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 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
27 Typically if OpenSSL has no EC or DH algorithms then it cannot support
28 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
29 through providers. Therefore third party providers may supply group
30 implementations even where there are no built-in ones. Attempting to create
31 TLS connections in such a build without also disabling TLSv1.3 at run time or
32 using third party provider groups may result in handshake failures. TLSv1.3
33 can be disabled at compile time using the "no-tls1_3" Configure option.
37 * The undocumented function X509_certificate_type() has been deprecated;
38 applications can use X509_get0_pubkey() and X509_get0_signature() to
39 get the same information.
43 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
44 functions. They are identical to BN_rand() and BN_rand_range()
49 * Deprecated the obsolete X9.31 RSA key generation related functions
50 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
51 BN_X931_generate_prime_ex().
55 * Deprecate EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
56 as they are not useful with non-deprecated functions.
60 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
61 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
62 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
63 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
64 were used to collect all necessary data to form a HTTP request, and to
65 perform the HTTP transfer with that request. With OpenSSL 3.0, the
66 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
67 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
68 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
69 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
70 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
71 OSSL_HTTP_REQ_CTX_set_max_response_length().
73 *Rich Salz and Richard Levitte*
75 * Validation of SM2 keys has been separated from the validation of regular EC
76 keys, allowing to improve the SM2 validation process to reject loaded private
77 keys that are not conforming to the SM2 ISO standard.
78 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
83 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
84 switches: a validation failure triggers an early exit, returning a failure
85 exit status to the parent process.
89 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
90 to ignore unknown ciphers.
94 * The `-cipher-commands` and `-digest-commands` options
95 of the command line utility `list` have been deprecated.
96 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
100 * All of the low level EC_KEY functions have been deprecated including:
102 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
103 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
104 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
105 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
106 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
107 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
108 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
109 EC_KEY_METHOD_get_verify,
110 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
111 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
112 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
113 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
114 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
115 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
116 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
117 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
118 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
119 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
120 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
121 Applications that need to implement an EC_KEY_METHOD need to consider
122 implementation of the functionality in a special provider.
123 For replacement of the functions manipulating the EC_KEY objects
124 see the EVP_PKEY-EC(7) manual page.
126 Additionally functions that read and write EC_KEY objects such as
127 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
128 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
129 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
130 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
131 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
132 have also been deprecated. Applications should instead use the
133 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
135 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
136 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
137 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
138 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
139 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
141 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
143 * Deprecated all the libcrypto and libssl error string loading
144 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
145 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
146 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
147 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
148 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
149 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
150 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
151 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
152 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
153 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
155 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
156 now loads error strings automatically.
160 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
161 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
162 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
163 are to be used by servers requiring ephemeral DH keys. Instead applications
164 should consider using the built-in DH parameters that are available by
165 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
166 necessary then applications can use the alternative functions
167 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
168 replacement for the "callback" functions. The callback was originally useful
169 in order to have different parameters for export and non-export ciphersuites.
170 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
171 functions should be replaced by one of the other methods described above.
175 * The `-crypt` option to the `passwd` command line tool has been removed.
179 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
184 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
186 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
187 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
188 The inverse ciphers use AES decryption for wrapping, and
189 AES encryption for unwrapping.
193 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
194 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
195 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
196 OpenSSL 3.0 these are replaced by the more generic functions
197 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
198 The old versions have been converted to deprecated macros that just call the
203 * The security callback, which can be customised by application code, supports
204 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
205 in the "other" parameter. In most places this is what is passed. All these
206 places occur server side. However there was one client side call of this
207 security operation and it passed a DH object instead. This is incorrect
208 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
209 of the other locations. Therefore this client side call has been changed to
210 pass an EVP_PKEY instead.
214 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
215 interface. Their functionality remains unchanged.
219 * Added new option for 'openssl list', '-providers', which will display the
220 list of loaded providers, their names, version and status. It optionally
221 displays their gettable parameters.
225 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
226 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
227 type is internally recognised so the workaround is no longer needed.
229 Functionality is still retained as it is, but will only work with
230 EVP_PKEYs with a legacy internal key.
234 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
235 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
239 * Changed all "STACK" functions to be macros instead of inline functions. Macro
240 parameters are still checked for type safety at compile time via helper
245 * Remove the RAND_DRBG API
247 The RAND_DRBG API did not fit well into the new provider concept as
248 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
249 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
250 and some of its API calls are rather low-level. This holds in particular
251 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
253 Adding a compatibility layer to continue supporting the RAND_DRBG API as
254 a legacy API for a regular deprecation period turned out to come at the
255 price of complicating the new provider API unnecessarily. Since the
256 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
259 *Paul Dale and Matthias St. Pierre*
261 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
262 as well as actual hostnames.
266 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
267 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
268 conversely, silently ignore DTLS protocol version bounds when configuring
269 TLS-based contexts. The commands can be repeated to set bounds of both
270 types. The same applies with the corresponding "min_protocol" and
271 "max_protocol" command-line switches, in case some application uses both TLS
274 SSL_CTX instances that are created for a fixed protocol version (e.g.
275 `TLSv1_server_method()`) also silently ignore version bounds. Previously
276 attempts to apply bounds to these protocol versions would result in an
277 error. Now only the "version-flexible" SSL_CTX instances are subject to
278 limits in configuration files in command-line options.
282 * Deprecated the `ENGINE` API. Engines should be replaced with providers
287 * Reworked the recorded ERR codes to make better space for system errors.
288 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
289 given code is a system error (true) or an OpenSSL error (false).
293 * Reworked the test perl framework to better allow parallel testing.
295 *Nicola Tuveri and David von Oheimb*
297 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
298 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
302 * 'Configure' has been changed to figure out the configuration target if
303 none is given on the command line. Consequently, the 'config' script is
304 now only a mere wrapper. All documentation is changed to only mention
307 *Rich Salz and Richard Levitte*
309 * Added a library context `OSSL_LIB_CTX` that applications as well as
310 other libraries can use to form a separate context within which
311 libcrypto operations are performed.
313 There are two ways this can be used:
315 - Directly, by passing a library context to functions that take
316 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
318 - Indirectly, by creating a new library context and then assigning
319 it as the new default, with `OSSL_LIB_CTX_set0_default`.
321 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
322 apart from the functions directly related to `OSSL_LIB_CTX`, accept
323 NULL to indicate that the default library context should be used.
325 Library code that changes the default library context using
326 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
327 second call before returning to the caller.
329 _(Note: the library context was initially called `OPENSSL_CTX` and
330 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
334 * Handshake now fails if Extended Master Secret extension is dropped
339 * Dropped interactive mode from the `openssl` program. From now on,
340 running it without arguments is equivalent to `openssl help`.
344 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
345 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
346 While the old function names have been retained for backward compatibility
347 they should not be used in new developments
348 because their return values are confusing: Unlike other `_cmp()` functions
349 they do not return 0 in case their arguments are equal.
353 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
354 `EC_GROUP_get_field_type()`.
358 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
359 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
360 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
361 Applications should rely on the library automatically assigning a suitable
362 EC_METHOD internally upon EC_GROUP construction.
366 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
367 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
368 assigned internally without application intervention.
369 Users of EC_GROUP_new() should switch to a different suitable constructor.
373 * Add CAdES-BES signature verification support, mostly derived
374 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
376 *Filipe Raimundo da Silva*
378 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
382 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
383 functions are not widely used and now OpenSSL automatically perform this
384 conversion when needed.
388 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
389 EC_KEY_precompute_mult(). These functions are not widely used and
390 applications should instead switch to named curves which OpenSSL has
391 hardcoded lookup tables for.
395 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
396 should instead use the L<EC_POINT_mul(3)> function.
400 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
401 that are not applicable to the new provider model. Applications should
402 instead use EVP_default_properties_is_fips_enabled() and
403 EVP_default_properties_enable_fips().
407 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
408 is set, an unexpected EOF is ignored, it pretends a close notify was received
409 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
413 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
414 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
415 used and applications should instead use the
416 L<EC_POINT_set_affine_coordinates(3)> and
417 L<EC_POINT_get_affine_coordinates(3)> functions.
421 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
422 arrays to be more easily constructed via a series of utility functions.
423 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
424 the various push functions and finally convert to a passable OSSL_PARAM
425 array using OSSL_PARAM_BLD_to_param().
429 * The security strength of SHA1 and MD5 based signatures in TLS has been
430 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
431 working at the default security level of 1 and instead requires security
432 level 0. The security level can be changed either using the cipher string
433 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
437 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
438 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
439 internal keys, if they correspond to one of those built in types.
443 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
444 contain a provider side internal key.
448 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
449 They are old functions that we don't use, and that you could disable with
450 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
454 * Project text documents not yet having a proper file name extension
455 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
456 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
460 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
461 have been converted to Markdown with the goal to produce documents
462 which not only look pretty when viewed online in the browser, but
463 remain well readable inside a plain text editor.
465 To achieve this goal, a 'minimalistic' Markdown style has been applied
466 which avoids formatting elements that interfere too much with the
467 reading flow in the text file. For example, it
469 * avoids [ATX headings][] and uses [setext headings][] instead
470 (which works for `<h1>` and `<h2>` headings only).
471 * avoids [inline links][] and uses [reference links][] instead.
472 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
474 [ATX headings]: https://github.github.com/gfm/#atx-headings
475 [setext headings]: https://github.github.com/gfm/#setext-headings
476 [inline links]: https://github.github.com/gfm/#inline-link
477 [reference links]: https://github.github.com/gfm/#reference-link
478 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
479 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
481 *Matthias St. Pierre*
483 * The test suite is changed to preserve results of each test recipe.
484 A new directory test-runs/ with subdirectories named like the
485 test recipes are created in the build tree for this purpose.
489 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
490 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
491 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
493 *David von Oheimb, Martin Peylo*
495 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
496 The legacy OCSP-focused and only partly documented API is retained for
497 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
501 * Added `util/check-format.pl`, a tool for checking adherence to the
502 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
503 The checks performed are incomplete and yield some false positives.
504 Still the tool should be useful for detecting most typical glitches.
508 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
509 If domain name resolution yields multiple IP addresses all of them are tried
510 after `connect()` failures.
514 * All of the low level RSA functions have been deprecated including:
516 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
517 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
518 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
519 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
520 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
521 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
522 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
523 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
524 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
525 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
526 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
527 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
528 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
529 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
530 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
531 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
532 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
533 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
534 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
535 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
536 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
537 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
538 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
539 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
540 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
541 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
542 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
543 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
545 Use of these low level functions has been informally discouraged for a long
546 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
547 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
548 L<EVP_PKEY_decrypt(3)>.
552 * X509 certificates signed using SHA1 are no longer allowed at security
554 In TLS/SSL the default security level is 1. It can be set either
555 using the cipher string with `@SECLEVEL`, or calling
556 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
557 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
559 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
560 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
561 options of the commands.
565 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
566 modified to use PKEY APIs. These commands are now in maintenance mode
567 and no new features will be added to them.
571 * The command line utility rsautl has been deprecated.
572 Instead use the pkeyutl program.
576 * The command line utilities genrsa and rsa have been modified to use PKEY
577 APIs. They now write PKCS#8 keys by default. These commands are now in
578 maintenance mode and no new features will be added to them.
582 * All of the low level DH functions have been deprecated including:
584 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
585 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
586 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
587 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
588 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
589 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
590 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
591 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
592 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
593 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
594 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
595 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
596 DH_meth_get_generate_params and DH_meth_set_generate_params.
598 Use of these low level functions has been informally discouraged for a long
599 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
600 and L<EVP_PKEY_derive(3)>.
602 Additionally functions that read and write DH objects such as d2i_DHparams,
603 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
604 functions have also been deprecated. Applications should instead use the
605 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
607 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
608 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
609 `EVP_PKEY_set1_DH()` are also deprecated.
610 Applications should instead either read or write an
611 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
612 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
614 *Paul Dale and Matt Caswell*
616 * All of the low level DSA functions have been deprecated including:
618 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
619 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
620 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
621 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
622 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
623 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
624 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
625 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
626 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
627 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
628 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
629 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
630 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
631 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
632 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
633 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
634 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
636 Use of these low level functions has been informally discouraged for a long
637 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
638 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
642 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
643 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
644 This means that applications don't have to look at the curve NID and
645 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
646 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
647 a no-op when the EVP_PKEY is already of the given type.
649 Parameter and key generation is also reworked to make it possible
650 to generate EVP_PKEY_SM2 parameters and keys without having to go
651 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
652 However, code that does the latter will still work as before.
656 * Deprecated low level ECDH and ECDSA functions. These include:
658 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
659 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
662 Use of these low level functions has been informally discouraged for a long
663 time. Instead applications should use the EVP_PKEY_derive(3),
664 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
668 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
669 and EVP_PKEY_decrypt() instead.
670 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
671 and EVP_PKEY_encrypt() instead.
675 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
676 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
677 a new formulation to include all the things it can be used for,
678 as well as words of caution.
682 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
683 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
687 * All of the low level HMAC functions have been deprecated including:
689 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
690 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
693 Use of these low level functions has been informally discouraged for a long
694 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
695 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
696 and L<EVP_MAC_final(3)>.
700 * Over two thousand fixes were made to the documentation, including:
701 - Common options (such as -rand/-writerand, TLS version control, etc)
702 were refactored and point to newly-enhanced descriptions in openssl.pod.
703 - Added style conformance for all options (with help from Richard Levitte),
704 documented all reported missing options, added a CI build to check
705 that all options are documented and that no unimplemented options
707 - Documented some internals, such as all use of environment variables.
708 - Addressed all internal broken L<> references.
712 * All of the low level CMAC functions have been deprecated including:
714 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
715 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
717 Use of these low level functions has been informally discouraged for a long
718 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
719 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
720 and L<EVP_MAC_final(3)>.
724 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
725 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
728 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
729 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
730 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
731 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
732 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
733 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
734 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
735 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
736 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
737 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
739 Use of these low level functions has been informally discouraged
740 for a long time. Applications should use the EVP_DigestInit_ex(3),
741 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
745 * Corrected the documentation of the return values from the `EVP_DigestSign*`
746 set of functions. The documentation mentioned negative values for some
747 errors, but this was never the case, so the mention of negative values
750 Code that followed the documentation and thereby check with something
751 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
755 * All of the low level cipher functions have been deprecated including:
757 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
758 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
759 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
760 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
761 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
762 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
763 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
764 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
765 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
766 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
767 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
768 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
769 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
770 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
771 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
772 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
773 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
774 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
775 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
776 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
777 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
778 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
779 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
780 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
781 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
782 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
783 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
784 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
785 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
787 Use of these low level functions has been informally discouraged for
788 a long time. Applications should use the high level EVP APIs, e.g.
789 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
790 equivalently named decrypt functions instead.
792 *Matt Caswell and Paul Dale*
794 * Removed include/openssl/opensslconf.h.in and replaced it with
795 include/openssl/configuration.h.in, which differs in not including
796 <openssl/macros.h>. A short header include/openssl/opensslconf.h
797 was added to include both.
799 This allows internal hacks where one might need to modify the set
800 of configured macros, for example this if deprecated symbols are
801 still supposed to be available internally:
803 #include <openssl/configuration.h>
805 #undef OPENSSL_NO_DEPRECATED
806 #define OPENSSL_SUPPRESS_DEPRECATED
808 #include <openssl/macros.h>
810 This should not be used by applications that use the exported
811 symbols, as that will lead to linking errors.
815 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
816 used in exponentiation with 512-bit moduli. No EC algorithms are
817 affected. Analysis suggests that attacks against 2-prime RSA1024,
818 3-prime RSA1536, and DSA1024 as a result of this defect would be very
819 difficult to perform and are not believed likely. Attacks against DH512
820 are considered just feasible. However, for an attack the target would
821 have to re-use the DH512 private key, which is not recommended anyway.
822 Also applications directly using the low level API BN_mod_exp may be
823 affected if they use BN_FLG_CONSTTIME.
828 * Most memory-debug features have been deprecated, and the functionality
829 replaced with no-ops.
833 * Added documentation for the STACK API.
837 * Introduced a new method type and API, OSSL_ENCODER, to
838 represent generic encoders. An implementation is expected to
839 be able to encode an object associated with a given name (such
840 as an algorithm name for an asymmetric key) into forms given by
841 implementation properties.
843 Encoders are primarily used from inside libcrypto, through
844 calls to functions like EVP_PKEY_print_private(),
845 PEM_write_bio_PrivateKey() and similar.
847 Encoders are specified in such a way that they can be made to
848 directly handle the provider side portion of an object, if this
849 provider side part comes from the same provider as the encoder
850 itself, but can also be made to handle objects in parametrized
851 form (as an OSSL_PARAM array of data). This allows a provider to
852 offer generic encoders as a service for any other provider.
856 * Added a .pragma directive to the syntax of configuration files, to
857 allow varying behavior in a supported and predictable manner.
858 Currently added pragma:
862 This allows dollar signs to be a keyword character unless it's
863 followed by a opening brace or parenthesis. This is useful for
864 platforms where dollar signs are commonly used in names, such as
865 volume names and system directory names on VMS.
869 * Added functionality to create an EVP_PKEY from user data. This
870 is effectively the same as creating a RSA, DH or DSA object and
871 then assigning them to an EVP_PKEY, but directly using algorithm
872 agnostic EVP functions. A benefit is that this should be future
873 proof for public key algorithms to come.
877 * Change the interpretation of the '--api' configuration option to
878 mean that this is a desired API compatibility level with no
879 further meaning. The previous interpretation, that this would
880 also mean to remove all deprecated symbols up to and including
881 the given version, no requires that 'no-deprecated' is also used
882 in the configuration.
884 When building applications, the desired API compatibility level
885 can be set with the OPENSSL_API_COMPAT macro like before. For
886 API compatibility version below 3.0, the old style numerical
887 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
888 For version 3.0 and on, the value is expected to be the decimal
889 value calculated from the major and minor version like this:
891 MAJOR * 10000 + MINOR * 100
895 -DOPENSSL_API_COMPAT=30000 For 3.0
896 -DOPENSSL_API_COMPAT=30200 For 3.2
898 To hide declarations that are deprecated up to and including the
899 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
900 given when building the application as well.
904 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
905 access to certificate and CRL stores via URIs and OSSL_STORE
908 This adds the following functions:
910 - X509_LOOKUP_store()
911 - X509_STORE_load_file()
912 - X509_STORE_load_path()
913 - X509_STORE_load_store()
914 - SSL_add_store_cert_subjects_to_stack()
915 - SSL_CTX_set_default_verify_store()
916 - SSL_CTX_load_verify_file()
917 - SSL_CTX_load_verify_dir()
918 - SSL_CTX_load_verify_store()
922 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
923 The presence of this system service is determined at run-time.
927 * Added functionality to create an EVP_PKEY context based on data
928 for methods from providers. This takes an algorithm name and a
929 property query string and simply stores them, with the intent
930 that any operation that uses this context will use those strings
931 to fetch the needed methods implicitly, thereby making the port
932 of application written for pre-3.0 OpenSSL easier.
936 * The undocumented function NCONF_WIN32() has been deprecated; for
937 conversion details see the HISTORY section of doc/man5/config.pod
941 * Introduced the new functions EVP_DigestSignInit_ex() and
942 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
943 EVP_DigestVerifyUpdate() have been converted to functions. See the man
944 pages for further details.
948 * Over two thousand fixes were made to the documentation, including:
949 adding missing command flags, better style conformance, documentation
952 *Rich Salz, Richard Levitte*
954 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
955 X25519, X448, Ed25519 and Ed448.
959 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
964 * Deprecated the public definition of `ERR_STATE` as well as the function
965 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
970 * Added ERR functionality to give callers access to the stored function
971 names that have replaced the older function code based functions.
973 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
974 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
975 ERR_peek_error_all() and ERR_peek_last_error_all().
977 These functions have become deprecated: ERR_get_error_line(),
978 ERR_get_error_line_data(), ERR_peek_error_line_data(),
979 ERR_peek_last_error_line_data() and ERR_func_error_string().
981 Users are recommended to use ERR_get_error_all(), or to pick information
982 with ERR_peek functions and finish off with getting the error code by using
987 * Extended testing to be verbose for failing tests only. The make variables
988 VERBOSE_FAILURE or VF can be used to enable this:
990 $ make VF=1 test # Unix
991 $ mms /macro=(VF=1) test ! OpenVMS
992 $ nmake VF=1 test # Windows
996 * Added the `-copy_extensions` option to the `x509` command for use with
997 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
998 all extensions in the request are copied to the certificate or vice versa.
1000 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1002 * Added the `-copy_extensions` option to the `req` command for use with
1003 `-x509`. When given with the `copy` or `copyall` argument,
1004 all extensions in the certification request are copied to the certificate.
1008 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1009 they generate are by default RFC 5280 compliant in the following sense:
1010 There is a subjectKeyIdentifier extension with a hash value of the public key
1011 and for not self-signed certs there is an authorityKeyIdentifier extension
1012 with a keyIdentifier field or issuer information identifying the signing key.
1013 This is done unless some configuration overrides the new default behavior,
1014 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1018 * Added several checks to `X509_verify_cert()` according to requirements in
1019 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1020 (which may be done by using the CLI option `-x509_strict`):
1021 * The basicConstraints of CA certificates must be marked critical.
1022 * CA certificates must explicitly include the keyUsage extension.
1023 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1024 * The issuer name of any certificate must not be empty.
1025 * The subject name of CA certs, certs with keyUsage crlSign,
1026 and certs without subjectAlternativeName must not be empty.
1027 * If a subjectAlternativeName extension is given it must not be empty.
1028 * The signatureAlgorithm field and the cert signature must be consistent.
1029 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1030 must not be marked critical.
1031 * The authorityKeyIdentifier must be given for X.509v3 certs
1032 unless they are self-signed.
1033 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1037 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1038 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1042 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1043 used even when parsing explicit parameters, when loading a encoded key
1044 or calling `EC_GROUP_new_from_ecpkparameters()`/
1045 `EC_GROUP_new_from_ecparameters()`.
1046 This prevents bypass of security hardening and performance gains,
1047 especially for curves with specialized EC_METHODs.
1048 By default, if a key encoded with explicit parameters is loaded and later
1049 encoded, the output is still encoded with explicit parameters, even if
1050 internally a "named" EC_GROUP is used for computation.
1054 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1055 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1056 NULL. After this change, only the cofactor parameter can be NULL. It also
1057 does some minimal sanity checks on the passed order.
