LICENSE
18718 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18719 o merged COPYRIGHT into LICENSE
18720 o removed obsolete TODO file
18721 o renamed MICROSOFT to INSTALL.W32
18723 *Ralf S. Engelschall*
18725 * Removed dummy files from the 0.9.1b source tree:
18727 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18728 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18729 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18730 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18731 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18734 *Ralf S. Engelschall*
18736 * Added various platform portability fixes.
18740 * The Genesis of the OpenSSL rpject:
18741 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18742 Young and Tim J. Hudson created while they were working for C2Net until
18745 *The OpenSSL Project*
18747 ### Changes between 0.9.0b and 0.9.1b [not released]
18749 * Updated a few CA certificates under certs/
18753 * Changed some BIGNUM api stuff.
18757 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18758 DGUX x86, Linux Alpha, etc.
18762 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18763 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18768 * Add -strparse option to asn1pars program which parses nested
18771 *Dr Stephen Henson <shenson@bigfoot.com>*
18773 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18777 * DSA fix for "ca" program.
18781 * Added "-genkey" option to "dsaparam" program.
18785 * Added RIPE MD160 (rmd160) message digest.
18789 * Added -a (all) option to "ssleay version" command.
18793 * Added PLATFORM define which is the id given to Configure.
18797 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18801 * Extended the ASN.1 parser routines.
18805 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18809 * Added a BN_CTX to the BN library.
18813 * Fixed the weak key values in DES library
18817 * Changed API in EVP library for cipher aliases.
18821 * Added support for RC2/64bit cipher.
18825 * Converted the lhash library to the crypto/mem.c functions.
18829 * Added more recognized ASN.1 object ids.
18833 * Added more RSA padding checks for SSL/TLS.
18837 * Added BIO proxy/filter functionality.
18841 * Added extra_certs to SSL_CTX which can be used
18842 send extra CA certificates to the client in the CA cert chain sending
18843 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18847 * Now Fortezza is denied in the authentication phase because
18848 this is key exchange mechanism is not supported by SSLeay at all.
18852 * Additional PKCS1 checks.
18856 * Support the string "TLSv1" for all TLS v1 ciphers.
18860 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18861 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18865 * Fixed a few memory leaks.
18869 * Fixed various code and comment typos.
18873 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18874 bytes sent in the client random.
18876 *Edward Bishop <ebishop@spyglass.com>*
18880 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18881 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18882 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18883 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18884 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18885 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18886 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18887 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18888 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18889 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18890 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18891 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18892 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18893 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18894 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18895 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18896 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18897 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18898 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18899 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18900 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18901 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18902 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18903 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18904 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18905 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18906 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18907 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18908 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18909 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18910 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18911 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18912 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18913 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18914 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18915 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18916 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18917 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18918 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18919 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18920 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18921 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18922 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18923 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18924 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18925 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18926 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18927 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18928 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18929 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18930 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18931 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18932 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18933 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18934 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18935 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18936 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18937 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18938 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18939 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18940 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18941 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18942 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18943 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18944 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18945 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18946 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18947 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18948 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18949 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18950 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18951 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18952 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18953 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18954 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18955 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18956 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18957 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18958 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18959 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18960 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18961 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18962 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18963 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18964 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18965 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18966 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18967 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18968 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18969 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18970 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18971 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18972 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18973 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18974 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18975 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18976 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18977 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18978 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18979 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18980 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18981 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18982 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18983 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18984 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18985 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18986 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18987 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18988 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18989 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18990 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18991 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18992 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18993 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18994 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18995 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18996 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18997 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18998 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18999 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19000 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19001 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19002 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19003 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19004 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19005 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19006 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19007 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19008 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19009 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19010 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19011 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19012 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19013 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19014 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19015 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19016 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19017 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19018 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19019 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19020 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19021 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19022 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19023 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19024 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19025 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19026 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19027 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19028 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19029 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19030 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19031 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19032 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19033 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19034 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19035 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19036 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19037 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19038 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19039 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19040 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19041 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655