1062 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1063 An attack is simple, if the first CMS_recipientInfo is valid but the
1064 second CMS_recipientInfo is chosen ciphertext. If the second
1065 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1066 encryption key will be replaced by garbage, and the message cannot be
1067 decoded, but if the RSA decryption fails, the correct encryption key is
1068 used and the recipient will not notice the attack.
1069 As a work around for this potential attack the length of the decrypted
1070 key must be equal to the cipher default key length, in case the
1071 certifiate is not given and all recipientInfo are tried out.
1072 The old behaviour can be re-enabled in the CMS code by setting the
1073 CMS_DEBUG_DECRYPT flag.
1077 * Early start up entropy quality from the DEVRANDOM seed source has been
1078 improved for older Linux systems. The RAND subsystem will wait for
1079 /dev/random to be producing output before seeding from /dev/urandom.
1080 The seeded state is stored for future library initialisations using
1081 a system global shared memory segment. The shared memory identifier
1082 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1083 the desired value. The default identifier is 114.
1087 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1088 when primes for RSA keys are computed.
1089 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1090 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1091 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1092 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1093 This avoids possible fingerprinting of newly generated RSA modules.
1097 * Correct the extended master secret constant on EBCDIC systems. Without this
1098 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1099 negotiate EMS will fail. Unfortunately this also means that TLS connections
1100 between EBCDIC systems with this fix, and EBCDIC systems without this
1101 fix will fail if they negotiate EMS.
1105 * Changed the library initialisation so that the config file is now loaded
1106 by default. This was already the case for libssl. It now occurs for both
1107 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1108 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1112 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1113 where the former acts as a replacement for `ERR_put_error()`, and the
1114 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1115 `ERR_raise_data()` adds more flexibility by taking a format string and
1116 an arbitrary number of arguments following it, to be processed with
1121 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1122 to check if a named provider is loaded and available. When called, it
1123 will also activate all fallback providers if such are still present.
1127 * Enforce a minimum DH modulus size of 512 bits.
1131 * Changed DH parameters to generate the order q subgroup instead of 2q.
1132 Previously generated DH parameters are still accepted by DH_check
1133 but DH_generate_key works around that by clearing bit 0 of the
1134 private key for those. This avoids leaking bit 0 of the private key.
1138 * Significantly reduce secure memory usage by the randomness pools.
1142 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1147 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1148 algorithms. An implementation of a key exchange algorithm can be obtained
1149 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1150 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1151 the older EVP_PKEY_derive_init() function. See the man pages for the new
1152 functions for further details.
1156 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1160 * Removed the function names from error messages and deprecated the
1163 * Removed NextStep support and the macro OPENSSL_UNISTD
1167 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1168 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1169 Also removed "export var as function" capability; we do not export
1170 variables, only functions.
1174 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1175 an error and 1 indicating success. In previous versions of OpenSSL this
1176 was a void type. If a key was set longer than the maximum possible this
1181 * Support SM2 signing and verification schemes with X509 certificate.
1185 * Use SHA256 as the default digest for TS query in the `ts` app.
1189 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1190 This checks that the salt length is at least 128 bits, the derived key
1191 length is at least 112 bits, and that the iteration count is at least 1000.
1192 For backwards compatibility these checks are disabled by default in the
1193 default provider, but are enabled by default in the fips provider.
1194 To enable or disable these checks use the control
1195 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1199 * Default cipher lists/suites are now available via a function, the
1200 #defines are deprecated.
1204 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1205 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1206 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1210 * Join the directories crypto/x509 and crypto/x509v3
1214 * Added command 'openssl kdf' that uses the EVP_KDF API.
1218 * Added command 'openssl mac' that uses the EVP_MAC API.
1222 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1223 as default directories. Also added the command 'openssl info'
1224 for scripting purposes.
1228 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1229 deprecated. These undocumented functions were never integrated into the EVP
1230 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1231 Bi-directional IGE mode. These modes were never formally standardised and
1232 usage of these functions is believed to be very small. In particular
1233 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1234 is ever used. The security implications are believed to be minimal, but
1235 this issue was never fixed for backwards compatibility reasons. New code
1236 should not use these modes.
1240 * Add prediction resistance to the DRBG reseeding process.
1244 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1245 mandated by IEEE Std 1619-2018.
1249 * Added newline escaping functionality to a filename when using openssl dgst.
1250 This output format is to replicate the output format found in the `*sum`
1251 checksum programs. This aims to preserve backward compatibility.
1253 *Matt Eaton, Richard Levitte, and Paul Dale*
1255 * Removed the heartbeat message in DTLS feature, as it has very
1256 little usage and doesn't seem to fulfill a valuable purpose.
1257 The configuration option is now deprecated.
1261 * Changed the output of 'openssl {digestname} < file' to display the
1262 digest name in its output.
1266 * Added a new generic trace API which provides support for enabling
1267 instrumentation through trace output. This feature is mainly intended
1268 as an aid for developers and is disabled by default. To utilize it,
1269 OpenSSL needs to be configured with the `enable-trace` option.
1271 If the tracing API is enabled, the application can activate trace output
1272 by registering BIOs as trace channels for a number of tracing and debugging
1275 The `openssl` program has been expanded to enable any of the types
1276 available via environment variables defined by the user, and serves as
1277 one possible example on how to use this functionality.
1279 *Richard Levitte & Matthias St. Pierre*
1281 * Added build tests for C++. These are generated files that only do one
1282 thing, to include one public OpenSSL head file each. This tests that
1283 the public header files can be usefully included in a C++ application.
1285 This test isn't enabled by default. It can be enabled with the option
1286 'enable-buildtest-c++'.
1290 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1294 * Add KMAC to EVP_MAC.
1298 * Added property based algorithm implementation selection framework to
1303 * Added SCA hardening for modular field inversion in EC_GROUP through
1304 a new dedicated field_inv() pointer in EC_METHOD.
1305 This also addresses a leakage affecting conversions from projective
1306 to affine coordinates.
1308 *Billy Bob Brumley, Nicola Tuveri*
1310 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1311 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1312 those algorithms that were already supported through the EVP_PKEY API
1313 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1314 and scrypt are now wrappers that call EVP_KDF.
1318 * Build devcrypto engine as a dynamic engine.
1320 *Eneas U de Queiroz*
1322 * Add keyed BLAKE2 to EVP_MAC.
1326 * Fix a bug in the computation of the endpoint-pair shared secret used
1327 by DTLS over SCTP. This breaks interoperability with older versions
1328 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1329 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1330 interoperability with such broken implementations. However, enabling
1331 this switch breaks interoperability with correct implementations.
1333 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1334 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1338 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1342 * Change the license to the Apache License v2.0.
1346 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1348 - Major releases (indicated by incrementing the MAJOR release number)
1349 may introduce incompatible API/ABI changes.
1350 - Minor releases (indicated by incrementing the MINOR release number)
1351 may introduce new features but retain API/ABI compatibility.
1352 - Patch releases (indicated by incrementing the PATCH number)
1353 are intended for bug fixes and other improvements of existing
1354 features only (like improving performance or adding documentation)
1355 and retain API/ABI compatibility.
1359 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1363 * Remove the 'dist' target and add a tarball building script. The
1364 'dist' target has fallen out of use, and it shouldn't be
1365 necessary to configure just to create a source distribution.
1369 * Recreate the OS390-Unix config target. It no longer relies on a
1370 special script like it did for OpenSSL pre-1.1.0.
1374 * Instead of having the source directories listed in Configure, add
1375 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1380 * Add GMAC to EVP_MAC.
1384 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1388 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1389 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1390 to facilitate the continued use of MACs through raw private keys in
1391 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1395 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1396 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1400 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1401 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1402 are retained for backwards compatibility.
1406 * AES-XTS mode now enforces that its two keys are different to mitigate
1407 the attacked described in "Efficient Instantiations of Tweakable
1408 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1409 Details of this attack can be obtained from:
1410 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1414 * Rename the object files, i.e. give them other names than in previous
1415 versions. Their names now include the name of the final product, as
1416 well as its type mnemonic (bin, lib, shlib).
1420 * Added new option for 'openssl list', '-objects', which will display the
1421 list of built in objects, i.e. OIDs with names.
1425 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1426 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1431 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1432 improves application performance by removing data copies and providing
1433 applications with zero-copy system calls such as sendfile and splice.
1437 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1438 option is set, openssl cleanses (zeroize) plaintext bytes from
1439 internal buffers after delivering them to the application. Note,
1440 the application is still responsible for cleansing other copies
1441 (e.g.: data received by SSL_read(3)).
1445 * `PKCS12_parse` now maintains the order of the parsed certificates
1446 when outputting them via `*ca` (rather than reversing it).
1450 * Deprecated pthread fork support methods. These were unused so no
1451 replacement is required.
1453 - OPENSSL_fork_prepare()
1454 - OPENSSL_fork_parent()
1455 - OPENSSL_fork_child()
1462 ### Changes between 1.1.1i and 1.1.1j [xx XXX xxxx]
1464 * Fixed SRP_Calc_client_key so that it uses constant time. The previous
1465 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1466 could be exploited in a side channel attack to recover the password. Since
1467 the attack is local host only this is outside of the current OpenSSL
1468 threat model and therefore no CVE is assigned.
1470 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1475 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1477 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1478 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1479 If an attacker can control both items being compared then this could lead
1480 to a possible denial of service attack. OpenSSL itself uses the
1481 GENERAL_NAME_cmp function for two purposes:
1482 1) Comparing CRL distribution point names between an available CRL and a
1483 CRL distribution point embedded in an X509 certificate
1484 2) When verifying that a timestamp response token signer matches the
1485 timestamp authority name (exposed via the API functions
1486 TS_RESP_verify_response and TS_RESP_verify_token)
1491 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1493 * Certificates with explicit curve parameters are now disallowed in
1494 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1498 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1499 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1500 conversely, silently ignore DTLS protocol version bounds when configuring
1501 TLS-based contexts. The commands can be repeated to set bounds of both
1502 types. The same applies with the corresponding "min_protocol" and
1503 "max_protocol" command-line switches, in case some application uses both TLS
1506 SSL_CTX instances that are created for a fixed protocol version (e.g.
1507 TLSv1_server_method()) also silently ignore version bounds. Previously
1508 attempts to apply bounds to these protocol versions would result in an
1509 error. Now only the "version-flexible" SSL_CTX instances are subject to
1510 limits in configuration files in command-line options.
1514 * Handshake now fails if Extended Master Secret extension is dropped
1519 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1521 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1523 * Fixed segmentation fault in SSL_check_chain()
1524 Server or client applications that call the SSL_check_chain() function
1525 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1526 dereference as a result of incorrect handling of the
1527 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1528 or unrecognised signature algorithm is received from the peer. This could
1529 be exploited by a malicious peer in a Denial of Service attack.
1534 * Added AES consttime code for no-asm configurations
1535 an optional constant time support for AES was added
1536 when building openssl for no-asm.
1537 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1538 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1539 At this time this feature is by default disabled.
1540 It will be enabled by default in 3.0.
1544 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1546 * Revert the change of EOF detection while reading in libssl to avoid
1547 regressions in applications depending on the current way of reporting
1548 the EOF. As the existing method is not fully accurate the change to
1549 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1550 branch and will be present in the 3.0 release.
1554 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1555 when primes for RSA keys are computed.
1556 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1557 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1558 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1559 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1560 This avoids possible fingerprinting of newly generated RSA modules.
1564 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1566 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1567 while reading in libssl then we would report an error back to the
1568 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1569 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1570 therefore give a hint as to what went wrong.
1574 * Check that ed25519 and ed448 are allowed by the security level. Previously
1575 signature algorithms not using an MD were not being checked that they were
1576 allowed by the security level.
1580 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1581 was not quite right. The behaviour was not consistent between resumption
1582 and normal handshakes, and also not quite consistent with historical
1583 behaviour. The behaviour in various scenarios has been clarified and
1584 it has been updated to make it match historical behaviour as closely as
1589 * *[VMS only]* The header files that the VMS compilers include automatically,
1590 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1591 that the C++ compiler doesn't understand. This is a shortcoming in the
1592 compiler, but can be worked around with `__cplusplus` guards.
1594 C++ applications that use OpenSSL libraries must be compiled using the
1595 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1596 functions. Otherwise, only functions with symbols of less than 31
1597 characters can be used, as the linker will not be able to successfully
1598 resolve symbols with longer names.
1602 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1603 The presence of this system service is determined at run-time.
1607 * Added newline escaping functionality to a filename when using openssl dgst.
1608 This output format is to replicate the output format found in the `*sum`
1609 checksum programs. This aims to preserve backward compatibility.
1611 *Matt Eaton, Richard Levitte, and Paul Dale*
1613 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1618 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1620 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1621 number generator (RNG). This was intended to include protection in the
1622 event of a fork() system call in order to ensure that the parent and child
1623 processes did not share the same RNG state. However this protection was not
1624 being used in the default case.
1626 A partial mitigation for this issue is that the output from a high
1627 precision timer is mixed into the RNG state so the likelihood of a parent
1628 and child process sharing state is significantly reduced.
1630 If an application already calls OPENSSL_init_crypto() explicitly using
1631 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1634 *Matthias St. Pierre*
1636 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1637 used even when parsing explicit parameters, when loading a encoded key
1638 or calling `EC_GROUP_new_from_ecpkparameters()`/
1639 `EC_GROUP_new_from_ecparameters()`.
1640 This prevents bypass of security hardening and performance gains,
1641 especially for curves with specialized EC_METHODs.
1642 By default, if a key encoded with explicit parameters is loaded and later
1643 encoded, the output is still encoded with explicit parameters, even if
1644 internally a "named" EC_GROUP is used for computation.
1648 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1649 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1650 NULL. After this change, only the cofactor parameter can be NULL. It also
1651 does some minimal sanity checks on the passed order.
1656 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1657 An attack is simple, if the first CMS_recipientInfo is valid but the
1658 second CMS_recipientInfo is chosen ciphertext. If the second
1659 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1660 encryption key will be replaced by garbage, and the message cannot be
1661 decoded, but if the RSA decryption fails, the correct encryption key is
1662 used and the recipient will not notice the attack.
1663 As a work around for this potential attack the length of the decrypted
1664 key must be equal to the cipher default key length, in case the
1665 certifiate is not given and all recipientInfo are tried out.
1666 The old behaviour can be re-enabled in the CMS code by setting the
1667 CMS_DEBUG_DECRYPT flag.
1672 * Early start up entropy quality from the DEVRANDOM seed source has been
1673 improved for older Linux systems. The RAND subsystem will wait for
1674 /dev/random to be producing output before seeding from /dev/urandom.
1675 The seeded state is stored for future library initialisations using
1676 a system global shared memory segment. The shared memory identifier
1677 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1678 the desired value. The default identifier is 114.
1682 * Correct the extended master secret constant on EBCDIC systems. Without this
1683 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1684 negotiate EMS will fail. Unfortunately this also means that TLS connections
1685 between EBCDIC systems with this fix, and EBCDIC systems without this
1686 fix will fail if they negotiate EMS.
1690 * Use Windows installation paths in the mingw builds
1692 Mingw isn't a POSIX environment per se, which means that Windows
1693 paths should be used for installation.
1698 * Changed DH_check to accept parameters with order q and 2q subgroups.
1699 With order 2q subgroups the bit 0 of the private key is not secret
1700 but DH_generate_key works around that by clearing bit 0 of the
1701 private key for those. This avoids leaking bit 0 of the private key.
1705 * Significantly reduce secure memory usage by the randomness pools.
1709 * Revert the DEVRANDOM_WAIT feature for Linux systems
1711 The DEVRANDOM_WAIT feature added a select() call to wait for the
1712 /dev/random device to become readable before reading from the
1713 /dev/urandom device.
1715 It turned out that this change had negative side effects on
1716 performance which were not acceptable. After some discussion it
1717 was decided to revert this feature and leave it up to the OS
1718 resp. the platform maintainer to ensure a proper initialization
1719 during early boot time.
1721 *Matthias St. Pierre*
1723 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1725 * Add build tests for C++. These are generated files that only do one
1726 thing, to include one public OpenSSL head file each. This tests that
1727 the public header files can be usefully included in a C++ application.
1729 This test isn't enabled by default. It can be enabled with the option
1730 'enable-buildtest-c++'.
1734 * Enable SHA3 pre-hashing for ECDSA and DSA.
1738 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1739 This changes the size when using the `genpkey` command when no size is given.
1740 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1741 generation commands to use 2048 bits by default.
1745 * Reorganize the manual pages to consistently have RETURN VALUES,
1746 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1747 util/fix-doc-nits accordingly.
1749 *Paul Yang, Joshua Lock*
1751 * Add the missing accessor EVP_PKEY_get0_engine()
1755 * Have commands like `s_client` and `s_server` output the signature scheme
1756 along with other cipher suite parameters when debugging.
1760 * Make OPENSSL_config() error agnostic again.
1764 * Do the error handling in RSA decryption constant time.
1768 * Prevent over long nonces in ChaCha20-Poly1305.
1770 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1771 for every encryption operation. RFC 7539 specifies that the nonce value
1772 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1773 and front pads the nonce with 0 bytes if it is less than 12
1774 bytes. However it also incorrectly allows a nonce to be set of up to 16
1775 bytes. In this case only the last 12 bytes are significant and any
1776 additional leading bytes are ignored.
1778 It is a requirement of using this cipher that nonce values are
1779 unique. Messages encrypted using a reused nonce value are susceptible to
1780 serious confidentiality and integrity attacks. If an application changes
1781 the default nonce length to be longer than 12 bytes and then makes a
1782 change to the leading bytes of the nonce expecting the new value to be a
1783 new unique nonce then such an application could inadvertently encrypt
1784 messages with a reused nonce.
1786 Additionally the ignored bytes in a long nonce are not covered by the
1787 integrity guarantee of this cipher. Any application that relies on the
1788 integrity of these ignored leading bytes of a long nonce may be further
1789 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1790 is safe because no such use sets such a long nonce value. However user
1791 applications that use this cipher directly and set a non-default nonce
1792 length to be longer than 12 bytes may be vulnerable.
1794 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1800 * Add DEVRANDOM_WAIT feature for Linux systems
1802 On older Linux systems where the getrandom() system call is not available,
1803 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1804 Contrary to getrandom(), the /dev/urandom device will not block during
1805 early boot when the kernel CSPRNG has not been seeded yet.
1807 To mitigate this known weakness, use select() to wait for /dev/random to
1808 become readable before reading from /dev/urandom.
1810 * Ensure that SM2 only uses SM3 as digest algorithm
1814 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1816 * Change the info callback signals for the start and end of a post-handshake
1817 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1818 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1819 confused by this and assume that a TLSv1.2 renegotiation has started. This
1820 can break KeyUpdate handling. Instead we no longer signal the start and end
1821 of a post handshake message exchange (although the messages themselves are
1822 still signalled). This could break some applications that were expecting
1823 the old signals. However without this KeyUpdate is not usable for many
1828 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1830 * Timing vulnerability in DSA signature generation
1832 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1833 timing side channel attack. An attacker could use variations in the signing
1834 algorithm to recover the private key.
1836 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1841 * Timing vulnerability in ECDSA signature generation
1843 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1844 timing side channel attack. An attacker could use variations in the signing
1845 algorithm to recover the private key.
1847 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1852 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1853 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1854 of two gigabytes and the error handling improved.
1856 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1857 categorized as a normal bug, not a security issue, because the DRBG reseeds
1858 automatically and is fully functional even without additional randomness
1859 provided by the application.
1861 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1863 * Add a new ClientHello callback. Provides a callback interface that gives
1864 the application the ability to adjust the nascent SSL object at the
1865 earliest stage of ClientHello processing, immediately after extensions have
1866 been collected but before they have been processed. In particular, this
1867 callback can adjust the supported TLS versions in response to the contents
1872 * Add SM2 base algorithm support.
1876 * s390x assembly pack: add (improved) hardware-support for the following
1877 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1878 aes-cfb/cfb8, aes-ecb.
1882 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1883 parameter is no longer accepted, as it leads to a corrupt table. NULL
1884 pem_str is reserved for alias entries only.
1888 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1889 step for prime curves. The new implementation is based on formulae from
1890 differential addition-and-doubling in homogeneous projective coordinates
1891 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1892 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1893 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1894 to work in projective coordinates.
1896 *Billy Bob Brumley, Nicola Tuveri*
1898 * Change generating and checking of primes so that the error rate of not
1899 being prime depends on the intended use based on the size of the input.
1900 For larger primes this will result in more rounds of Miller-Rabin.
1901 The maximal error rate for primes with more than 1080 bits is lowered
1904 *Kurt Roeckx, Annie Yousar*
1906 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1910 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1911 moving between systems, and to avoid confusion when a Windows build is
1912 done with mingw vs with MSVC. For POSIX installs, there's still a
1913 symlink or copy named 'tsget' to avoid that confusion as well.
1917 * Revert blinding in ECDSA sign and instead make problematic addition
1918 length-invariant. Switch even to fixed-length Montgomery multiplication.
1922 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1923 step for binary curves. The new implementation is based on formulae from
1924 differential addition-and-doubling in mixed Lopez-Dahab projective
1925 coordinates, modified to independently blind the operands.
1927 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1929 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1930 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1931 EC_METHODs to implement their own specialized "ladder step", to take
1932 advantage of more favorable coordinate systems or more efficient
1933 differential addition-and-doubling algorithms.
1935 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1937 * Modified the random device based seed sources to keep the relevant
1938 file descriptors open rather than reopening them on each access.
1939 This allows such sources to operate in a chroot() jail without
1940 the associated device nodes being available. This behaviour can be
1941 controlled using RAND_keep_random_devices_open().
1945 * Numerous side-channel attack mitigations have been applied. This may have
1946 performance impacts for some algorithms for the benefit of improved
1947 security. Specific changes are noted in this change log by their respective
1952 * AIX shared library support overhaul. Switch to AIX "natural" way of
1953 handling shared libraries, which means collecting shared objects of
1954 different versions and bitnesses in one common archive. This allows to
1955 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1956 doesn't affect the way 3rd party applications are linked, only how
1957 multi-version installation is managed.
1961 * Make ec_group_do_inverse_ord() more robust and available to other
1962 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1963 mitigations are applied to the fallback BN_mod_inverse().
1964 When using this function rather than BN_mod_inverse() directly, new
1965 EC cryptosystem implementations are then safer-by-default.
1969 * Add coordinate blinding for EC_POINT and implement projective
1970 coordinate blinding for generic prime curves as a countermeasure to
1971 chosen point SCA attacks.
1973 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1975 * Add blinding to ECDSA and DSA signatures to protect against side channel
1976 attacks discovered by Keegan Ryan (NCC Group).
1980 * Enforce checking in the `pkeyutl` command to ensure that the input
1981 length does not exceed the maximum supported digest length when performing
1982 a sign, verify or verifyrecover operation.
1986 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1987 I/O in combination with something like select() or poll() will hang. This
1988 can be turned off again using SSL_CTX_clear_mode().
1989 Many applications do not properly handle non-application data records, and
1990 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1991 around the problems in those applications, but can also break some.
1992 It's recommended to read the manpages about SSL_read(), SSL_write(),
1993 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1994 SSL_CTX_set_read_ahead() again.
1998 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1999 now allow empty (zero character) pass phrases.
2003 * Apply blinding to binary field modular inversion and remove patent
2004 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2008 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2009 binary and prime elliptic curves.
2013 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2014 constant time fixed point multiplication.
2018 * Revise elliptic curve scalar multiplication with timing attack
2019 defenses: ec_wNAF_mul redirects to a constant time implementation
2020 when computing fixed point and variable point multiplication (which
2021 in OpenSSL are mostly used with secret scalars in keygen, sign,
2022 ECDH derive operations).
2023 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2026 * Updated CONTRIBUTING
2030 * Updated DRBG / RAND to request nonce and additional low entropy
2031 randomness from the system.
2033 *Matthias St. Pierre*
2035 * Updated 'openssl rehash' to use OpenSSL consistent default.
2039 * Moved the load of the ssl_conf module to libcrypto, which helps
2040 loading engines that libssl uses before libssl is initialised.
2044 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2048 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2050 *Ingo Schwarze, Rich Salz*
2052 * Added output of accepting IP address and port for 'openssl s_server'
2056 * Added a new API for TLSv1.3 ciphersuites:
2057 SSL_CTX_set_ciphersuites()
2058 SSL_set_ciphersuites()
2062 * Memory allocation failures consistently add an error to the error
2067 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2068 in libcrypto when run as setuid/setgid.
2072 * Load any config file by default when libssl is used.
2076 * Added new public header file <openssl/rand_drbg.h> and documentation
2077 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2079 *Matthias St. Pierre*
2081 * QNX support removed (cannot find contributors to get their approval
2082 for the license change).
2086 * TLSv1.3 replay protection for early data has been implemented. See the
2087 SSL_read_early_data() man page for further details.
2091 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2092 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2093 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2094 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2095 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2096 configuration has been separated out. See the ciphers man page or the
2097 SSL_CTX_set_ciphersuites() man page for more information.
2101 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2102 in responder mode now supports the new "-multi" option, which
2103 spawns the specified number of child processes to handle OCSP
2104 requests. The "-timeout" option now also limits the OCSP
2105 responder's patience to wait to receive the full client request
2106 on a newly accepted connection. Child processes are respawned
2107 as needed, and the CA index file is automatically reloaded
2108 when changed. This makes it possible to run the "ocsp" responder
2109 as a long-running service, making the OpenSSL CA somewhat more
2110 feature-complete. In this mode, most diagnostic messages logged
2111 after entering the event loop are logged via syslog(3) rather than
2116 * Added support for X448 and Ed448. Heavily based on original work by
2121 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2122 objects loaded. This adds the functions OSSL_STORE_expect() and
2123 OSSL_STORE_find() as well as needed tools to construct searches and
2124 get the search data out of them.
2128 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2129 version of OpenSSL should review their configuration settings to ensure
2130 that they are still appropriate for TLSv1.3. For further information see:
2131 <https://wiki.openssl.org/index.php/TLS1.3>
2135 * Grand redesign of the OpenSSL random generator
2137 The default RAND method now utilizes an AES-CTR DRBG according to
2138 NIST standard SP 800-90Ar1. The new random generator is essentially
2139 a port of the default random generator from the OpenSSL FIPS 2.0
2140 object module. It is a hybrid deterministic random bit generator
2141 using an AES-CTR bit stream and which seeds and reseeds itself
2142 automatically using trusted system entropy sources.
2144 Some of its new features are:
2145 - Support for multiple DRBG instances with seed chaining.
2146 - The default RAND method makes use of a DRBG.
2147 - There is a public and private DRBG instance.
2148 - The DRBG instances are fork-safe.
2149 - Keep all global DRBG instances on the secure heap if it is enabled.
2150 - The public and private DRBG instance are per thread for lock free
2153 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2155 * Changed Configure so it only says what it does and doesn't dump
2156 so much data. Instead, ./configdata.pm should be used as a script
2157 to display all sorts of configuration data.
2161 * Added processing of "make variables" to Configure.
2165 * Added SHA512/224 and SHA512/256 algorithm support.
2169 * The last traces of Netware support, first removed in 1.1.0, have
2174 * Get rid of Makefile.shared, and in the process, make the processing
2175 of certain files (rc.obj, or the .def/.map/.opt files produced from
2176 the ordinal files) more visible and hopefully easier to trace and
2177 debug (or make silent).
2181 * Make it possible to have environment variable assignments as
2182 arguments to config / Configure.
2186 * Add multi-prime RSA (RFC 8017) support.
2190 * Add SM3 implemented according to GB/T 32905-2016
2191 *Jack Lloyd <jack.lloyd@ribose.com>,*
2192 *Ronald Tse <ronald.tse@ribose.com>,*
2193 *Erick Borsboom <erick.borsboom@ribose.com>*
2195 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2196 as documented in RFC6066.
2197 Based on a patch from Tomasz Moń
2199 *Filipe Raimundo da Silva*
2201 * Add SM4 implemented according to GB/T 32907-2016.
2202 *Jack Lloyd <jack.lloyd@ribose.com>,*
2203 *Ronald Tse <ronald.tse@ribose.com>,*
2204 *Erick Borsboom <erick.borsboom@ribose.com>*
2206 * Reimplement -newreq-nodes and ERR_error_string_n; the
2207 original author does not agree with the license change.
2211 * Add ARIA AEAD TLS support.
2215 * Some macro definitions to support VS6 have been removed. Visual
2216 Studio 6 has not worked since 1.1.0
2220 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2221 without clearing the errors.
2225 * Add "atfork" functions. If building on a system that without
2226 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2227 requirements. The RAND facility now uses/requires this.
2235 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2236 not possible to disable entirely. However, it's still possible to
2237 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2240 To disable, configure with 'no-ui-console'. 'no-ui' is still
2241 possible to use as an alias. Check at compile time with the
2242 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2243 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2247 * Add a STORE module, which implements a uniform and URI based reader of
2248 stores that can contain keys, certificates, CRLs and numerous other
2249 objects. The main API is loosely based on a few stdio functions,
2250 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2251 OSSL_STORE_error and OSSL_STORE_close.
2252 The implementation uses backends called "loaders" to implement arbitrary
2253 URI schemes. There is one built in "loader" for the 'file' scheme.
2257 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2258 then adjusted to work on FreeBSD 8.4 as well.
2259 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2260 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2264 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2265 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2266 error code calls like this:
2268 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2270 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2271 that can be encoded in C. For the foreseeable future, this will only
2274 *Richard Levitte and Tim Hudson*
2276 * Removed BSD cryptodev engine.
2280 * Add a build target 'build_all_generated', to build all generated files
2281 and only that. This can be used to prepare everything that requires
2282 things like perl for a system that lacks perl and then move everything
2283 to that system and do the rest of the build there.
2287 * In the UI interface, make it possible to duplicate the user data. This
2288 can be used by engines that need to retain the data for a longer time
2289 than just the call where this user data is passed.
2293 * Ignore the '-named_curve auto' value for compatibility of applications
2296 *Tomáš Mráz <tmraz@fedoraproject.org>*
2298 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2299 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2300 alerts across multiple records (some of which could be empty). In practice
2301 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2302 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2303 support this at all. Supporting it adds significant complexity to the
2304 record layer, and its removal is unlikely to cause interoperability
2309 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2310 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2311 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2316 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2317 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2319 *Richard Levitte, Andy Polyakov*
2321 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2326 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2327 platform rather than 'mingw'.
2331 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2332 success if they are asked to add an object which already exists
2333 in the store. This change cascades to other functions which load
2334 certificates and CRLs.
2338 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2339 facilitate stack unwinding even from assembly subroutines.
2343 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2344 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2348 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2349 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2350 which is the minimum version we support.
2354 * Certificate time validation (X509_cmp_time) enforces stricter
2355 compliance with RFC 5280. Fractional seconds and timezone offsets
2356 are no longer allowed.
2360 * Add support for ARIA
2364 * s_client will now send the Server Name Indication (SNI) extension by
2365 default unless the new "-noservername" option is used. The server name is
2366 based on the host provided to the "-connect" option unless overridden by
2367 using "-servername".
2371 * Add support for SipHash
2375 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2376 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2377 prevent issues where no progress is being made and the peer continually
2378 sends unrecognised record types, using up resources processing them.
2382 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2383 using the algorithm defined in
2384 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2388 * Heartbeat support has been removed; the ABI is changed for now.
2390 *Richard Levitte, Rich Salz*
2392 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2396 * The RSA "null" method, which was partially supported to avoid patent
2397 issues, has been replaced to always returns NULL.
2404 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2406 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2407 used even when parsing explicit parameters, when loading a encoded key
2408 or calling `EC_GROUP_new_from_ecpkparameters()`/
2409 `EC_GROUP_new_from_ecparameters()`.
2410 This prevents bypass of security hardening and performance gains,
2411 especially for curves with specialized EC_METHODs.
2412 By default, if a key encoded with explicit parameters is loaded and later
2413 encoded, the output is still encoded with explicit parameters, even if
2414 internally a "named" EC_GROUP is used for computation.
2418 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2419 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2420 NULL. After this change, only the cofactor parameter can be NULL. It also
2421 does some minimal sanity checks on the passed order.
2426 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2427 An attack is simple, if the first CMS_recipientInfo is valid but the
2428 second CMS_recipientInfo is chosen ciphertext. If the second
2429 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2430 encryption key will be replaced by garbage, and the message cannot be
2431 decoded, but if the RSA decryption fails, the correct encryption key is
2432 used and the recipient will not notice the attack.
2433 As a work around for this potential attack the length of the decrypted
2434 key must be equal to the cipher default key length, in case the
2435 certifiate is not given and all recipientInfo are tried out.
2436 The old behaviour can be re-enabled in the CMS code by setting the
2437 CMS_DEBUG_DECRYPT flag.
2442 * Use Windows installation paths in the mingw builds
2444 Mingw isn't a POSIX environment per se, which means that Windows
2445 paths should be used for installation.
2450 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2452 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2453 This changes the size when using the `genpkey` command when no size is given.
2454 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2455 generation commands to use 2048 bits by default.
2459 * Prevent over long nonces in ChaCha20-Poly1305.
2461 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2462 for every encryption operation. RFC 7539 specifies that the nonce value
2463 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2464 and front pads the nonce with 0 bytes if it is less than 12
2465 bytes. However it also incorrectly allows a nonce to be set of up to 16
2466 bytes. In this case only the last 12 bytes are significant and any
2467 additional leading bytes are ignored.
2469 It is a requirement of using this cipher that nonce values are
2470 unique. Messages encrypted using a reused nonce value are susceptible to
2471 serious confidentiality and integrity attacks. If an application changes
2472 the default nonce length to be longer than 12 bytes and then makes a
2473 change to the leading bytes of the nonce expecting the new value to be a
2474 new unique nonce then such an application could inadvertently encrypt
2475 messages with a reused nonce.
2477 Additionally the ignored bytes in a long nonce are not covered by the
2478 integrity guarantee of this cipher. Any application that relies on the
2479 integrity of these ignored leading bytes of a long nonce may be further
2480 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2481 is safe because no such use sets such a long nonce value. However user
2482 applications that use this cipher directly and set a non-default nonce
2483 length to be longer than 12 bytes may be vulnerable.
2485 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2491 * Added SCA hardening for modular field inversion in EC_GROUP through
2492 a new dedicated field_inv() pointer in EC_METHOD.
2493 This also addresses a leakage affecting conversions from projective
2494 to affine coordinates.
2496 *Billy Bob Brumley, Nicola Tuveri*
2498 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2499 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2503 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2507 * Remove the 'dist' target and add a tarball building script. The
2508 'dist' target has fallen out of use, and it shouldn't be
2509 necessary to configure just to create a source distribution.
2513 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2515 * Timing vulnerability in DSA signature generation
2517 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2518 timing side channel attack. An attacker could use variations in the signing
2519 algorithm to recover the private key.
2521 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2526 * Timing vulnerability in ECDSA signature generation
2528 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2529 timing side channel attack. An attacker could use variations in the signing
2530 algorithm to recover the private key.
2532 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2537 * Add coordinate blinding for EC_POINT and implement projective
2538 coordinate blinding for generic prime curves as a countermeasure to
2539 chosen point SCA attacks.
2541 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2543 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2545 * Client DoS due to large DH parameter
2547 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2548 malicious server can send a very large prime value to the client. This will
2549 cause the client to spend an unreasonably long period of time generating a
2550 key for this prime resulting in a hang until the client has finished. This
2551 could be exploited in a Denial Of Service attack.
2553 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2558 * Cache timing vulnerability in RSA Key Generation
2560 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2561 a cache timing side channel attack. An attacker with sufficient access to
2562 mount cache timing attacks during the RSA key generation process could
2563 recover the private key.
2565 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2566 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2571 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2572 parameter is no longer accepted, as it leads to a corrupt table. NULL
2573 pem_str is reserved for alias entries only.
2577 * Revert blinding in ECDSA sign and instead make problematic addition
2578 length-invariant. Switch even to fixed-length Montgomery multiplication.
2582 * Change generating and checking of primes so that the error rate of not
2583 being prime depends on the intended use based on the size of the input.
2584 For larger primes this will result in more rounds of Miller-Rabin.
2585 The maximal error rate for primes with more than 1080 bits is lowered
2588 *Kurt Roeckx, Annie Yousar*
2590 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2594 * Add blinding to ECDSA and DSA signatures to protect against side channel
2595 attacks discovered by Keegan Ryan (NCC Group).
2599 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2600 now allow empty (zero character) pass phrases.
2604 * Certificate time validation (X509_cmp_time) enforces stricter
2605 compliance with RFC 5280. Fractional seconds and timezone offsets
2606 are no longer allowed.
2610 * Fixed a text canonicalisation bug in CMS
2612 Where a CMS detached signature is used with text content the text goes
2613 through a canonicalisation process first prior to signing or verifying a
2614 signature. This process strips trailing space at the end of lines, converts
2615 line terminators to CRLF and removes additional trailing line terminators
2616 at the end of a file. A bug in the canonicalisation process meant that
2617 some characters, such as form-feed, were incorrectly treated as whitespace
2618 and removed. This is contrary to the specification (RFC5485). This fix
2619 could mean that detached text data signed with an earlier version of
2620 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2621 signed with a fixed OpenSSL may fail to verify with an earlier version of
2622 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2623 and use the "-binary" flag (for the "cms" command line application) or set
2624 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2628 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2630 * Constructed ASN.1 types with a recursive definition could exceed the stack
2632 Constructed ASN.1 types with a recursive definition (such as can be found
2633 in PKCS7) could eventually exceed the stack given malicious input with
2634 excessive recursion. This could result in a Denial Of Service attack. There
2635 are no such structures used within SSL/TLS that come from untrusted sources
2636 so this is considered safe.
2638 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2644 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2646 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2647 effectively reduced to only comparing the least significant bit of each
2648 byte. This allows an attacker to forge messages that would be considered as
2649 authenticated in an amount of tries lower than that guaranteed by the
2650 security claims of the scheme. The module can only be compiled by the
2651 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2653 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2659 * Add a build target 'build_all_generated', to build all generated files
2660 and only that. This can be used to prepare everything that requires
2661 things like perl for a system that lacks perl and then move everything
2662 to that system and do the rest of the build there.
2666 * Backport SSL_OP_NO_RENGOTIATION
2668 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2669 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2670 changes this is no longer possible in 1.1.0. Therefore the new
2671 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2672 1.1.0 to provide equivalent functionality.
2674 Note that if an application built against 1.1.0h headers (or above) is run
2675 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2676 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2680 * Removed the OS390-Unix config target. It relied on a script that doesn't
2685 * rsaz_1024_mul_avx2 overflow bug on x86_64
2687 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2688 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2689 Analysis suggests that attacks against RSA and DSA as a result of this
2690 defect would be very difficult to perform and are not believed likely.
2691 Attacks against DH1024 are considered just feasible, because most of the
2692 work necessary to deduce information about a private key may be performed
2693 offline. The amount of resources required for such an attack would be
2694 significant. However, for an attack on TLS to be meaningful, the server
2695 would have to share the DH1024 private key among multiple clients, which is
2696 no longer an option since CVE-2016-0701.
2698 This only affects processors that support the AVX2 but not ADX extensions
2699 like Intel Haswell (4th generation).
2701 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2702 was originally found via the OSS-Fuzz project.
2707 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2709 * bn_sqrx8x_internal carry bug on x86_64
2711 There is a carry propagating bug in the x86_64 Montgomery squaring
2712 procedure. No EC algorithms are affected. Analysis suggests that attacks
2713 against RSA and DSA as a result of this defect would be very difficult to
2714 perform and are not believed likely. Attacks against DH are considered just
2715 feasible (although very difficult) because most of the work necessary to
2716 deduce information about a private key may be performed offline. The amount
2717 of resources required for such an attack would be very significant and
2718 likely only accessible to a limited number of attackers. An attacker would
2719 additionally need online access to an unpatched system using the target
2720 private key in a scenario with persistent DH parameters and a private
2721 key that is shared between multiple clients.
2723 This only affects processors that support the BMI1, BMI2 and ADX extensions
2724 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2726 This issue was reported to OpenSSL by the OSS-Fuzz project.
2731 * Malformed X.509 IPAddressFamily could cause OOB read
2733 If an X.509 certificate has a malformed IPAddressFamily extension,
2734 OpenSSL could do a one-byte buffer overread. The most likely result
2735 would be an erroneous display of the certificate in text format.
2737 This issue was reported to OpenSSL by the OSS-Fuzz project.
2742 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2744 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2745 platform rather than 'mingw'.
2749 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2750 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2751 which is the minimum version we support.
2755 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2757 * Encrypt-Then-Mac renegotiation crash
2759 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2760 negotiated where it was not in the original handshake (or vice-versa) then
2761 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2762 and servers are affected.
2764 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2769 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2771 * Truncated packet could crash via OOB read
2773 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2774 cipher is being used, then a truncated packet can cause that host to
2775 perform an out-of-bounds read, usually resulting in a crash.
2777 This issue was reported to OpenSSL by Robert Święcki of Google.
2782 * Bad (EC)DHE parameters cause a client crash
2784 If a malicious server supplies bad parameters for a DHE or ECDHE key
2785 exchange then this can result in the client attempting to dereference a
2786 NULL pointer leading to a client crash. This could be exploited in a Denial
2789 This issue was reported to OpenSSL by Guido Vranken.
2794 * BN_mod_exp may produce incorrect results on x86_64
2796 There is a carry propagating bug in the x86_64 Montgomery squaring
2797 procedure. No EC algorithms are affected. Analysis suggests that attacks
2798 against RSA and DSA as a result of this defect would be very difficult to
2799 perform and are not believed likely. Attacks against DH are considered just
2800 feasible (although very difficult) because most of the work necessary to
2801 deduce information about a private key may be performed offline. The amount
2802 of resources required for such an attack would be very significant and
2803 likely only accessible to a limited number of attackers. An attacker would
2804 additionally need online access to an unpatched system using the target
2805 private key in a scenario with persistent DH parameters and a private
2806 key that is shared between multiple clients. For example this can occur by
2807 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2808 similar to CVE-2015-3193 but must be treated as a separate problem.
2810 This issue was reported to OpenSSL by the OSS-Fuzz project.
2815 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2817 * ChaCha20/Poly1305 heap-buffer-overflow
2819 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2820 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2821 crash. This issue is not considered to be exploitable beyond a DoS.
2823 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2828 * CMS Null dereference
2830 Applications parsing invalid CMS structures can crash with a NULL pointer
2831 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2832 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2833 structure callback if an attempt is made to free certain invalid encodings.
2834 Only CHOICE structures using a callback which do not handle NULL value are
2837 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2842 * Montgomery multiplication may produce incorrect results
2844 There is a carry propagating bug in the Broadwell-specific Montgomery
2845 multiplication procedure that handles input lengths divisible by, but
2846 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2847 and DH private keys are impossible. This is because the subroutine in
2848 question is not used in operations with the private key itself and an input
2849 of the attacker's direct choice. Otherwise the bug can manifest itself as
2850 transient authentication and key negotiation failures or reproducible
2851 erroneous outcome of public-key operations with specially crafted input.
2852 Among EC algorithms only Brainpool P-512 curves are affected and one
2853 presumably can attack ECDH key negotiation. Impact was not analyzed in
2854 detail, because pre-requisites for attack are considered unlikely. Namely
2855 multiple clients have to choose the curve in question and the server has to
2856 share the private key among them, neither of which is default behaviour.
2857 Even then only clients that chose the curve will be affected.
2859 This issue was publicly reported as transient failures and was not
2860 initially recognized as a security issue. Thanks to Richard Morgan for
2861 providing reproducible case.
2866 * Removed automatic addition of RPATH in shared libraries and executables,
2867 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2871 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2873 * Fix Use After Free for large message sizes
2875 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2876 message larger than approx 16k is received then the underlying buffer to
2877 store the incoming message is reallocated and moved. Unfortunately a
2878 dangling pointer to the old location is left which results in an attempt to
2879 write to the previously freed location. This is likely to result in a
2880 crash, however it could potentially lead to execution of arbitrary code.
2882 This issue only affects OpenSSL 1.1.0a.
2884 This issue was reported to OpenSSL by Robert Święcki.
2889 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2891 * OCSP Status Request extension unbounded memory growth
2893 A malicious client can send an excessively large OCSP Status Request
2894 extension. If that client continually requests renegotiation, sending a
2895 large OCSP Status Request extension each time, then there will be unbounded
2896 memory growth on the server. This will eventually lead to a Denial Of
2897 Service attack through memory exhaustion. Servers with a default
2898 configuration are vulnerable even if they do not support OCSP. Builds using
2899 the "no-ocsp" build time option are not affected.
2901 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2906 * SSL_peek() hang on empty record
2908 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2909 sends an empty record. This could be exploited by a malicious peer in a
2910 Denial Of Service attack.
2912 This issue was reported to OpenSSL by Alex Gaynor.
2917 * Excessive allocation of memory in tls_get_message_header() and
2918 dtls1_preprocess_fragment()
2920 A (D)TLS message includes 3 bytes for its length in the header for the
2921 message. This would allow for messages up to 16Mb in length. Messages of
2922 this length are excessive and OpenSSL includes a check to ensure that a
2923 peer is sending reasonably sized messages in order to avoid too much memory
2924 being consumed to service a connection. A flaw in the logic of version
2925 1.1.0 means that memory for the message is allocated too early, prior to
2926 the excessive message length check. Due to way memory is allocated in
2927 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2928 to service a connection. This could lead to a Denial of Service through
2929 memory exhaustion. However, the excessive message length check still takes
2930 place, and this would cause the connection to immediately fail. Assuming
2931 that the application calls SSL_free() on the failed connection in a timely
2932 manner then the 21Mb of allocated memory will then be immediately freed
2933 again. Therefore the excessive memory allocation will be transitory in
2934 nature. This then means that there is only a security impact if:
2936 1) The application does not call SSL_free() in a timely manner in the event
2937 that the connection fails
2939 2) The application is working in a constrained environment where there is
2940 very little free memory
2942 3) The attacker initiates multiple connection attempts such that there are
2943 multiple connections in a state where memory has been allocated for the
2944 connection; SSL_free() has not yet been called; and there is insufficient
2945 memory to service the multiple requests.
2947 Except in the instance of (1) above any Denial Of Service is likely to be
2948 transitory because as soon as the connection fails the memory is
2949 subsequently freed again in the SSL_free() call. However there is an
2950 increased risk during this period of application crashes due to the lack of
2951 memory - which would then mean a more serious Denial of Service.
2953 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2954 (CVE-2016-6307 and CVE-2016-6308)
2958 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2959 had to be removed. Primary reason is that vendor assembler can't
2960 assemble our modules with -KPIC flag. As result it, assembly
2961 support, was not even available as option. But its lack means
2962 lack of side-channel resistant code, which is incompatible with
2963 security by todays standards. Fortunately gcc is readily available
2964 prepackaged option, which we firmly point at...
2968 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2970 * Windows command-line tool supports UTF-8 opt-in option for arguments
2971 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2972 (to any value) allows Windows user to access PKCS#12 file generated
2973 with Windows CryptoAPI and protected with non-ASCII password, as well
2974 as files generated under UTF-8 locale on Linux also protected with
2979 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2980 have been disabled by default and removed from DEFAULT, just like RC4.
2981 See the RC4 item below to re-enable both.
2985 * The method for finding the storage location for the Windows RAND seed file
2986 has changed. First we check %RANDFILE%. If that is not set then we check
2987 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2988 all else fails we fall back to C:\.
2992 * The EVP_EncryptUpdate() function has had its return type changed from void
2993 to int. A return of 0 indicates and error while a return of 1 indicates
2998 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2999 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3000 off the constant time implementation for RSA, DSA and DH have been made
3001 no-ops and deprecated.
3005 * Windows RAND implementation was simplified to only get entropy by
3006 calling CryptGenRandom(). Various other RAND-related tickets
3009 *Joseph Wylie Yandle, Rich Salz*
3011 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3012 and `OPENSSL_LH_`, respectively. The old names are available
3013 with API compatibility. They new names are now completely documented.
3017 * Unify TYPE_up_ref(obj) methods signature.
3018 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3019 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3020 int (instead of void) like all others TYPE_up_ref() methods.
3021 So now these methods also check the return value of CRYPTO_atomic_add(),
3022 and the validity of object reference counter.
3024 *fdasilvayy@gmail.com*
3026 * With Windows Visual Studio builds, the .pdb files are installed
3027 alongside the installed libraries and executables. For a static
3028 library installation, ossl_static.pdb is the associate compiler
3029 generated .pdb file to be used when linking programs.
3033 * Remove openssl.spec. Packaging files belong with the packagers.
3037 * Automatic Darwin/OSX configuration has had a refresh, it will now
3038 recognise x86_64 architectures automatically. You can still decide
3039 to build for a different bitness with the environment variable
3040 KERNEL_BITS (can be 32 or 64), for example:
3042 KERNEL_BITS=32 ./config
3046 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3047 256 bit AES and HMAC with SHA256.
3051 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3055 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3059 * To enable users to have their own config files and build file templates,
3060 Configure looks in the directory indicated by the environment variable
3061 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3062 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3063 name and is used as is.
3067 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3068 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3069 X509_CERT_FILE_CTX was removed.
3073 * "shared" builds are now the default. To create only static libraries use
3074 the "no-shared" Configure option.
3078 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3079 All of these option have not worked for some while and are fundamental
3084 * Make various cleanup routines no-ops and mark them as deprecated. Most
3085 global cleanup functions are no longer required because they are handled
3086 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3087 Explicitly de-initing can cause problems (e.g. where a library that uses
3088 OpenSSL de-inits, but an application is still using it). The affected
3089 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3090 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3091 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3092 COMP_zlib_cleanup().
3096 * --strict-warnings no longer enables runtime debugging options
3097 such as REF_DEBUG. Instead, debug options are automatically
3098 enabled with '--debug' builds.
3100 *Andy Polyakov, Emilia Käsper*
3102 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3103 have been moved out of the public header files. New functions for managing
3104 these have been added.
3108 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3109 objects have been moved out of the public header files. New
3110 functions for managing these have been added.
3114 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3115 have been moved out of the public header files. New functions for managing
3116 these have been added.
3120 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3121 moved out of the public header files. New functions for managing these
3126 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3130 * Removed the mk1mf build scripts.
3134 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3135 it is always safe to #include a header now.
3139 * Removed the aged BC-32 config and all its supporting scripts
3143 * Removed support for Ultrix, Netware, and OS/2.
3147 * Add support for HKDF.
3149 *Alessandro Ghedini*
3151 * Add support for blake2b and blake2s
3155 * Added support for "pipelining". Ciphers that have the
3156 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3157 encryptions/decryptions simultaneously. There are currently no built-in
3158 ciphers with this property but the expectation is that engines will be able
3159 to offer it to significantly improve throughput. Support has been extended
3160 into libssl so that multiple records for a single connection can be
3161 processed in one go (for >=TLS 1.1).
3165 * Added the AFALG engine. This is an async capable engine which is able to
3166 offload work to the Linux kernel. In this initial version it only supports
3167 AES128-CBC. The kernel must be version 4.1.0 or greater.
3171 * OpenSSL now uses a new threading API. It is no longer necessary to
3172 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3173 are two supported threading models: pthreads and windows threads. It is
3174 also possible to configure OpenSSL at compile time for "no-threads". The
3175 old threading API should no longer be used. The functions have been
3176 replaced with "no-op" compatibility macros.
3178 *Alessandro Ghedini, Matt Caswell*
3180 * Modify behavior of ALPN to invoke callback after SNI/servername
3181 callback, such that updates to the SSL_CTX affect ALPN.
3185 * Add SSL_CIPHER queries for authentication and key-exchange.
3189 * Changes to the DEFAULT cipherlist:
3190 - Prefer (EC)DHE handshakes over plain RSA.
3191 - Prefer AEAD ciphers over legacy ciphers.
3192 - Prefer ECDSA over RSA when both certificates are available.
3193 - Prefer TLSv1.2 ciphers/PRF.
3194 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3199 * Change the ECC default curve list to be this, in order: x25519,
3200 secp256r1, secp521r1, secp384r1.
3204 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3205 disabled by default. They can be re-enabled using the
3206 enable-weak-ssl-ciphers option to Configure.
3210 * If the server has ALPN configured, but supports no protocols that the
3211 client advertises, send a fatal "no_application_protocol" alert.
3212 This behaviour is SHALL in RFC 7301, though it isn't universally
3213 implemented by other servers.
3217 * Add X25519 support.
3218 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3219 for public and private key encoding using the format documented in
3220 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3221 key generation and key derivation.
3223 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3228 * Deprecate SRP_VBASE_get_by_user.
3229 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3230 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3231 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3232 seed, even if the seed is configured.
3234 Users should use SRP_VBASE_get1_by_user instead. Note that in
3235 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3236 also that even though configuring the SRP seed attempts to hide
3237 invalid usernames by continuing the handshake with fake
3238 credentials, this behaviour is not constant time and no strong
3239 guarantees are made that the handshake is indistinguishable from
3240 that of a valid user.
3244 * Configuration change; it's now possible to build dynamic engines
3245 without having to build shared libraries and vice versa. This
3246 only applies to the engines in `engines/`, those in `crypto/engine/`
3247 will always be built into libcrypto (i.e. "static").
3249 Building dynamic engines is enabled by default; to disable, use
3250 the configuration option "disable-dynamic-engine".
3252 The only requirements for building dynamic engines are the
3253 presence of the DSO module and building with position independent
3254 code, so they will also automatically be disabled if configuring
3255 with "disable-dso" or "disable-pic".
3257 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3258 are also taken away from openssl/opensslconf.h, as they are
3263 * Configuration change; if there is a known flag to compile
3264 position independent code, it will always be applied on the
3265 libcrypto and libssl object files, and never on the application
3266 object files. This means other libraries that use routines from
3267 libcrypto / libssl can be made into shared libraries regardless
3268 of how OpenSSL was configured.
3270 If this isn't desirable, the configuration options "disable-pic"
3271 or "no-pic" can be used to disable the use of PIC. This will
3272 also disable building shared libraries and dynamic engines.
3276 * Removed JPAKE code. It was experimental and has no wide use.
3280 * The INSTALL_PREFIX Makefile variable has been renamed to
3281 DESTDIR. That makes for less confusion on what this variable
3282 is for. Also, the configuration option --install_prefix is
3287 * Heartbeat for TLS has been removed and is disabled by default
3288 for DTLS; configure with enable-heartbeats. Code that uses the
3289 old #define's might need to be updated.
3291 *Emilia Käsper, Rich Salz*
3293 * Rename REF_CHECK to REF_DEBUG.
3297 * New "unified" build system
3299 The "unified" build system is aimed to be a common system for all
3300 platforms we support. With it comes new support for VMS.
3302 This system builds supports building in a different directory tree
3303 than the source tree. It produces one Makefile (for unix family
3304 or lookalikes), or one descrip.mms (for VMS).
3306 The source of information to make the Makefile / descrip.mms is
3307 small files called 'build.info', holding the necessary
3308 information for each directory with source to compile, and a
3309 template in Configurations, like unix-Makefile.tmpl or
3312 With this change, the library names were also renamed on Windows
3313 and on VMS. They now have names that are closer to the standard
3314 on Unix, and include the major version number, and in certain
3315 cases, the architecture they are built for. See "Notes on shared
3316 libraries" in INSTALL.
3318 We rely heavily on the perl module Text::Template.
3322 * Added support for auto-initialisation and de-initialisation of the library.
3323 OpenSSL no longer requires explicit init or deinit routines to be called,
3324 except in certain circumstances. See the OPENSSL_init_crypto() and
3325 OPENSSL_init_ssl() man pages for further information.
3329 * The arguments to the DTLSv1_listen function have changed. Specifically the
3330 "peer" argument is now expected to be a BIO_ADDR object.
3332 * Rewrite of BIO networking library. The BIO library lacked consistent
3333 support of IPv6, and adding it required some more extensive
3334 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3335 which hold all types of addresses and chains of address information.
3336 It also introduces a new API, with functions like BIO_socket,
3337 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3338 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3339 have been adapted accordingly.
3343 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3348 * CRIME protection: disable compression by default, even if OpenSSL is
3349 compiled with zlib enabled. Applications can still enable compression
3350 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3351 using the SSL_CONF library to configure compression.
3355 * The signature of the session callback configured with
3356 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3357 was explicitly marked as `const unsigned char*` instead of
3362 * Always DPURIFY. Remove the use of uninitialized memory in the
3363 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3367 * Removed many obsolete configuration items, including
3368 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3369 MD2_CHAR, MD2_INT, MD2_LONG
3371 IDEA_SHORT, IDEA_LONG
3372 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3374 *Rich Salz, with advice from Andy Polyakov*
3376 * Many BN internals have been moved to an internal header file.
3378 *Rich Salz with help from Andy Polyakov*
3380 * Configuration and writing out the results from it has changed.
3381 Files such as Makefile include/openssl/opensslconf.h and are now
3382 produced through general templates, such as Makefile.in and
3383 crypto/opensslconf.h.in and some help from the perl module
3386 Also, the center of configuration information is no longer
3387 Makefile. Instead, Configure produces a perl module in
3388 configdata.pm which holds most of the config data (in the hash
3389 table %config), the target data that comes from the target
3390 configuration in one of the `Configurations/*.conf` files (in
3395 * To clarify their intended purposes, the Configure options
3396 --prefix and --openssldir change their semantics, and become more
3397 straightforward and less interdependent.
3399 --prefix shall be used exclusively to give the location INSTALLTOP
3400 where programs, scripts, libraries, include files and manuals are
3401 going to be installed. The default is now /usr/local.
3403 --openssldir shall be used exclusively to give the default
3404 location OPENSSLDIR where certificates, private keys, CRLs are
3405 managed. This is also where the default openssl.cnf gets
3407 If the directory given with this option is a relative path, the
3408 values of both the --prefix value and the --openssldir value will
3409 be combined to become OPENSSLDIR.
3410 The default for --openssldir is INSTALLTOP/ssl.
3412 Anyone who uses --openssldir to specify where OpenSSL is to be
3413 installed MUST change to use --prefix instead.
3417 * The GOST engine was out of date and therefore it has been removed. An up
3418 to date GOST engine is now being maintained in an external repository.
3419 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3420 support for GOST ciphersuites (these are only activated if a GOST engine
3425 * EGD is no longer supported by default; use enable-egd when
3428 *Ben Kaduk and Rich Salz*
3430 * The distribution now has Makefile.in files, which are used to
3431 create Makefile's when Configure is run. *Configure must be run
3432 before trying to build now.*
3436 * The return value for SSL_CIPHER_description() for error conditions
3441 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3443 Obtaining and performing DNSSEC validation of TLSA records is
3444 the application's responsibility. The application provides
3445 the TLSA records of its choice to OpenSSL, and these are then
3446 used to authenticate the peer.
3448 The TLSA records need not even come from DNS. They can, for
3449 example, be used to implement local end-entity certificate or
3450 trust-anchor "pinning", where the "pin" data takes the form
3451 of TLSA records, which can augment or replace verification
3452 based on the usual WebPKI public certification authorities.
3456 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3457 continues to support deprecated interfaces in default builds.
3458 However, applications are strongly advised to compile their
3459 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3460 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3461 or the 1.1.0 releases.
3463 In environments in which all applications have been ported to
3464 not use any deprecated interfaces OpenSSL's Configure script
3465 should be used with the --api=1.1.0 option to entirely remove
3466 support for the deprecated features from the library and
3467 unconditionally disable them in the installed headers.
3468 Essentially the same effect can be achieved with the "no-deprecated"
3469 argument to Configure, except that this will always restrict
3470 the build to just the latest API, rather than a fixed API
3473 As applications are ported to future revisions of the API,
3474 they should update their compile-time OPENSSL_API_COMPAT define
3475 accordingly, but in most cases should be able to continue to
3476 compile with later releases.
3478 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3479 0x10000000L and 0x00908000L, respectively. However those
3480 versions did not support the OPENSSL_API_COMPAT feature, and
3481 so applications are not typically tested for explicit support
3482 of just the undeprecated features of either release.
3486 * Add support for setting the minimum and maximum supported protocol.
3487 It can bet set via the SSL_set_min_proto_version() and
3488 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3489 MaxProtocol. It's recommended to use the new APIs to disable
3490 protocols instead of disabling individual protocols using
3491 SSL_set_options() or SSL_CONF's Protocol. This change also
3492 removes support for disabling TLS 1.2 in the OpenSSL TLS
3493 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3497 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3501 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3502 and integrates ECDSA and ECDH functionality into EC. Implementations can
3503 now redirect key generation and no longer need to convert to or from
3506 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3507 include the ec.h header file instead.
3511 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3512 ciphers who are no longer supported and drops support the ephemeral RSA key
3513 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3517 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3518 opaque. For HMAC_CTX, the following constructors and destructors
3521 HMAC_CTX *HMAC_CTX_new(void);
3522 void HMAC_CTX_free(HMAC_CTX *ctx);
3524 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3525 destroy such methods has been added. See EVP_MD_meth_new(3) and
3526 EVP_CIPHER_meth_new(3) for documentation.
3529 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3530 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3531 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3532 an already created structure.
3533 2) For consistency with the majority of our object creators and
3534 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3535 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3536 for deprecated builds.
3540 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3541 cryptographic operations to be performed asynchronously as long as an
3542 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3543 further details. Libssl has also had this capability integrated with the
3544 introduction of the new mode SSL_MODE_ASYNC and associated error
3545 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3546 pages. This work was developed in partnership with Intel Corp.
3550 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3551 always enabled now. If you want to disable the support you should
3552 exclude it using the list of supported ciphers. This also means that the
3553 "-no_ecdhe" option has been removed from s_server.
3557 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3558 SSL_{CTX_}set1_curves() which can set a list.
3562 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3563 curve you want to support using SSL_{CTX_}set1_curves().
3567 * State machine rewrite. The state machine code has been significantly
3568 refactored in order to remove much duplication of code and solve issues
3569 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3570 further details). This change does have some associated API changes.
3571 Notably the SSL_state() function has been removed and replaced by
3572 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3573 SSL_set_state() has been removed altogether. The previous handshake states
3574 defined in ssl.h and ssl3.h have also been removed.
3578 * All instances of the string "ssleay" in the public API were replaced
3579 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3580 Some error codes related to internal RSA_eay API's were renamed.
3584 * The demo files in crypto/threads were moved to demo/threads.
3588 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3591 *Matt Caswell, Rich Salz*
3593 * New ASN.1 embed macro.
3595 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3596 structure is not allocated: it is part of the parent. That is instead of
3604 This reduces memory fragmentation and make it impossible to accidentally
3605 set a mandatory field to NULL.
3607 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3608 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3609 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3614 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3618 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3619 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3620 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3621 DES and RC4 ciphersuites.
3625 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3626 This changes the decoding behaviour for some invalid messages,
3627 though the change is mostly in the more lenient direction, and
3628 legacy behaviour is preserved as much as possible.
3632 * Fix no-stdio build.
3633 *David Woodhouse <David.Woodhouse@intel.com> and also*
3634 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3636 * New testing framework
3637 The testing framework has been largely rewritten and is now using
3638 perl and the perl modules Test::Harness and an extended variant of
3639 Test::More called OpenSSL::Test to do its work. All test scripts in
3640 test/ have been rewritten into test recipes, and all direct calls to
3641 executables in test/Makefile have become individual recipes using the
3642 simplified testing OpenSSL::Test::Simple.
3644 For documentation on our testing modules, do:
3646 perldoc test/testlib/OpenSSL/Test/Simple.pm
3647 perldoc test/testlib/OpenSSL/Test.pm
3651 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3652 are used; the latter aborts on memory leaks (usually checked on exit).
3653 Some undocumented "set malloc, etc., hooks" functions were removed
3654 and others were changed. All are now documented.
3658 * In DSA_generate_parameters_ex, if the provided seed is too short,
3661 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3663 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3664 from RFC4279, RFC4785, RFC5487, RFC5489.
3666 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3667 original RSA_PSK patch.
3671 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3672 era flag was never set throughout the codebase (only read). Also removed
3673 SSL3_FLAGS_POP_BUFFER which was only used if
3674 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3678 * Changed the default name options in the "ca", "crl", "req" and "x509"
3679 to be "oneline" instead of "compat".
3683 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3684 not aware of clients that still exhibit this bug, and the workaround
3685 hasn't been working properly for a while.
3689 * The return type of BIO_number_read() and BIO_number_written() as well as
3690 the corresponding num_read and num_write members in the BIO structure has
3691 changed from unsigned long to uint64_t. On platforms where an unsigned
3692 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3697 * Given the pervasive nature of TLS extensions it is inadvisable to run
3698 OpenSSL without support for them. It also means that maintaining
3699 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3700 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3704 * Removed support for the two export grade static DH ciphersuites
3705 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3706 were newly added (along with a number of other static DH ciphersuites) to
3707 1.0.2. However the two export ones have *never* worked since they were
3708 introduced. It seems strange in any case to be adding new export
3709 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3713 * Version negotiation has been rewritten. In particular SSLv23_method(),
3714 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3715 and turned into macros which simply call the new preferred function names
3716 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3717 should use the new names instead. Also as part of this change the ssl23.h
3718 header file has been removed.
3722 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3723 code and the associated standard is no longer considered fit-for-purpose.
3727 * RT2547 was closed. When generating a private key, try to make the
3728 output file readable only by the owner. This behavior change might
3729 be noticeable when interacting with other software.
3731 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3736 * Added HTTP GET support to the ocsp command.
3740 * Changed default digest for the dgst and enc commands from MD5 to
3745 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3749 * Added support for TLS extended master secret from
3750 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3751 initial patch which was a great help during development.
3755 * All libssl internal structures have been removed from the public header
3756 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3757 now redundant). Users should not attempt to access internal structures
3758 directly. Instead they should use the provided API functions.
3762 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3763 Access to deprecated functions can be re-enabled by running config with
3764 "enable-deprecated". In addition applications wishing to use deprecated
3765 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3766 will, by default, disable some transitive includes that previously existed
3767 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3771 * Added support for OCB mode. OpenSSL has been granted a patent license
3772 compatible with the OpenSSL license for use of OCB. Details are available
3773 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3774 for OCB can be removed by calling config with no-ocb.
3778 * SSLv2 support has been removed. It still supports receiving a SSLv2
3779 compatible client hello.
3783 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3784 done while fixing the error code for the key-too-small case.
3786 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3788 * CA.sh has been removed; use CA.pl instead.
3792 * Removed old DES API.
3796 * Remove various unsupported platforms:
3802 Sinix/ReliantUNIX RM400
3807 16-bit platforms such as WIN16
3811 * Clean up OPENSSL_NO_xxx #define's
3812 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3813 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3814 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3815 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3816 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3817 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3818 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3819 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3820 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3821 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3825 * Cleaned up dead code
3826 Remove all but one '#ifdef undef' which is to be looked at.
3830 * Clean up calling of xxx_free routines.
3831 Just like free(), fix most of the xxx_free routines to accept
3832 NULL. Remove the non-null checks from callers. Save much code.
3836 * Add secure heap for storage of private keys (when possible).
3837 Add BIO_s_secmem(), CBIGNUM, etc.
3838 Contributed by Akamai Technologies under our Corporate CLA.
3842 * Experimental support for a new, fast, unbiased prime candidate generator,
3843 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3845 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3847 * New output format NSS in the sess_id command line tool. This allows
3848 exporting the session id and the master key in NSS keylog format.
3850 *Martin Kaiser <martin@kaiser.cx>*
3852 * Harmonize version and its documentation. -f flag is used to display
3855 *mancha <mancha1@zoho.com>*
3857 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3858 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3860 *mancha <mancha1@zoho.com>*
3862 * Fix some double frees. These are not thought to be exploitable.
3864 *mancha <mancha1@zoho.com>*
3866 * A missing bounds check in the handling of the TLS heartbeat extension
3867 can be used to reveal up to 64k of memory to a connected client or
3870 Thanks for Neel Mehta of Google Security for discovering this bug and to
3871 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3872 preparing the fix ([CVE-2014-0160])
3874 *Adam Langley, Bodo Moeller*
3876 * Fix for the attack described in the paper "Recovering OpenSSL
3877 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3878 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3879 <http://eprint.iacr.org/2014/140>
3881 Thanks to Yuval Yarom and Naomi Benger for discovering this
3882 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3884 *Yuval Yarom and Naomi Benger*
3886 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3887 this fixes a limitation in previous versions of OpenSSL.
3891 * Experimental encrypt-then-mac support.
3893 Experimental support for encrypt then mac from
3894 draft-gutmann-tls-encrypt-then-mac-02.txt
3896 To enable it set the appropriate extension number (0x42 for the test
3897 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3899 For non-compliant peers (i.e. just about everything) this should have no
3902 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3906 * Add EVP support for key wrapping algorithms, to avoid problems with
3907 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3908 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3909 algorithms and include tests cases.
3913 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3918 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3919 MGF1 digest and OAEP label.
3923 * Make openssl verify return errors.
3925 *Chris Palmer <palmer@google.com> and Ben Laurie*
3927 * New function ASN1_TIME_diff to calculate the difference between two
3928 ASN1_TIME structures or one structure and the current time.
3932 * Update fips_test_suite to support multiple command line options. New
3933 test to induce all self test errors in sequence and check expected
3938 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3939 sign or verify all in one operation.
3943 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3944 test programs and fips_test_suite. Includes functionality to parse
3945 the minimal script output of fipsalgest.pl directly.
3949 * Add authorisation parameter to FIPS_module_mode_set().
3953 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3957 * Use separate DRBG fields for internal and external flags. New function
3958 FIPS_drbg_health_check() to perform on demand health checking. Add
3959 generation tests to fips_test_suite with reduced health check interval to
3960 demonstrate periodic health checking. Add "nodh" option to
3961 fips_test_suite to skip very slow DH test.
3965 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3970 * More extensive health check for DRBG checking many more failure modes.
3971 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3972 combination: call this in fips_test_suite.
3976 * Add support for canonical generation of DSA parameter 'g'. See
3979 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3980 POST to handle HMAC cases.
3984 * Add functions FIPS_module_version() and FIPS_module_version_text()
3985 to return numerical and string versions of the FIPS module number.
3989 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3990 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3991 outside the validated module in the FIPS capable OpenSSL.
3995 * Minor change to DRBG entropy callback semantics. In some cases
3996 there is no multiple of the block length between min_len and
3997 max_len. Allow the callback to return more than max_len bytes
3998 of entropy but discard any extra: it is the callback's responsibility
3999 to ensure that the extra data discarded does not impact the
4000 requested amount of entropy.
4004 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4005 information in FIPS186-3, SP800-57 and SP800-131A.
4009 * CCM support via EVP. Interface is very similar to GCM case except we
4010 must supply all data in one chunk (i.e. no update, final) and the
4011 message length must be supplied if AAD is used. Add algorithm test
4016 * Initial version of POST overhaul. Add POST callback to allow the status
4017 of POST to be monitored and/or failures induced. Modify fips_test_suite
4018 to use callback. Always run all selftests even if one fails.
4022 * XTS support including algorithm test driver in the fips_gcmtest program.
4023 Note: this does increase the maximum key length from 32 to 64 bytes but
4024 there should be no binary compatibility issues as existing applications
4025 will never use XTS mode.
4029 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4030 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4031 performs algorithm blocking for unapproved PRNG types. Also do not
4032 set PRNG type in FIPS_mode_set(): leave this to the application.
4033 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4034 the standard OpenSSL PRNG: set additional data to a date time vector.
4038 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4039 This shouldn't present any incompatibility problems because applications
4040 shouldn't be using these directly and any that are will need to rethink
4041 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4045 * Extensive self tests and health checking required by SP800-90 DRBG.
4046 Remove strength parameter from FIPS_drbg_instantiate and always
4047 instantiate at maximum supported strength.
4051 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4055 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4059 * New function DH_compute_key_padded() to compute a DH key and pad with
4060 leading zeroes if needed: this complies with SP800-56A et al.
4064 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4065 anything, incomplete, subject to change and largely untested at present.
4069 * Modify fipscanisteronly build option to only build the necessary object
4070 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4074 * Add experimental option FIPSSYMS to give all symbols in
4075 fipscanister.o and FIPS or fips prefix. This will avoid
4076 conflicts with future versions of OpenSSL. Add perl script
4077 util/fipsas.pl to preprocess assembly language source files
4078 and rename any affected symbols.
4082 * Add selftest checks and algorithm block of non-fips algorithms in
4083 FIPS mode. Remove DES2 from selftests.
4087 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4088 return internal method without any ENGINE dependencies. Add new
4089 tiny fips sign and verify functions.
4093 * New build option no-ec2m to disable characteristic 2 code.
4097 * New build option "fipscanisteronly". This only builds fipscanister.o
4098 and (currently) associated fips utilities. Uses the file Makefile.fips
4099 instead of Makefile.org as the prototype.
4103 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4104 Update fips_gcmtest to use IV generator.
4108 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4109 setting output buffer to NULL. The `*Final` function must be
4110 called although it will not retrieve any additional data. The tag
4111 can be set or retrieved with a ctrl. The IV length is by default 12
4112 bytes (96 bits) but can be set to an alternative value. If the IV
4113 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4118 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4119 underlying do_cipher function handles all cipher semantics itself
4120 including padding and finalisation. This is useful if (for example)
4121 an ENGINE cipher handles block padding itself. The behaviour of
4122 do_cipher is subtly changed if this flag is set: the return value
4123 is the number of characters written to the output buffer (zero is
4124 no longer an error code) or a negative error code. Also if the
4125 input buffer is NULL and length 0 finalisation should be performed.
4129 * If a candidate issuer certificate is already part of the constructed
4130 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4134 * Improve forward-security support: add functions
4136 void SSL_CTX_set_not_resumable_session_callback(
4137 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4138 void SSL_set_not_resumable_session_callback(
4139 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4141 for use by SSL/TLS servers; the callback function will be called whenever a
4142 new session is created, and gets to decide whether the session may be
4143 cached to make it resumable (return 0) or not (return 1). (As by the
4144 SSL/TLS protocol specifications, the session_id sent by the server will be
4145 empty to indicate that the session is not resumable; also, the server will
4146 not generate RFC 4507 (RFC 5077) session tickets.)
4148 A simple reasonable callback implementation is to return is_forward_secure.
4149 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4150 by the SSL/TLS server library, indicating whether it can provide forward
4153 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4155 * New -verify_name option in command line utilities to set verification
4160 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4161 Add CMAC pkey methods.
4165 * Experimental renegotiation in s_server -www mode. If the client
4166 browses /reneg connection is renegotiated. If /renegcert it is
4167 renegotiated requesting a certificate.
4171 * Add an "external" session cache for debugging purposes to s_server. This
4172 should help trace issues which normally are only apparent in deployed
4173 multi-process servers.
4177 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4178 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4179 BIO_set_cipher() and some obscure PEM functions were changed so they
4180 can now return an error. The RAND changes required a change to the
4181 RAND_METHOD structure.
4185 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4186 a gcc attribute to warn if the result of a function is ignored. This
4187 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4188 whose return value is often ignored.
4192 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4193 These allow SCTs (signed certificate timestamps) to be requested and
4194 validated when establishing a connection.
4196 *Rob Percival <robpercival@google.com>*
4201 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4203 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4204 used even when parsing explicit parameters, when loading a encoded key
4205 or calling `EC_GROUP_new_from_ecpkparameters()`/
4206 `EC_GROUP_new_from_ecparameters()`.
4207 This prevents bypass of security hardening and performance gains,
4208 especially for curves with specialized EC_METHODs.
4209 By default, if a key encoded with explicit parameters is loaded and later
4210 encoded, the output is still encoded with explicit parameters, even if
4211 internally a "named" EC_GROUP is used for computation.
4215 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4216 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4217 NULL. After this change, only the cofactor parameter can be NULL. It also
4218 does some minimal sanity checks on the passed order.
4223 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4224 An attack is simple, if the first CMS_recipientInfo is valid but the
4225 second CMS_recipientInfo is chosen ciphertext. If the second
4226 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4227 encryption key will be replaced by garbage, and the message cannot be
4228 decoded, but if the RSA decryption fails, the correct encryption key is
4229 used and the recipient will not notice the attack.
4230 As a work around for this potential attack the length of the decrypted
4231 key must be equal to the cipher default key length, in case the
4232 certifiate is not given and all recipientInfo are tried out.
4233 The old behaviour can be re-enabled in the CMS code by setting the
4234 CMS_DEBUG_DECRYPT flag.
4239 * Document issue with installation paths in diverse Windows builds
4241 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4242 binaries and run-time config file.
4247 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4249 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4250 This changes the size when using the `genpkey` command when no size is given.
4251 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4252 generation commands to use 2048 bits by default.
4256 * Add FIPS support for Android Arm 64-bit
4258 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4259 Module in Version 2.0.10. For some reason, the corresponding target
4260 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4261 built with FIPS support on Android Arm 64-bit. This omission has been
4264 *Matthias St. Pierre*
4266 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4268 * 0-byte record padding oracle
4270 If an application encounters a fatal protocol error and then calls
4271 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4272 then OpenSSL can respond differently to the calling application if a 0 byte
4273 record is received with invalid padding compared to if a 0 byte record is
4274 received with an invalid MAC. If the application then behaves differently
4275 based on that in a way that is detectable to the remote peer, then this
4276 amounts to a padding oracle that could be used to decrypt data.
4278 In order for this to be exploitable "non-stitched" ciphersuites must be in
4279 use. Stitched ciphersuites are optimised implementations of certain
4280 commonly used ciphersuites. Also the application must call SSL_shutdown()
4281 twice even if a protocol error has occurred (applications should not do
4282 this but some do anyway).
4284 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4285 Aviram, with additional investigation by Steven Collison and Andrew
4286 Hourselt. It was reported to OpenSSL on 10th December 2018.
4291 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4295 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4297 * Microarchitecture timing vulnerability in ECC scalar multiplication
4299 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4300 shown to be vulnerable to a microarchitecture timing side channel attack.
4301 An attacker with sufficient access to mount local timing attacks during
4302 ECDSA signature generation could recover the private key.
4304 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4305 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4311 * Timing vulnerability in DSA signature generation
4313 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4314 timing side channel attack. An attacker could use variations in the signing
4315 algorithm to recover the private key.
4317 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4322 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4323 Module, accidentally introduced while backporting security fixes from the
4324 development branch and hindering the use of ECC in FIPS mode.
4328 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4330 * Client DoS due to large DH parameter
4332 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4333 malicious server can send a very large prime value to the client. This will
4334 cause the client to spend an unreasonably long period of time generating a
4335 key for this prime resulting in a hang until the client has finished. This
4336 could be exploited in a Denial Of Service attack.
4338 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4343 * Cache timing vulnerability in RSA Key Generation
4345 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4346 a cache timing side channel attack. An attacker with sufficient access to
4347 mount cache timing attacks during the RSA key generation process could
4348 recover the private key.
4350 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4351 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4356 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4357 parameter is no longer accepted, as it leads to a corrupt table. NULL
4358 pem_str is reserved for alias entries only.
4362 * Revert blinding in ECDSA sign and instead make problematic addition
4363 length-invariant. Switch even to fixed-length Montgomery multiplication.
4367 * Change generating and checking of primes so that the error rate of not
4368 being prime depends on the intended use based on the size of the input.
4369 For larger primes this will result in more rounds of Miller-Rabin.
4370 The maximal error rate for primes with more than 1080 bits is lowered
4373 *Kurt Roeckx, Annie Yousar*
4375 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4379 * Add blinding to ECDSA and DSA signatures to protect against side channel
4380 attacks discovered by Keegan Ryan (NCC Group).
4384 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4385 now allow empty (zero character) pass phrases.
4389 * Certificate time validation (X509_cmp_time) enforces stricter
4390 compliance with RFC 5280. Fractional seconds and timezone offsets
4391 are no longer allowed.
4395 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4397 * Constructed ASN.1 types with a recursive definition could exceed the stack
4399 Constructed ASN.1 types with a recursive definition (such as can be found
4400 in PKCS7) could eventually exceed the stack given malicious input with
4401 excessive recursion. This could result in a Denial Of Service attack. There
4402 are no such structures used within SSL/TLS that come from untrusted sources
4403 so this is considered safe.
4405 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4411 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4413 * Read/write after SSL object in error state
4415 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4416 mechanism. The intent was that if a fatal error occurred during a handshake
4417 then OpenSSL would move into the error state and would immediately fail if
4418 you attempted to continue the handshake. This works as designed for the
4419 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4420 SSL_connect()), however due to a bug it does not work correctly if
4421 SSL_read() or SSL_write() is called directly. In that scenario, if the
4422 handshake fails then a fatal error will be returned in the initial function
4423 call. If SSL_read()/SSL_write() is subsequently called by the application
4424 for the same SSL object then it will succeed and the data is passed without
4425 being decrypted/encrypted directly from the SSL/TLS record layer.
4427 In order to exploit this issue an application bug would have to be present
4428 that resulted in a call to SSL_read()/SSL_write() being issued after having
4429 already received a fatal error.
4431 This issue was reported to OpenSSL by David Benjamin (Google).
4436 * rsaz_1024_mul_avx2 overflow bug on x86_64
4438 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4439 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4440 Analysis suggests that attacks against RSA and DSA as a result of this
4441 defect would be very difficult to perform and are not believed likely.
4442 Attacks against DH1024 are considered just feasible, because most of the
4443 work necessary to deduce information about a private key may be performed
4444 offline. The amount of resources required for such an attack would be
4445 significant. However, for an attack on TLS to be meaningful, the server
4446 would have to share the DH1024 private key among multiple clients, which is
4447 no longer an option since CVE-2016-0701.
4449 This only affects processors that support the AVX2 but not ADX extensions
4450 like Intel Haswell (4th generation).
4452 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4453 was originally found via the OSS-Fuzz project.
4458 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4460 * bn_sqrx8x_internal carry bug on x86_64
4462 There is a carry propagating bug in the x86_64 Montgomery squaring
4463 procedure. No EC algorithms are affected. Analysis suggests that attacks
4464 against RSA and DSA as a result of this defect would be very difficult to
4465 perform and are not believed likely. Attacks against DH are considered just
4466 feasible (although very difficult) because most of the work necessary to
4467 deduce information about a private key may be performed offline. The amount
4468 of resources required for such an attack would be very significant and
4469 likely only accessible to a limited number of attackers. An attacker would
4470 additionally need online access to an unpatched system using the target
4471 private key in a scenario with persistent DH parameters and a private
4472 key that is shared between multiple clients.
4474 This only affects processors that support the BMI1, BMI2 and ADX extensions
4475 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4477 This issue was reported to OpenSSL by the OSS-Fuzz project.
4482 * Malformed X.509 IPAddressFamily could cause OOB read
4484 If an X.509 certificate has a malformed IPAddressFamily extension,
4485 OpenSSL could do a one-byte buffer overread. The most likely result
4486 would be an erroneous display of the certificate in text format.
4488 This issue was reported to OpenSSL by the OSS-Fuzz project.
4493 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4495 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4496 platform rather than 'mingw'.
4500 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4502 * Truncated packet could crash via OOB read
4504 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4505 cipher is being used, then a truncated packet can cause that host to
4506 perform an out-of-bounds read, usually resulting in a crash.
4508 This issue was reported to OpenSSL by Robert Święcki of Google.
4513 * BN_mod_exp may produce incorrect results on x86_64
4515 There is a carry propagating bug in the x86_64 Montgomery squaring
4516 procedure. No EC algorithms are affected. Analysis suggests that attacks
4517 against RSA and DSA as a result of this defect would be very difficult to
4518 perform and are not believed likely. Attacks against DH are considered just
4519 feasible (although very difficult) because most of the work necessary to
4520 deduce information about a private key may be performed offline. The amount
4521 of resources required for such an attack would be very significant and
4522 likely only accessible to a limited number of attackers. An attacker would
4523 additionally need online access to an unpatched system using the target
4524 private key in a scenario with persistent DH parameters and a private
4525 key that is shared between multiple clients. For example this can occur by
4526 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4527 similar to CVE-2015-3193 but must be treated as a separate problem.
4529 This issue was reported to OpenSSL by the OSS-Fuzz project.
4534 * Montgomery multiplication may produce incorrect results
4536 There is a carry propagating bug in the Broadwell-specific Montgomery
4537 multiplication procedure that handles input lengths divisible by, but
4538 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4539 and DH private keys are impossible. This is because the subroutine in
4540 question is not used in operations with the private key itself and an input
4541 of the attacker's direct choice. Otherwise the bug can manifest itself as
4542 transient authentication and key negotiation failures or reproducible
4543 erroneous outcome of public-key operations with specially crafted input.
4544 Among EC algorithms only Brainpool P-512 curves are affected and one
4545 presumably can attack ECDH key negotiation. Impact was not analyzed in
4546 detail, because pre-requisites for attack are considered unlikely. Namely
4547 multiple clients have to choose the curve in question and the server has to
4548 share the private key among them, neither of which is default behaviour.
4549 Even then only clients that chose the curve will be affected.
4551 This issue was publicly reported as transient failures and was not
4552 initially recognized as a security issue. Thanks to Richard Morgan for
4553 providing reproducible case.
4558 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4559 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4560 prevent issues where no progress is being made and the peer continually
4561 sends unrecognised record types, using up resources processing them.
4565 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4567 * Missing CRL sanity check
4569 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4570 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4571 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4573 This issue only affects the OpenSSL 1.0.2i
4578 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4580 * OCSP Status Request extension unbounded memory growth
4582 A malicious client can send an excessively large OCSP Status Request
4583 extension. If that client continually requests renegotiation, sending a
4584 large OCSP Status Request extension each time, then there will be unbounded
4585 memory growth on the server. This will eventually lead to a Denial Of
4586 Service attack through memory exhaustion. Servers with a default
4587 configuration are vulnerable even if they do not support OCSP. Builds using
4588 the "no-ocsp" build time option are not affected.
4590 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4595 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4598 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4604 * OOB write in MDC2_Update()
4606 An overflow can occur in MDC2_Update() either if called directly or
4607 through the EVP_DigestUpdate() function using MDC2. If an attacker
4608 is able to supply very large amounts of input data after a previous
4609 call to EVP_EncryptUpdate() with a partial block then a length check
4610 can overflow resulting in a heap corruption.
4612 The amount of data needed is comparable to SIZE_MAX which is impractical
4615 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4620 * Malformed SHA512 ticket DoS
4622 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4623 DoS attack where a malformed ticket will result in an OOB read which will
4626 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4627 a custom server callback and ticket lookup mechanism.
4629 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4634 * OOB write in BN_bn2dec()
4636 The function BN_bn2dec() does not check the return value of BN_div_word().
4637 This can cause an OOB write if an application uses this function with an
4638 overly large BIGNUM. This could be a problem if an overly large certificate
4639 or CRL is printed out from an untrusted source. TLS is not affected because
4640 record limits will reject an oversized certificate before it is parsed.
4642 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4647 * OOB read in TS_OBJ_print_bio()
4649 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4650 the total length the OID text representation would use and not the amount
4651 of data written. This will result in OOB reads when large OIDs are
4654 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4659 * Pointer arithmetic undefined behaviour
4661 Avoid some undefined pointer arithmetic
4663 A common idiom in the codebase is to check limits in the following manner:
4666 Where "p" points to some malloc'd data of SIZE bytes and
4669 "len" here could be from some externally supplied data (e.g. from a TLS
4672 The rules of C pointer arithmetic are such that "p + len" is only well
4673 defined where len <= SIZE. Therefore the above idiom is actually
4674 undefined behaviour.
4676 For example this could cause problems if some malloc implementation
4677 provides an address for "p" such that "p + len" actually overflows for
4678 values of len that are too big and therefore p + len < limit.
4680 This issue was reported to OpenSSL by Guido Vranken
4685 * Constant time flag not preserved in DSA signing
4687 Operations in the DSA signing algorithm should run in constant time in
4688 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4689 implementation means that a non-constant time codepath is followed for
4690 certain operations. This has been demonstrated through a cache-timing
4691 attack to be sufficient for an attacker to recover the private DSA key.
4693 This issue was reported by César Pereida (Aalto University), Billy Brumley
4694 (Tampere University of Technology), and Yuval Yarom (The University of
4695 Adelaide and NICTA).
4700 * DTLS buffered message DoS
4702 In a DTLS connection where handshake messages are delivered out-of-order
4703 those messages that OpenSSL is not yet ready to process will be buffered
4704 for later use. Under certain circumstances, a flaw in the logic means that
4705 those messages do not get removed from the buffer even though the handshake
4706 has been completed. An attacker could force up to approx. 15 messages to
4707 remain in the buffer when they are no longer required. These messages will
4708 be cleared when the DTLS connection is closed. The default maximum size for
4709 a message is 100k. Therefore the attacker could force an additional 1500k
4710 to be consumed per connection. By opening many simulataneous connections an
4711 attacker could cause a DoS attack through memory exhaustion.
4713 This issue was reported to OpenSSL by Quan Luo.
4718 * DTLS replay protection DoS
4720 A flaw in the DTLS replay attack protection mechanism means that records
4721 that arrive for future epochs update the replay protection "window" before
4722 the MAC for the record has been validated. This could be exploited by an
4723 attacker by sending a record for the next epoch (which does not have to
4724 decrypt or have a valid MAC), with a very large sequence number. This means
4725 that all subsequent legitimate packets are dropped causing a denial of
4726 service for a specific DTLS connection.
4728 This issue was reported to OpenSSL by the OCAP audit team.
4733 * Certificate message OOB reads
4735 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4736 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4737 theoretical DoS risk but this has not been observed in practice on common
4740 The messages affected are client certificate, client certificate request
4741 and server certificate. As a result the attack can only be performed
4742 against a client or a server which enables client authentication.
4744 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4749 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4751 * Prevent padding oracle in AES-NI CBC MAC check
4753 A MITM attacker can use a padding oracle attack to decrypt traffic
4754 when the connection uses an AES CBC cipher and the server support
4757 This issue was introduced as part of the fix for Lucky 13 padding
4758 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4759 constant time by making sure that always the same bytes are read and
4760 compared against either the MAC or padding bytes. But it no longer
4761 checked that there was enough data to have both the MAC and padding
4764 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4769 * Fix EVP_EncodeUpdate overflow
4771 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4772 Base64 encoding of binary data. If an attacker is able to supply very large
4773 amounts of input data then a length check can overflow resulting in a heap
4776 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4777 the `PEM_write_bio*` family of functions. These are mainly used within the
4778 OpenSSL command line applications, so any application which processes data
4779 from an untrusted source and outputs it as a PEM file should be considered
4780 vulnerable to this issue. User applications that call these APIs directly
4781 with large amounts of untrusted data may also be vulnerable.
4783 This issue was reported by Guido Vranken.
4788 * Fix EVP_EncryptUpdate overflow
4790 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4791 is able to supply very large amounts of input data after a previous call to
4792 EVP_EncryptUpdate() with a partial block then a length check can overflow
4793 resulting in a heap corruption. Following an analysis of all OpenSSL
4794 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4795 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4796 the first called function after an EVP_EncryptInit(), and therefore that
4797 specific call must be safe. The second form is where the length passed to
4798 EVP_EncryptUpdate() can be seen from the code to be some small value and
4799 therefore there is no possibility of an overflow. Since all instances are
4800 one of these two forms, it is believed that there can be no overflows in
4801 internal code due to this problem. It should be noted that
4802 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4803 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4804 of these calls have also been analysed too and it is believed there are no
4805 instances in internal usage where an overflow could occur.
4807 This issue was reported by Guido Vranken.
4812 * Prevent ASN.1 BIO excessive memory allocation
4814 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4815 a short invalid encoding can cause allocation of large amounts of memory
4816 potentially consuming excessive resources or exhausting memory.
4818 Any application parsing untrusted data through d2i BIO functions is
4819 affected. The memory based functions such as d2i_X509() are *not* affected.
4820 Since the memory based functions are used by the TLS library, TLS
4821 applications are not affected.
4823 This issue was reported by Brian Carpenter.
4830 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4831 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4832 in arbitrary stack data being returned in the buffer.
4834 This issue was reported by Guido Vranken.
4839 * Modify behavior of ALPN to invoke callback after SNI/servername
4840 callback, such that updates to the SSL_CTX affect ALPN.
4844 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4849 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4850 methods are enabled and ssl2 is disabled the methods return NULL.
4854 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4856 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4857 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4858 provide any "EXPORT" or "LOW" strength ciphers.
4862 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4863 is by default disabled at build-time. Builds that are not configured with
4864 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4865 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4866 will need to explicitly call either of:
4868 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4870 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4872 as appropriate. Even if either of those is used, or the application
4873 explicitly uses the version-specific SSLv2_method() or its client and
4874 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4875 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4876 ciphers, and SSLv2 56-bit DES are no longer available.
4881 * Fix a double-free in DSA code
4883 A double free bug was discovered when OpenSSL parses malformed DSA private
4884 keys and could lead to a DoS attack or memory corruption for applications
4885 that receive DSA private keys from untrusted sources. This scenario is
4888 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4894 * Disable SRP fake user seed to address a server memory leak.
4896 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4898 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4899 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4900 was changed to ignore the "fake user" SRP seed, even if the seed
4903 Users should use SRP_VBASE_get1_by_user instead. Note that in
4904 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4905 also that even though configuring the SRP seed attempts to hide
4906 invalid usernames by continuing the handshake with fake
4907 credentials, this behaviour is not constant time and no strong
4908 guarantees are made that the handshake is indistinguishable from
4909 that of a valid user.
4914 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4916 In the BN_hex2bn function the number of hex digits is calculated using an
4917 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4918 large values of `i` this can result in `bn_expand` not allocating any
4919 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4920 field as NULL leading to a subsequent NULL ptr deref. For very large values
4921 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4922 In this case memory is allocated to the internal BIGNUM data field, but it
4923 is insufficiently sized leading to heap corruption. A similar issue exists
4924 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4925 is ever called by user applications with very large untrusted hex/dec data.
4926 This is anticipated to be a rare occurrence.
4928 All OpenSSL internal usage of these functions use data that is not expected
4929 to be untrusted, e.g. config file data or application command line
4930 arguments. If user developed applications generate config file data based
4931 on untrusted data then it is possible that this could also lead to security
4932 consequences. This is also anticipated to be rare.
4934 This issue was reported to OpenSSL by Guido Vranken.
4939 * Fix memory issues in `BIO_*printf` functions
4941 The internal `fmtstr` function used in processing a "%s" format string in
4942 the `BIO_*printf` functions could overflow while calculating the length of a
4943 string and cause an OOB read when printing very long strings.
4945 Additionally the internal `doapr_outch` function can attempt to write to an
4946 OOB memory location (at an offset from the NULL pointer) in the event of a
4947 memory allocation failure. In 1.0.2 and below this could be caused where
4948 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4949 could be in processing a very long "%s" format string. Memory leaks can
4952 The first issue may mask the second issue dependent on compiler behaviour.
4953 These problems could enable attacks where large amounts of untrusted data
4954 is passed to the `BIO_*printf` functions. If applications use these functions
4955 in this way then they could be vulnerable. OpenSSL itself uses these
4956 functions when printing out human-readable dumps of ASN.1 data. Therefore
4957 applications that print this data could be vulnerable if the data is from
4958 untrusted sources. OpenSSL command line applications could also be
4959 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4960 as command line arguments.
4962 Libssl is not considered directly vulnerable. Additionally certificates etc
4963 received via remote connections via libssl are also unlikely to be able to
4964 trigger these issues because of message size limits enforced within libssl.
4966 This issue was reported to OpenSSL Guido Vranken.
4971 * Side channel attack on modular exponentiation
4973 A side-channel attack was found which makes use of cache-bank conflicts on
4974 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4975 of RSA keys. The ability to exploit this issue is limited as it relies on
4976 an attacker who has control of code in a thread running on the same
4977 hyper-threaded core as the victim thread which is performing decryptions.
4979 This issue was reported to OpenSSL by Yuval Yarom, The University of
4980 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4981 Nadia Heninger, University of Pennsylvania with more information at
4982 <http://cachebleed.info>.
4987 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4988 if no keysize is specified with default_bits. This fixes an
4989 omission in an earlier change that changed all RSA/DSA key generation
4990 commands to use 2048 bits by default.
4994 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4996 * DH small subgroups
4998 Historically OpenSSL only ever generated DH parameters based on "safe"
4999 primes. More recently (in version 1.0.2) support was provided for
5000 generating X9.42 style parameter files such as those required for RFC 5114
5001 support. The primes used in such files may not be "safe". Where an
5002 application is using DH configured with parameters based on primes that are
5003 not "safe" then an attacker could use this fact to find a peer's private
5004 DH exponent. This attack requires that the attacker complete multiple
5005 handshakes in which the peer uses the same private DH exponent. For example
5006 this could be used to discover a TLS server's private DH exponent if it's
5007 reusing the private DH exponent or it's using a static DH ciphersuite.
5009 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5010 TLS. It is not on by default. If the option is not set then the server
5011 reuses the same private DH exponent for the life of the server process and
5012 would be vulnerable to this attack. It is believed that many popular
5013 applications do set this option and would therefore not be at risk.
5015 The fix for this issue adds an additional check where a "q" parameter is
5016 available (as is the case in X9.42 based parameters). This detects the
5017 only known attack, and is the only possible defense for static DH
5018 ciphersuites. This could have some performance impact.
5020 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5021 default and cannot be disabled. This could have some performance impact.
5023 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5028 * SSLv2 doesn't block disabled ciphers
5030 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5031 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5032 been disabled, provided that the SSLv2 protocol was not also disabled via
5035 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5036 and Sebastian Schinzel.
5041 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5043 * BN_mod_exp may produce incorrect results on x86_64
5045 There is a carry propagating bug in the x86_64 Montgomery squaring
5046 procedure. No EC algorithms are affected. Analysis suggests that attacks
5047 against RSA and DSA as a result of this defect would be very difficult to
5048 perform and are not believed likely. Attacks against DH are considered just
5049 feasible (although very difficult) because most of the work necessary to
5050 deduce information about a private key may be performed offline. The amount
5051 of resources required for such an attack would be very significant and
5052 likely only accessible to a limited number of attackers. An attacker would
5053 additionally need online access to an unpatched system using the target
5054 private key in a scenario with persistent DH parameters and a private
5055 key that is shared between multiple clients. For example this can occur by
5056 default in OpenSSL DHE based SSL/TLS ciphersuites.
5058 This issue was reported to OpenSSL by Hanno Böck.
5063 * Certificate verify crash with missing PSS parameter
5065 The signature verification routines will crash with a NULL pointer
5066 dereference if presented with an ASN.1 signature using the RSA PSS
5067 algorithm and absent mask generation function parameter. Since these
5068 routines are used to verify certificate signature algorithms this can be
5069 used to crash any certificate verification operation and exploited in a
5070 DoS attack. Any application which performs certificate verification is
5071 vulnerable including OpenSSL clients and servers which enable client
5074 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5079 * X509_ATTRIBUTE memory leak
5081 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5082 memory. This structure is used by the PKCS#7 and CMS routines so any
5083 application which reads PKCS#7 or CMS data from untrusted sources is
5084 affected. SSL/TLS is not affected.
5086 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5092 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5093 This changes the decoding behaviour for some invalid messages,
5094 though the change is mostly in the more lenient direction, and
5095 legacy behaviour is preserved as much as possible.
5099 * In DSA_generate_parameters_ex, if the provided seed is too short,
5102 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5104 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5106 * Alternate chains certificate forgery
5108 During certificate verification, OpenSSL will attempt to find an
5109 alternative certificate chain if the first attempt to build such a chain
5110 fails. An error in the implementation of this logic can mean that an
5111 attacker could cause certain checks on untrusted certificates to be
5112 bypassed, such as the CA flag, enabling them to use a valid leaf
5113 certificate to act as a CA and "issue" an invalid certificate.
5115 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5120 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5122 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5123 incompatibility in the handling of HMAC. The previous ABI has now been
5128 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5130 * Malformed ECParameters causes infinite loop
5132 When processing an ECParameters structure OpenSSL enters an infinite loop
5133 if the curve specified is over a specially malformed binary polynomial
5136 This can be used to perform denial of service against any
5137 system which processes public keys, certificate requests or
5138 certificates. This includes TLS clients and TLS servers with
5139 client authentication enabled.
5141 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5146 * Exploitable out-of-bounds read in X509_cmp_time
5148 X509_cmp_time does not properly check the length of the ASN1_TIME
5149 string and can read a few bytes out of bounds. In addition,
5150 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5153 An attacker can use this to craft malformed certificates and CRLs of
5154 various sizes and potentially cause a segmentation fault, resulting in
5155 a DoS on applications that verify certificates or CRLs. TLS clients
5156 that verify CRLs are affected. TLS clients and servers with client
5157 authentication enabled may be affected if they use custom verification
5160 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5161 independently by Hanno Böck.
5166 * PKCS7 crash with missing EnvelopedContent
5168 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5169 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5170 with missing content and trigger a NULL pointer dereference on parsing.
5172 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5173 structures from untrusted sources are affected. OpenSSL clients and
5174 servers are not affected.
5176 This issue was reported to OpenSSL by Michal Zalewski (Google).
5181 * CMS verify infinite loop with unknown hash function
5183 When verifying a signedData message the CMS code can enter an infinite loop
5184 if presented with an unknown hash function OID. This can be used to perform
5185 denial of service against any system which verifies signedData messages using
5187 This issue was reported to OpenSSL by Johannes Bauer.
5192 * Race condition handling NewSessionTicket
5194 If a NewSessionTicket is received by a multi-threaded client when attempting to
5195 reuse a previous ticket then a race condition can occur potentially leading to
5196 a double free of the ticket data.
5201 * Only support 256-bit or stronger elliptic curves with the
5202 'ecdh_auto' setting (server) or by default (client). Of supported
5203 curves, prefer P-256 (both).
5207 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5209 * ClientHello sigalgs DoS fix
5211 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5212 invalid signature algorithms extension a NULL pointer dereference will
5213 occur. This can be exploited in a DoS attack against the server.
5215 This issue was was reported to OpenSSL by David Ramos of Stanford
5219 *Stephen Henson and Matt Caswell*
5221 * Multiblock corrupted pointer fix
5223 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5224 feature only applies on 64 bit x86 architecture platforms that support AES
5225 NI instructions. A defect in the implementation of "multiblock" can cause
5226 OpenSSL's internal write buffer to become incorrectly set to NULL when
5227 using non-blocking IO. Typically, when the user application is using a
5228 socket BIO for writing, this will only result in a failed connection.
5229 However if some other BIO is used then it is likely that a segmentation
5230 fault will be triggered, thus enabling a potential DoS attack.
5232 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5237 * Segmentation fault in DTLSv1_listen fix
5239 The DTLSv1_listen function is intended to be stateless and processes the
5240 initial ClientHello from many peers. It is common for user code to loop
5241 over the call to DTLSv1_listen until a valid ClientHello is received with
5242 an associated cookie. A defect in the implementation of DTLSv1_listen means
5243 that state is preserved in the SSL object from one invocation to the next
5244 that can lead to a segmentation fault. Errors processing the initial
5245 ClientHello can trigger this scenario. An example of such an error could be
5246 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5249 This issue was reported to OpenSSL by Per Allansson.
5254 * Segmentation fault in ASN1_TYPE_cmp fix
5256 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5257 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5258 certificate signature algorithm consistency this can be used to crash any
5259 certificate verification operation and exploited in a DoS attack. Any
5260 application which performs certificate verification is vulnerable including
5261 OpenSSL clients and servers which enable client authentication.
5266 * Segmentation fault for invalid PSS parameters fix
5268 The signature verification routines will crash with a NULL pointer
5269 dereference if presented with an ASN.1 signature using the RSA PSS
5270 algorithm and invalid parameters. Since these routines are used to verify
5271 certificate signature algorithms this can be used to crash any
5272 certificate verification operation and exploited in a DoS attack. Any
5273 application which performs certificate verification is vulnerable including
5274 OpenSSL clients and servers which enable client authentication.
5276 This issue was was reported to OpenSSL by Brian Carpenter.
5281 * ASN.1 structure reuse memory corruption fix
5283 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5284 memory corruption via an invalid write. Such reuse is and has been
5285 strongly discouraged and is believed to be rare.
5287 Applications that parse structures containing CHOICE or ANY DEFINED BY
5288 components may be affected. Certificate parsing (d2i_X509 and related
5289 functions) are however not affected. OpenSSL clients and servers are
5295 * PKCS7 NULL pointer dereferences fix
5297 The PKCS#7 parsing code does not handle missing outer ContentInfo
5298 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5299 missing content and trigger a NULL pointer dereference on parsing.
5301 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5302 otherwise parse PKCS#7 structures from untrusted sources are
5303 affected. OpenSSL clients and servers are not affected.
5305 This issue was reported to OpenSSL by Michal Zalewski (Google).
5310 * DoS via reachable assert in SSLv2 servers fix
5312 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5313 servers that both support SSLv2 and enable export cipher suites by sending
5314 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5316 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5317 (OpenSSL development team).
5322 * Empty CKE with client auth and DHE fix
5324 If client auth is used then a server can seg fault in the event of a DHE
5325 ciphersuite being selected and a zero length ClientKeyExchange message
5326 being sent by the client. This could be exploited in a DoS attack.
5331 * Handshake with unseeded PRNG fix
5333 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5334 with an unseeded PRNG. The conditions are:
5335 - The client is on a platform where the PRNG has not been seeded
5336 automatically, and the user has not seeded manually
5337 - A protocol specific client method version has been used (i.e. not
5338 SSL_client_methodv23)
5339 - A ciphersuite is used that does not require additional random data from
5340 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5342 If the handshake succeeds then the client random that has been used will
5343 have been generated from a PRNG with insufficient entropy and therefore the
5344 output may be predictable.
5346 For example using the following command with an unseeded openssl will
5347 succeed on an unpatched platform:
5349 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5354 * Use After Free following d2i_ECPrivatekey error fix
5356 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5357 could cause a use after free condition. This, in turn, could cause a double
5358 free in several private key parsing functions (such as d2i_PrivateKey
5359 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5360 for applications that receive EC private keys from untrusted
5361 sources. This scenario is considered rare.
5363 This issue was discovered by the BoringSSL project and fixed in their
5369 * X509_to_X509_REQ NULL pointer deref fix
5371 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5372 the certificate key is invalid. This function is rarely used in practice.
5374 This issue was discovered by Brian Carpenter.
5379 * Removed the export ciphers from the DEFAULT ciphers
5383 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5385 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5386 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5387 So far those who have to target multiple platforms would compromise
5388 and argue that binary targeting say ARMv5 would still execute on
5389 ARMv8. "Universal" build resolves this compromise by providing
5390 near-optimal performance even on newer platforms.
5394 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5395 (other platforms pending).
5397 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5399 * Add support for the SignedCertificateTimestampList certificate and
5400 OCSP response extensions from RFC6962.
5404 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5405 for corner cases. (Certain input points at infinity could lead to
5406 bogus results, with non-infinity inputs mapped to infinity too.)
5410 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5411 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5412 common cases are optimized and there still is room for further
5413 improvements. Vector Permutation AES for Altivec is also added.
5417 * Add support for little-endian ppc64 Linux target.
5419 *Marcelo Cerri (IBM)*
5421 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5422 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5423 are optimized and there still is room for further improvements.
5424 Both 32- and 64-bit modes are supported.
5426 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5428 * Improved ARMv7 NEON support.
5432 * Support for SPARC Architecture 2011 crypto extensions, first
5433 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5434 SHA256/512, MD5, GHASH and modular exponentiation.
5436 *Andy Polyakov, David Miller*
5438 * Accelerated modular exponentiation for Intel processors, a.k.a.
5441 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5443 * Support for new and upcoming Intel processors, including AVX2,
5444 BMI and SHA ISA extensions. This includes additional "stitched"
5445 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5448 This work was sponsored by Intel Corp.
5452 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5453 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5454 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5458 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5459 this fixes a limitation in previous versions of OpenSSL.
5463 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5464 MGF1 digest and OAEP label.
5468 * Add EVP support for key wrapping algorithms, to avoid problems with
5469 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5470 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5471 algorithms and include tests cases.
5475 * Add functions to allocate and set the fields of an ECDSA_METHOD
5478 *Douglas E. Engert, Steve Henson*
5480 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5481 difference in days and seconds between two tm or ASN1_TIME structures.
5485 * Add -rev test option to s_server to just reverse order of characters
5486 received by client and send back to server. Also prints an abbreviated
5487 summary of the connection parameters.
5491 * New option -brief for s_client and s_server to print out a brief summary
5492 of connection parameters.
5496 * Add callbacks for arbitrary TLS extensions.
5498 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5500 * New option -crl_download in several openssl utilities to download CRLs
5501 from CRLDP extension in certificates.
5505 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5509 * New function X509_CRL_diff to generate a delta CRL from the difference
5510 of two full CRLs. Add support to "crl" utility.
5514 * New functions to set lookup_crls function and to retrieve
5515 X509_STORE from X509_STORE_CTX.
5519 * Print out deprecated issuer and subject unique ID fields in
5524 * Extend OCSP I/O functions so they can be used for simple general purpose
5525 HTTP as well as OCSP. New wrapper function which can be used to download
5526 CRLs using the OCSP API.
5530 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5534 * `SSL_CONF*` functions. These provide a common framework for application
5535 configuration using configuration files or command lines.
5539 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5540 message callback and prints the results. Needs compile time option
5541 "enable-ssl-trace". New options to s_client and s_server to enable
5546 * New ctrl and macro to retrieve supported points extensions.
5547 Print out extension in s_server and s_client.
5551 * New functions to retrieve certificate signature and signature
5556 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5561 * New Suite B modes for TLS code. These use and enforce the requirements
5562 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5563 only use Suite B curves. The Suite B modes can be set by using the
5564 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5568 * New chain verification flags for Suite B levels of security. Check
5569 algorithms are acceptable when flags are set in X509_verify_cert.
5573 * Make tls1_check_chain return a set of flags indicating checks passed
5574 by a certificate chain. Add additional tests to handle client
5575 certificates: checks for matching certificate type and issuer name
5580 * If an attempt is made to use a signature algorithm not in the peer
5581 preference list abort the handshake. If client has no suitable
5582 signature algorithms in response to a certificate request do not
5583 use the certificate.
5587 * If server EC tmp key is not in client preference list abort handshake.
5591 * Add support for certificate stores in CERT structure. This makes it
5592 possible to have different stores per SSL structure or one store in
5593 the parent SSL_CTX. Include distinct stores for certificate chain
5594 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5595 to build and store a certificate chain in CERT structure: returning
5596 an error if the chain cannot be built: this will allow applications
5597 to test if a chain is correctly configured.
5599 Note: if the CERT based stores are not set then the parent SSL_CTX
5600 store is used to retain compatibility with existing behaviour.
5604 * New function ssl_set_client_disabled to set a ciphersuite disabled
5605 mask based on the current session, check mask when sending client
5606 hello and checking the requested ciphersuite.
5610 * New ctrls to retrieve and set certificate types in a certificate
5611 request message. Print out received values in s_client. If certificate
5612 types is not set with custom values set sensible values based on
5613 supported signature algorithms.
5617 * Support for distinct client and server supported signature algorithms.
5621 * Add certificate callback. If set this is called whenever a certificate
5622 is required by client or server. An application can decide which
5623 certificate chain to present based on arbitrary criteria: for example
5624 supported signature algorithms. Add very simple example to s_server.
5625 This fixes many of the problems and restrictions of the existing client
5626 certificate callback: for example you can now clear an existing
5627 certificate and specify the whole chain.
5631 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5632 the certificate can be used for (if anything). Set valid_flags field
5633 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5634 to have similar checks in it.
5636 Add new "cert_flags" field to CERT structure and include a "strict mode".
5637 This enforces some TLS certificate requirements (such as only permitting
5638 certificate signature algorithms contained in the supported algorithms
5639 extension) which some implementations ignore: this option should be used
5640 with caution as it could cause interoperability issues.
5644 * Update and tidy signature algorithm extension processing. Work out
5645 shared signature algorithms based on preferences and peer algorithms
5646 and print them out in s_client and s_server. Abort handshake if no
5647 shared signature algorithms.
5651 * Add new functions to allow customised supported signature algorithms
5652 for SSL and SSL_CTX structures. Add options to s_client and s_server
5657 * New function SSL_certs_clear() to delete all references to certificates
5658 from an SSL structure. Before this once a certificate had been added
5659 it couldn't be removed.
5663 * Integrate hostname, email address and IP address checking with certificate
5664 verification. New verify options supporting checking in openssl utility.
5668 * Fixes and wildcard matching support to hostname and email checking
5669 functions. Add manual page.
5671 *Florian Weimer (Red Hat Product Security Team)*
5673 * New functions to check a hostname email or IP address against a
5674 certificate. Add options x509 utility to print results of checks against
5679 * Fix OCSP checking.
5681 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5683 * Initial experimental support for explicitly trusted non-root CAs.
5684 OpenSSL still tries to build a complete chain to a root but if an
5685 intermediate CA has a trust setting included that is used. The first
5686 setting is used: whether to trust (e.g., -addtrust option to the x509
5691 * Add -trusted_first option which attempts to find certificates in the
5692 trusted store even if an untrusted chain is also supplied.
5696 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5697 platform support for Linux and Android.
5701 * Support for linux-x32, ILP32 environment in x86_64 framework.
5705 * Experimental multi-implementation support for FIPS capable OpenSSL.
5706 When in FIPS mode the approved implementations are used as normal,
5707 when not in FIPS mode the internal unapproved versions are used instead.
5708 This means that the FIPS capable OpenSSL isn't forced to use the
5709 (often lower performance) FIPS implementations outside FIPS mode.
5713 * Transparently support X9.42 DH parameters when calling
5714 PEM_read_bio_DHparameters. This means existing applications can handle
5715 the new parameter format automatically.
5719 * Initial experimental support for X9.42 DH parameter format: mainly
5720 to support use of 'q' parameter for RFC5114 parameters.
5724 * Add DH parameters from RFC5114 including test data to dhtest.
5728 * Support for automatic EC temporary key parameter selection. If enabled
5729 the most preferred EC parameters are automatically used instead of
5730 hardcoded fixed parameters. Now a server just has to call:
5731 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5732 support ECDH and use the most appropriate parameters.
5736 * Enhance and tidy EC curve and point format TLS extension code. Use
5737 static structures instead of allocation if default values are used.
5738 New ctrls to set curves we wish to support and to retrieve shared curves.
5739 Print out shared curves in s_server. New options to s_server and s_client
5740 to set list of supported curves.
5744 * New ctrls to retrieve supported signature algorithms and
5745 supported curve values as an array of NIDs. Extend openssl utility
5746 to print out received values.
5750 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5751 between NIDs and the more common NIST names such as "P-256". Enhance
5752 ecparam utility and ECC method to recognise the NIST names for curves.
5756 * Enhance SSL/TLS certificate chain handling to support different
5757 chains for each certificate instead of one chain in the parent SSL_CTX.
5761 * Support for fixed DH ciphersuite client authentication: where both
5762 server and client use DH certificates with common parameters.
5766 * Support for fixed DH ciphersuites: those requiring DH server
5771 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5773 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5774 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5775 X509_CINF_get_signature were reverted post internal team review.
5780 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5782 * OCSP Status Request extension unbounded memory growth
5784 A malicious client can send an excessively large OCSP Status Request
5785 extension. If that client continually requests renegotiation, sending a
5786 large OCSP Status Request extension each time, then there will be unbounded
5787 memory growth on the server. This will eventually lead to a Denial Of
5788 Service attack through memory exhaustion. Servers with a default
5789 configuration are vulnerable even if they do not support OCSP. Builds using
5790 the "no-ocsp" build time option are not affected.
5792 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5797 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5800 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5806 * OOB write in MDC2_Update()
5808 An overflow can occur in MDC2_Update() either if called directly or
5809 through the EVP_DigestUpdate() function using MDC2. If an attacker
5810 is able to supply very large amounts of input data after a previous
5811 call to EVP_EncryptUpdate() with a partial block then a length check
5812 can overflow resulting in a heap corruption.
5814 The amount of data needed is comparable to SIZE_MAX which is impractical
5817 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5822 * Malformed SHA512 ticket DoS
5824 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5825 DoS attack where a malformed ticket will result in an OOB read which will
5828 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5829 a custom server callback and ticket lookup mechanism.
5831 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5836 * OOB write in BN_bn2dec()
5838 The function BN_bn2dec() does not check the return value of BN_div_word().
5839 This can cause an OOB write if an application uses this function with an
5840 overly large BIGNUM. This could be a problem if an overly large certificate
5841 or CRL is printed out from an untrusted source. TLS is not affected because
5842 record limits will reject an oversized certificate before it is parsed.
5844 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5849 * OOB read in TS_OBJ_print_bio()
5851 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5852 the total length the OID text representation would use and not the amount
5853 of data written. This will result in OOB reads when large OIDs are
5856 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5861 * Pointer arithmetic undefined behaviour
5863 Avoid some undefined pointer arithmetic
5865 A common idiom in the codebase is to check limits in the following manner:
5868 Where "p" points to some malloc'd data of SIZE bytes and
5871 "len" here could be from some externally supplied data (e.g. from a TLS
5874 The rules of C pointer arithmetic are such that "p + len" is only well
5875 defined where len <= SIZE. Therefore the above idiom is actually
5876 undefined behaviour.
5878 For example this could cause problems if some malloc implementation
5879 provides an address for "p" such that "p + len" actually overflows for
5880 values of len that are too big and therefore p + len < limit.
5882 This issue was reported to OpenSSL by Guido Vranken
5887 * Constant time flag not preserved in DSA signing
5889 Operations in the DSA signing algorithm should run in constant time in
5890 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5891 implementation means that a non-constant time codepath is followed for
5892 certain operations. This has been demonstrated through a cache-timing
5893 attack to be sufficient for an attacker to recover the private DSA key.
5895 This issue was reported by César Pereida (Aalto University), Billy Brumley
5896 (Tampere University of Technology), and Yuval Yarom (The University of
5897 Adelaide and NICTA).
5902 * DTLS buffered message DoS
5904 In a DTLS connection where handshake messages are delivered out-of-order
5905 those messages that OpenSSL is not yet ready to process will be buffered
5906 for later use. Under certain circumstances, a flaw in the logic means that
5907 those messages do not get removed from the buffer even though the handshake
5908 has been completed. An attacker could force up to approx. 15 messages to
5909 remain in the buffer when they are no longer required. These messages will
5910 be cleared when the DTLS connection is closed. The default maximum size for
5911 a message is 100k. Therefore the attacker could force an additional 1500k
5912 to be consumed per connection. By opening many simulataneous connections an
5913 attacker could cause a DoS attack through memory exhaustion.
5915 This issue was reported to OpenSSL by Quan Luo.
5920 * DTLS replay protection DoS
5922 A flaw in the DTLS replay attack protection mechanism means that records
5923 that arrive for future epochs update the replay protection "window" before
5924 the MAC for the record has been validated. This could be exploited by an
5925 attacker by sending a record for the next epoch (which does not have to
5926 decrypt or have a valid MAC), with a very large sequence number. This means
5927 that all subsequent legitimate packets are dropped causing a denial of
5928 service for a specific DTLS connection.
5930 This issue was reported to OpenSSL by the OCAP audit team.
5935 * Certificate message OOB reads
5937 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5938 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5939 theoretical DoS risk but this has not been observed in practice on common
5942 The messages affected are client certificate, client certificate request
5943 and server certificate. As a result the attack can only be performed
5944 against a client or a server which enables client authentication.
5946 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5951 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5953 * Prevent padding oracle in AES-NI CBC MAC check
5955 A MITM attacker can use a padding oracle attack to decrypt traffic
5956 when the connection uses an AES CBC cipher and the server support
5959 This issue was introduced as part of the fix for Lucky 13 padding
5960 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5961 constant time by making sure that always the same bytes are read and
5962 compared against either the MAC or padding bytes. But it no longer
5963 checked that there was enough data to have both the MAC and padding
5966 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5971 * Fix EVP_EncodeUpdate overflow
5973 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5974 Base64 encoding of binary data. If an attacker is able to supply very large
5975 amounts of input data then a length check can overflow resulting in a heap
5978 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5979 the `PEM_write_bio*` family of functions. These are mainly used within the
5980 OpenSSL command line applications, so any application which processes data
5981 from an untrusted source and outputs it as a PEM file should be considered
5982 vulnerable to this issue. User applications that call these APIs directly
5983 with large amounts of untrusted data may also be vulnerable.
5985 This issue was reported by Guido Vranken.
5990 * Fix EVP_EncryptUpdate overflow
5992 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5993 is able to supply very large amounts of input data after a previous call to
5994 EVP_EncryptUpdate() with a partial block then a length check can overflow
5995 resulting in a heap corruption. Following an analysis of all OpenSSL
5996 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5997 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5998 the first called function after an EVP_EncryptInit(), and therefore that
5999 specific call must be safe. The second form is where the length passed to
6000 EVP_EncryptUpdate() can be seen from the code to be some small value and
6001 therefore there is no possibility of an overflow. Since all instances are
6002 one of these two forms, it is believed that there can be no overflows in
6003 internal code due to this problem. It should be noted that
6004 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6005 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6006 of these calls have also been analysed too and it is believed there are no
6007 instances in internal usage where an overflow could occur.
6009 This issue was reported by Guido Vranken.
6014 * Prevent ASN.1 BIO excessive memory allocation
6016 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6017 a short invalid encoding can casuse allocation of large amounts of memory
6018 potentially consuming excessive resources or exhausting memory.
6020 Any application parsing untrusted data through d2i BIO functions is
6021 affected. The memory based functions such as d2i_X509() are *not* affected.
6022 Since the memory based functions are used by the TLS library, TLS
6023 applications are not affected.
6025 This issue was reported by Brian Carpenter.
6032 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6033 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6034 in arbitrary stack data being returned in the buffer.
6036 This issue was reported by Guido Vranken.
6041 * Modify behavior of ALPN to invoke callback after SNI/servername
6042 callback, such that updates to the SSL_CTX affect ALPN.
6046 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6051 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6052 methods are enabled and ssl2 is disabled the methods return NULL.
6056 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6058 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6059 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6060 provide any "EXPORT" or "LOW" strength ciphers.
6064 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6065 is by default disabled at build-time. Builds that are not configured with
6066 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6067 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6068 will need to explicitly call either of:
6070 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6072 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6074 as appropriate. Even if either of those is used, or the application
6075 explicitly uses the version-specific SSLv2_method() or its client and
6076 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6077 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6078 ciphers, and SSLv2 56-bit DES are no longer available.
6083 * Fix a double-free in DSA code
6085 A double free bug was discovered when OpenSSL parses malformed DSA private
6086 keys and could lead to a DoS attack or memory corruption for applications
6087 that receive DSA private keys from untrusted sources. This scenario is
6090 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6096 * Disable SRP fake user seed to address a server memory leak.
6098 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6100 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6101 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6102 was changed to ignore the "fake user" SRP seed, even if the seed
6105 Users should use SRP_VBASE_get1_by_user instead. Note that in
6106 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6107 also that even though configuring the SRP seed attempts to hide
6108 invalid usernames by continuing the handshake with fake
6109 credentials, this behaviour is not constant time and no strong
6110 guarantees are made that the handshake is indistinguishable from
6111 that of a valid user.
6116 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6118 In the BN_hex2bn function the number of hex digits is calculated using an
6119 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6120 large values of `i` this can result in `bn_expand` not allocating any
6121 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6122 field as NULL leading to a subsequent NULL ptr deref. For very large values
6123 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6124 In this case memory is allocated to the internal BIGNUM data field, but it
6125 is insufficiently sized leading to heap corruption. A similar issue exists
6126 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6127 is ever called by user applications with very large untrusted hex/dec data.
6128 This is anticipated to be a rare occurrence.
6130 All OpenSSL internal usage of these functions use data that is not expected
6131 to be untrusted, e.g. config file data or application command line
6132 arguments. If user developed applications generate config file data based
6133 on untrusted data then it is possible that this could also lead to security
6134 consequences. This is also anticipated to be rare.
6136 This issue was reported to OpenSSL by Guido Vranken.
6141 * Fix memory issues in `BIO_*printf` functions
6143 The internal `fmtstr` function used in processing a "%s" format string in
6144 the `BIO_*printf` functions could overflow while calculating the length of a
6145 string and cause an OOB read when printing very long strings.
6147 Additionally the internal `doapr_outch` function can attempt to write to an
6148 OOB memory location (at an offset from the NULL pointer) in the event of a
6149 memory allocation failure. In 1.0.2 and below this could be caused where
6150 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6151 could be in processing a very long "%s" format string. Memory leaks can
6154 The first issue may mask the second issue dependent on compiler behaviour.
6155 These problems could enable attacks where large amounts of untrusted data
6156 is passed to the `BIO_*printf` functions. If applications use these functions
6157 in this way then they could be vulnerable. OpenSSL itself uses these
6158 functions when printing out human-readable dumps of ASN.1 data. Therefore
6159 applications that print this data could be vulnerable if the data is from
6160 untrusted sources. OpenSSL command line applications could also be
6161 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6162 as command line arguments.
6164 Libssl is not considered directly vulnerable. Additionally certificates etc
6165 received via remote connections via libssl are also unlikely to be able to
6166 trigger these issues because of message size limits enforced within libssl.
6168 This issue was reported to OpenSSL Guido Vranken.
6173 * Side channel attack on modular exponentiation
6175 A side-channel attack was found which makes use of cache-bank conflicts on
6176 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6177 of RSA keys. The ability to exploit this issue is limited as it relies on
6178 an attacker who has control of code in a thread running on the same
6179 hyper-threaded core as the victim thread which is performing decryptions.
6181 This issue was reported to OpenSSL by Yuval Yarom, The University of
6182 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6183 Nadia Heninger, University of Pennsylvania with more information at
6184 <http://cachebleed.info>.
6189 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6190 if no keysize is specified with default_bits. This fixes an
6191 omission in an earlier change that changed all RSA/DSA key generation
6192 commands to use 2048 bits by default.
6196 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6198 * Protection for DH small subgroup attacks
6200 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6201 switched on by default and cannot be disabled. This could have some
6206 * SSLv2 doesn't block disabled ciphers
6208 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6209 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6210 been disabled, provided that the SSLv2 protocol was not also disabled via
6213 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6214 and Sebastian Schinzel.
6219 * Reject DH handshakes with parameters shorter than 1024 bits.
6223 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6225 * Certificate verify crash with missing PSS parameter
6227 The signature verification routines will crash with a NULL pointer
6228 dereference if presented with an ASN.1 signature using the RSA PSS
6229 algorithm and absent mask generation function parameter. Since these
6230 routines are used to verify certificate signature algorithms this can be
6231 used to crash any certificate verification operation and exploited in a
6232 DoS attack. Any application which performs certificate verification is
6233 vulnerable including OpenSSL clients and servers which enable client
6236 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6241 * X509_ATTRIBUTE memory leak
6243 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6244 memory. This structure is used by the PKCS#7 and CMS routines so any
6245 application which reads PKCS#7 or CMS data from untrusted sources is
6246 affected. SSL/TLS is not affected.
6248 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6254 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6255 This changes the decoding behaviour for some invalid messages,
6256 though the change is mostly in the more lenient direction, and
6257 legacy behaviour is preserved as much as possible.
6261 * In DSA_generate_parameters_ex, if the provided seed is too short,
6262 use a random seed, as already documented.
6264 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6266 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6268 * Alternate chains certificate forgery
6270 During certificate verfification, OpenSSL will attempt to find an
6271 alternative certificate chain if the first attempt to build such a chain
6272 fails. An error in the implementation of this logic can mean that an
6273 attacker could cause certain checks on untrusted certificates to be
6274 bypassed, such as the CA flag, enabling them to use a valid leaf
6275 certificate to act as a CA and "issue" an invalid certificate.
6277 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6283 * Race condition handling PSK identify hint
6285 If PSK identity hints are received by a multi-threaded client then
6286 the values are wrongly updated in the parent SSL_CTX structure. This can
6287 result in a race condition potentially leading to a double free of the
6293 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6295 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6296 incompatibility in the handling of HMAC. The previous ABI has now been
6299 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6301 * Malformed ECParameters causes infinite loop
6303 When processing an ECParameters structure OpenSSL enters an infinite loop
6304 if the curve specified is over a specially malformed binary polynomial
6307 This can be used to perform denial of service against any
6308 system which processes public keys, certificate requests or
6309 certificates. This includes TLS clients and TLS servers with
6310 client authentication enabled.
6312 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6317 * Exploitable out-of-bounds read in X509_cmp_time
6319 X509_cmp_time does not properly check the length of the ASN1_TIME
6320 string and can read a few bytes out of bounds. In addition,
6321 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6324 An attacker can use this to craft malformed certificates and CRLs of
6325 various sizes and potentially cause a segmentation fault, resulting in
6326 a DoS on applications that verify certificates or CRLs. TLS clients
6327 that verify CRLs are affected. TLS clients and servers with client
6328 authentication enabled may be affected if they use custom verification
6331 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6332 independently by Hanno Böck.
6337 * PKCS7 crash with missing EnvelopedContent
6339 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6340 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6341 with missing content and trigger a NULL pointer dereference on parsing.
6343 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6344 structures from untrusted sources are affected. OpenSSL clients and
6345 servers are not affected.
6347 This issue was reported to OpenSSL by Michal Zalewski (Google).
6352 * CMS verify infinite loop with unknown hash function
6354 When verifying a signedData message the CMS code can enter an infinite loop
6355 if presented with an unknown hash function OID. This can be used to perform
6356 denial of service against any system which verifies signedData messages using
6358 This issue was reported to OpenSSL by Johannes Bauer.
6363 * Race condition handling NewSessionTicket
6365 If a NewSessionTicket is received by a multi-threaded client when attempting to
6366 reuse a previous ticket then a race condition can occur potentially leading to
6367 a double free of the ticket data.
6372 * Reject DH handshakes with parameters shorter than 768 bits.
6374 *Kurt Roeckx and Emilia Kasper*
6376 * dhparam: generate 2048-bit parameters by default.
6378 *Kurt Roeckx and Emilia Kasper*
6380 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6382 * Segmentation fault in ASN1_TYPE_cmp fix
6384 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6385 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6386 certificate signature algorithm consistency this can be used to crash any
6387 certificate verification operation and exploited in a DoS attack. Any
6388 application which performs certificate verification is vulnerable including
6389 OpenSSL clients and servers which enable client authentication.
6394 * ASN.1 structure reuse memory corruption fix
6396 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6397 memory corruption via an invalid write. Such reuse is and has been
6398 strongly discouraged and is believed to be rare.
6400 Applications that parse structures containing CHOICE or ANY DEFINED BY
6401 components may be affected. Certificate parsing (d2i_X509 and related
6402 functions) are however not affected. OpenSSL clients and servers are
6408 * PKCS7 NULL pointer dereferences fix
6410 The PKCS#7 parsing code does not handle missing outer ContentInfo
6411 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6412 missing content and trigger a NULL pointer dereference on parsing.
6414 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6415 otherwise parse PKCS#7 structures from untrusted sources are
6416 affected. OpenSSL clients and servers are not affected.
6418 This issue was reported to OpenSSL by Michal Zalewski (Google).
6423 * DoS via reachable assert in SSLv2 servers fix
6425 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6426 servers that both support SSLv2 and enable export cipher suites by sending
6427 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6429 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6430 (OpenSSL development team).
6435 * Use After Free following d2i_ECPrivatekey error fix
6437 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6438 could cause a use after free condition. This, in turn, could cause a double
6439 free in several private key parsing functions (such as d2i_PrivateKey
6440 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6441 for applications that receive EC private keys from untrusted
6442 sources. This scenario is considered rare.
6444 This issue was discovered by the BoringSSL project and fixed in their
6450 * X509_to_X509_REQ NULL pointer deref fix
6452 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6453 the certificate key is invalid. This function is rarely used in practice.
6455 This issue was discovered by Brian Carpenter.
6460 * Removed the export ciphers from the DEFAULT ciphers
6464 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6466 * Build fixes for the Windows and OpenVMS platforms
6468 *Matt Caswell and Richard Levitte*
6470 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6472 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6473 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6474 dereference. This could lead to a Denial Of Service attack. Thanks to
6475 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6480 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6481 dtls1_buffer_record function under certain conditions. In particular this
6482 could occur if an attacker sent repeated DTLS records with the same
6483 sequence number but for the next epoch. The memory leak could be exploited
6484 by an attacker in a Denial of Service attack through memory exhaustion.
6485 Thanks to Chris Mueller for reporting this issue.
6490 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6491 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6492 method would be set to NULL which could later result in a NULL pointer
6493 dereference. Thanks to Frank Schmirler for reporting this issue.
6498 * Abort handshake if server key exchange message is omitted for ephemeral
6501 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6502 reporting this issue.
6507 * Remove non-export ephemeral RSA code on client and server. This code
6508 violated the TLS standard by allowing the use of temporary RSA keys in
6509 non-export ciphersuites and could be used by a server to effectively
6510 downgrade the RSA key length used to a value smaller than the server
6511 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6512 INRIA or reporting this issue.
6517 * Fixed issue where DH client certificates are accepted without verification.
6518 An OpenSSL server will accept a DH certificate for client authentication
6519 without the certificate verify message. This effectively allows a client to
6520 authenticate without the use of a private key. This only affects servers
6521 which trust a client certificate authority which issues certificates
6522 containing DH keys: these are extremely rare and hardly ever encountered.
6523 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6529 * Ensure that the session ID context of an SSL is updated when its
6530 SSL_CTX is updated via SSL_set_SSL_CTX.
6532 The session ID context is typically set from the parent SSL_CTX,
6533 and can vary with the CTX.
6537 * Fix various certificate fingerprint issues.
6539 By using non-DER or invalid encodings outside the signed portion of a
6540 certificate the fingerprint can be changed without breaking the signature.
6541 Although no details of the signed portion of the certificate can be changed
6542 this can cause problems with some applications: e.g. those using the
6543 certificate fingerprint for blacklists.
6545 1. Reject signatures with non zero unused bits.
6547 If the BIT STRING containing the signature has non zero unused bits reject
6548 the signature. All current signature algorithms require zero unused bits.
6550 2. Check certificate algorithm consistency.
6552 Check the AlgorithmIdentifier inside TBS matches the one in the
6553 certificate signature. NB: this will result in signature failure
6554 errors for some broken certificates.
6556 Thanks to Konrad Kraszewski from Google for reporting this issue.
6558 3. Check DSA/ECDSA signatures use DER.
6560 Re-encode DSA/ECDSA signatures and compare with the original received
6561 signature. Return an error if there is a mismatch.
6563 This will reject various cases including garbage after signature
6564 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6565 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6566 (negative or with leading zeroes).
6568 Further analysis was conducted and fixes were developed by Stephen Henson
6569 of the OpenSSL core team.
6575 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6576 results on some platforms, including x86_64. This bug occurs at random
6577 with a very low probability, and is not known to be exploitable in any
6578 way, though its exact impact is difficult to determine. Thanks to Pieter
6579 Wuille (Blockstream) who reported this issue and also suggested an initial
6580 fix. Further analysis was conducted by the OpenSSL development team and
6581 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6582 the OpenSSL core team.
6587 * Do not resume sessions on the server if the negotiated protocol
6588 version does not match the session's version. Resuming with a different
6589 version, while not strictly forbidden by the RFC, is of questionable
6590 sanity and breaks all known clients.
6592 *David Benjamin, Emilia Käsper*
6594 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6595 early CCS messages during renegotiation. (Note that because
6596 renegotiation is encrypted, this early CCS was not exploitable.)
6600 * Tighten client-side session ticket handling during renegotiation:
6601 ensure that the client only accepts a session ticket if the server sends
6602 the extension anew in the ServerHello. Previously, a TLS client would
6603 reuse the old extension state and thus accept a session ticket if one was
6604 announced in the initial ServerHello.
6606 Similarly, ensure that the client requires a session ticket if one
6607 was advertised in the ServerHello. Previously, a TLS client would
6608 ignore a missing NewSessionTicket message.
6612 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6616 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6617 sends a carefully crafted handshake message, to cause OpenSSL to fail
6618 to free up to 64k of memory causing a memory leak. This could be
6619 exploited in a Denial Of Service attack. This issue affects OpenSSL
6620 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6621 whether SRTP is used or configured. Implementations of OpenSSL that
6622 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6624 The fix was developed by the OpenSSL team.
6629 * Session Ticket Memory Leak.
6631 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6632 integrity of that ticket is first verified. In the event of a session
6633 ticket integrity check failing, OpenSSL will fail to free memory
6634 causing a memory leak. By sending a large number of invalid session
6635 tickets an attacker could exploit this issue in a Denial Of Service
6641 * Build option no-ssl3 is incomplete.
6643 When OpenSSL is configured with "no-ssl3" as a build option, servers
6644 could accept and complete a SSL 3.0 handshake, and clients could be
6645 configured to send them.
6648 *Akamai and the OpenSSL team*
6650 * Add support for TLS_FALLBACK_SCSV.
6651 Client applications doing fallback retries should call
6652 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6655 *Adam Langley, Bodo Moeller*
6657 * Add additional DigestInfo checks.
6659 Re-encode DigestInto in DER and check against the original when
6660 verifying RSA signature: this will reject any improperly encoded
6661 DigestInfo structures.
6663 Note: this is a precautionary measure and no attacks are currently known.
6667 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6669 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6670 SRP code can be overrun an internal buffer. Add sanity check that
6671 g, A, B < N to SRP code.
6673 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6674 Group for discovering this issue.
6679 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6680 TLS 1.0 instead of higher protocol versions when the ClientHello message
6681 is badly fragmented. This allows a man-in-the-middle attacker to force a
6682 downgrade to TLS 1.0 even if both the server and the client support a
6683 higher protocol version, by modifying the client's TLS records.
6685 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6686 researching this issue.
6691 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6692 to a denial of service attack. A malicious server can crash the client
6693 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6694 ciphersuite and sending carefully crafted handshake messages.
6696 Thanks to Felix Gröbert (Google) for discovering and researching this
6702 * By sending carefully crafted DTLS packets an attacker could cause openssl
6703 to leak memory. This can be exploited through a Denial of Service attack.
6704 Thanks to Adam Langley for discovering and researching this issue.
6709 * An attacker can force openssl to consume large amounts of memory whilst
6710 processing DTLS handshake messages. This can be exploited through a
6711 Denial of Service attack.
6712 Thanks to Adam Langley for discovering and researching this issue.
6717 * An attacker can force an error condition which causes openssl to crash
6718 whilst processing DTLS packets due to memory being freed twice. This
6719 can be exploited through a Denial of Service attack.
6720 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6726 * If a multithreaded client connects to a malicious server using a resumed
6727 session and the server sends an ec point format extension it could write
6728 up to 255 bytes to freed memory.
6730 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6736 * A malicious server can crash an OpenSSL client with a null pointer
6737 dereference (read) by specifying an SRP ciphersuite even though it was not
6738 properly negotiated with the client. This can be exploited through a
6739 Denial of Service attack.
6741 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6742 discovering and researching this issue.
6747 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6748 X509_name_oneline, X509_name_print_ex et al. to leak some information
6749 from the stack. Applications may be affected if they echo pretty printing
6750 output to the attacker.
6752 Thanks to Ivan Fratric (Google) for discovering this issue.
6755 *Emilia Käsper, and Steve Henson*
6757 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6758 for corner cases. (Certain input points at infinity could lead to
6759 bogus results, with non-infinity inputs mapped to infinity too.)
6763 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6765 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6766 handshake can force the use of weak keying material in OpenSSL
6767 SSL/TLS clients and servers.
6769 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6770 researching this issue. ([CVE-2014-0224])
6772 *KIKUCHI Masashi, Steve Henson*
6774 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6775 OpenSSL DTLS client the code can be made to recurse eventually crashing
6778 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6781 *Imre Rad, Steve Henson*
6783 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6784 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6785 client or server. This is potentially exploitable to run arbitrary
6786 code on a vulnerable client or server.
6788 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6790 *Jüri Aedla, Steve Henson*
6792 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6793 are subject to a denial of service attack.
6795 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6796 this issue. ([CVE-2014-3470])
6798 *Felix Gröbert, Ivan Fratric, Steve Henson*
6800 * Harmonize version and its documentation. -f flag is used to display
6803 *mancha <mancha1@zoho.com>*
6805 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6806 in i2d_ECPrivateKey.
6808 *mancha <mancha1@zoho.com>*
6810 * Fix some double frees. These are not thought to be exploitable.
6812 *mancha <mancha1@zoho.com>*
6814 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6816 * A missing bounds check in the handling of the TLS heartbeat extension
6817 can be used to reveal up to 64k of memory to a connected client or
6820 Thanks for Neel Mehta of Google Security for discovering this bug and to
6821 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6822 preparing the fix ([CVE-2014-0160])
6824 *Adam Langley, Bodo Moeller*
6826 * Fix for the attack described in the paper "Recovering OpenSSL
6827 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6828 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6829 <http://eprint.iacr.org/2014/140>
6831 Thanks to Yuval Yarom and Naomi Benger for discovering this
6832 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6834 *Yuval Yarom and Naomi Benger*
6836 * TLS pad extension: draft-agl-tls-padding-03
6838 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6839 TLS client Hello record length value would otherwise be > 255 and
6840 less that 512 pad with a dummy extension containing zeroes so it
6841 is at least 512 bytes long.
6843 *Adam Langley, Steve Henson*
6845 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6847 * Fix for TLS record tampering bug. A carefully crafted invalid
6848 handshake could crash OpenSSL with a NULL pointer exception.
6849 Thanks to Anton Johansson for reporting this issues.
6852 * Keep original DTLS digest and encryption contexts in retransmission
6853 structures so we can use the previous session parameters if they need
6854 to be resent. ([CVE-2013-6450])
6858 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6859 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6860 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6861 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6862 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6863 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6865 *Rob Stradling, Adam Langley*
6867 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6869 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6870 supporting platforms or when small records were transferred.
6872 *Andy Polyakov, Steve Henson*
6874 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6876 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6878 This addresses the flaw in CBC record processing discovered by
6879 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6880 at: <http://www.isg.rhul.ac.uk/tls/>
6882 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6883 Security Group at Royal Holloway, University of London
6884 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6885 Emilia Käsper for the initial patch.
6888 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6890 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6891 ciphersuites which can be exploited in a denial of service attack.
6892 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6893 and detecting this bug and to Wolfgang Ettlinger
6894 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6899 * Return an error when checking OCSP signatures when key is NULL.
6900 This fixes a DoS attack. ([CVE-2013-0166])
6904 * Make openssl verify return errors.
6906 *Chris Palmer <palmer@google.com> and Ben Laurie*
6908 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6909 the right response is stapled. Also change SSL_get_certificate()
6910 so it returns the certificate actually sent.
6911 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6913 *Rob Stradling <rob.stradling@comodo.com>*
6915 * Fix possible deadlock when decoding public keys.
6919 * Don't use TLS 1.0 record version number in initial client hello
6924 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6926 * Sanity check record length before skipping explicit IV in TLS
6927 1.2, 1.1 and DTLS to fix DoS attack.
6929 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6930 fuzzing as a service testing platform.
6935 * Initialise tkeylen properly when encrypting CMS messages.
6936 Thanks to Solar Designer of Openwall for reporting this issue.
6940 * In FIPS mode don't try to use composite ciphers as they are not
6945 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6947 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6948 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6949 mean any application compiled against OpenSSL 1.0.0 headers setting
6950 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6951 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6952 0x10000000L Any application which was previously compiled against
6953 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6954 will need to be recompiled as a result. Letting be results in
6955 inability to disable specifically TLS 1.1 and in client context,
6956 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6960 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6961 disable just protocol X, but all protocols above X *if* there are
6962 protocols *below* X still enabled. In more practical terms it means
6963 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6964 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6965 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6970 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6972 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6973 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6974 in CRYPTO_realloc_clean.
6976 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6977 issue and to Adam Langley <agl@chromium.org> for fixing it.
6980 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6982 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6986 * Workarounds for some broken servers that "hang" if a client hello
6987 record length exceeds 255 bytes.
6989 1. Do not use record version number > TLS 1.0 in initial client
6990 hello: some (but not all) hanging servers will now work.
6991 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6992 the number of ciphers sent in the client hello. This should be
6993 set to an even number, such as 50, for example by passing:
6994 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6995 Most broken servers should now work.
6996 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6997 TLS 1.2 client support entirely.
7001 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7005 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7007 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7008 STRING form instead of a DigestInfo.
7012 * The format used for MDC2 RSA signatures is inconsistent between EVP
7013 and the RSA_sign/RSA_verify functions. This was made more apparent when
7014 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7015 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7016 the correct format in RSA_verify so both forms transparently work.
7020 * Some servers which support TLS 1.0 can choke if we initially indicate
7021 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7022 encrypted premaster secret. As a workaround use the maximum permitted
7023 client version in client hello, this should keep such servers happy
7024 and still work with previous versions of OpenSSL.
7028 * Add support for TLS/DTLS heartbeats.
7030 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7032 * Add support for SCTP.
7034 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7036 * Improved PRNG seeding for VOS.
7038 *Paul Green <Paul.Green@stratus.com>*
7040 * Extensive assembler packs updates, most notably:
7042 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7043 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7044 - x86_64: bit-sliced AES implementation;
7045 - ARM: NEON support, contemporary platforms optimizations;
7046 - s390x: z196 support;
7047 - `*`: GHASH and GF(2^m) multiplication implementations;
7051 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7052 (removal of unnecessary code)
7054 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7056 * Add TLS key material exporter from RFC 5705.
7060 * Add DTLS-SRTP negotiation from RFC 5764.
7064 * Add Next Protocol Negotiation,
7065 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7066 disabled with a no-npn flag to config or Configure. Code donated
7069 *Adam Langley <agl@google.com> and Ben Laurie*
7071 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7072 NIST-P256, NIST-P521, with constant-time single point multiplication on
7073 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7074 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7075 Code made available under Apache License version 2.0.
7077 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7078 line to include this in your build of OpenSSL, and run "make depend" (or
7079 "make update"). This enables the following EC_METHODs:
7081 EC_GFp_nistp224_method()
7082 EC_GFp_nistp256_method()
7083 EC_GFp_nistp521_method()
7085 EC_GROUP_new_by_curve_name() will automatically use these (while
7086 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7089 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7091 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7092 all platforms. Move ssize_t definition from e_os.h to the public
7093 header file e_os2.h as it now appears in public header file cms.h
7097 * New -sigopt option to the ca, req and x509 utilities. Additional
7098 signature parameters can be passed using this option and in
7103 * Add RSA PSS signing function. This will generate and set the
7104 appropriate AlgorithmIdentifiers for PSS based on those in the
7105 corresponding EVP_MD_CTX structure. No application support yet.
7109 * Support for companion algorithm specific ASN1 signing routines.
7110 New function ASN1_item_sign_ctx() signs a pre-initialised
7111 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7112 the appropriate parameters.
7116 * Add new algorithm specific ASN1 verification initialisation function
7117 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7118 handling will be the same no matter what EVP_PKEY_METHOD is used.
7119 Add a PSS handler to support verification of PSS signatures: checked
7120 against a number of sample certificates.
7124 * Add signature printing for PSS. Add PSS OIDs.
7126 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7128 * Add algorithm specific signature printing. An individual ASN1 method
7129 can now print out signatures instead of the standard hex dump.
7131 More complex signatures (e.g. PSS) can print out more meaningful
7132 information. Include DSA version that prints out the signature
7137 * Password based recipient info support for CMS library: implementing
7142 * Split password based encryption into PBES2 and PBKDF2 functions. This
7143 neatly separates the code into cipher and PBE sections and is required
7144 for some algorithms that split PBES2 into separate pieces (such as
7145 password based CMS).
7149 * Session-handling fixes:
7150 - Fix handling of connections that are resuming with a session ID,
7151 but also support Session Tickets.
7152 - Fix a bug that suppressed issuing of a new ticket if the client
7153 presented a ticket with an expired session.
7154 - Try to set the ticket lifetime hint to something reasonable.
7155 - Make tickets shorter by excluding irrelevant information.
7156 - On the client side, don't ignore renewed tickets.
7158 *Adam Langley, Bodo Moeller (Google)*
7160 * Fix PSK session representation.
7164 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7166 This work was sponsored by Intel.
7170 * Add GCM support to TLS library. Some custom code is needed to split
7171 the IV between the fixed (from PRF) and explicit (from TLS record)
7172 portions. This adds all GCM ciphersuites supported by RFC5288 and
7173 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7174 add a special AESGCM string for GCM only.
7178 * Expand range of ctrls for AES GCM. Permit setting invocation
7179 field on decrypt and retrieval of invocation field only on encrypt.
7183 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7184 As required by RFC5289 these ciphersuites cannot be used if for
7185 versions of TLS earlier than 1.2.
7189 * For FIPS capable OpenSSL interpret a NULL default public key method
7190 as unset and return the appropriate default but do *not* set the default.
7191 This means we can return the appropriate method in applications that
7192 switch between FIPS and non-FIPS modes.
7196 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7197 ENGINE is used then we cannot handle that in the FIPS module so we
7198 keep original code iff non-FIPS operations are allowed.
7202 * Add -attime option to openssl utilities.
7204 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7206 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7210 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7211 FIPS EC methods unconditionally for now.
7215 * New build option no-ec2m to disable characteristic 2 code.
7219 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7220 all cases can be covered as some introduce binary incompatibilities.
7224 * Redirect RSA operations to FIPS module including keygen,
7225 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7229 * Add similar low level API blocking to ciphers.
7233 * Low level digest APIs are not approved in FIPS mode: any attempt
7234 to use these will cause a fatal error. Applications that *really* want
7235 to use them can use the `private_*` version instead.
7239 * Redirect cipher operations to FIPS module for FIPS builds.
7243 * Redirect digest operations to FIPS module for FIPS builds.
7247 * Update build system to add "fips" flag which will link in fipscanister.o
7248 for static and shared library builds embedding a signature if needed.
7252 * Output TLS supported curves in preference order instead of numerical
7253 order. This is currently hardcoded for the highest order curves first.
7254 This should be configurable so applications can judge speed vs strength.
7258 * Add TLS v1.2 server support for client authentication.
7262 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7267 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7268 FIPS modules versions.
7272 * Add TLS v1.2 client side support for client authentication. Keep cache
7273 of handshake records longer as we don't know the hash algorithm to use
7274 until after the certificate request message is received.
7278 * Initial TLS v1.2 client support. Add a default signature algorithms
7279 extension including all the algorithms we support. Parse new signature
7280 format in client key exchange. Relax some ECC signing restrictions for
7281 TLS v1.2 as indicated in RFC5246.
7285 * Add server support for TLS v1.2 signature algorithms extension. Switch
7286 to new signature format when needed using client digest preference.
7287 All server ciphersuites should now work correctly in TLS v1.2. No client
7288 support yet and no support for client certificates.
7292 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7293 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7294 ciphersuites. At present only RSA key exchange ciphersuites work with
7295 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7296 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7297 and version checking.
7301 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7302 with this defined it will not be affected by any changes to ssl internal
7303 structures. Add several utility functions to allow openssl application
7304 to work with OPENSSL_NO_SSL_INTERN defined.
7308 * A long standing patch to add support for SRP from EdelWeb (Peter
7309 Sylvester and Christophe Renou) was integrated.
7310 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7311 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7314 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7318 * Permit abbreviated handshakes when renegotiating using the function
7319 SSL_renegotiate_abbreviated().
7321 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7323 * Add call to ENGINE_register_all_complete() to
7324 ENGINE_load_builtin_engines(), so some implementations get used
7325 automatically instead of needing explicit application support.
7329 * Add support for TLS key exporter as described in RFC5705.
7331 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7333 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7334 a few changes are required:
7336 Add SSL_OP_NO_TLSv1_1 flag.
7337 Add TLSv1_1 methods.
7338 Update version checking logic to handle version 1.1.
7339 Add explicit IV handling (ported from DTLS code).
7340 Add command line options to s_client/s_server.
7347 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7349 * X509_ATTRIBUTE memory leak
7351 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7352 memory. This structure is used by the PKCS#7 and CMS routines so any
7353 application which reads PKCS#7 or CMS data from untrusted sources is
7354 affected. SSL/TLS is not affected.
7356 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7362 * Race condition handling PSK identify hint
7364 If PSK identity hints are received by a multi-threaded client then
7365 the values are wrongly updated in the parent SSL_CTX structure. This can
7366 result in a race condition potentially leading to a double free of the
7372 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7374 * Malformed ECParameters causes infinite loop
7376 When processing an ECParameters structure OpenSSL enters an infinite loop
7377 if the curve specified is over a specially malformed binary polynomial
7380 This can be used to perform denial of service against any
7381 system which processes public keys, certificate requests or
7382 certificates. This includes TLS clients and TLS servers with
7383 client authentication enabled.
7385 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7390 * Exploitable out-of-bounds read in X509_cmp_time
7392 X509_cmp_time does not properly check the length of the ASN1_TIME
7393 string and can read a few bytes out of bounds. In addition,
7394 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7397 An attacker can use this to craft malformed certificates and CRLs of
7398 various sizes and potentially cause a segmentation fault, resulting in
7399 a DoS on applications that verify certificates or CRLs. TLS clients
7400 that verify CRLs are affected. TLS clients and servers with client
7401 authentication enabled may be affected if they use custom verification
7404 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7405 independently by Hanno Böck.
7410 * PKCS7 crash with missing EnvelopedContent
7412 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7413 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7414 with missing content and trigger a NULL pointer dereference on parsing.
7416 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7417 structures from untrusted sources are affected. OpenSSL clients and
7418 servers are not affected.
7420 This issue was reported to OpenSSL by Michal Zalewski (Google).
7425 * CMS verify infinite loop with unknown hash function
7427 When verifying a signedData message the CMS code can enter an infinite loop
7428 if presented with an unknown hash function OID. This can be used to perform
7429 denial of service against any system which verifies signedData messages using
7431 This issue was reported to OpenSSL by Johannes Bauer.
7436 * Race condition handling NewSessionTicket
7438 If a NewSessionTicket is received by a multi-threaded client when attempting to
7439 reuse a previous ticket then a race condition can occur potentially leading to
7440 a double free of the ticket data.
7445 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7447 * Segmentation fault in ASN1_TYPE_cmp fix
7449 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7450 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7451 certificate signature algorithm consistency this can be used to crash any
7452 certificate verification operation and exploited in a DoS attack. Any
7453 application which performs certificate verification is vulnerable including
7454 OpenSSL clients and servers which enable client authentication.
7459 * ASN.1 structure reuse memory corruption fix
7461 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7462 memory corruption via an invalid write. Such reuse is and has been
7463 strongly discouraged and is believed to be rare.
7465 Applications that parse structures containing CHOICE or ANY DEFINED BY
7466 components may be affected. Certificate parsing (d2i_X509 and related
7467 functions) are however not affected. OpenSSL clients and servers are
7473 * PKCS7 NULL pointer dereferences fix
7475 The PKCS#7 parsing code does not handle missing outer ContentInfo
7476 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7477 missing content and trigger a NULL pointer dereference on parsing.
7479 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7480 otherwise parse PKCS#7 structures from untrusted sources are
7481 affected. OpenSSL clients and servers are not affected.
7483 This issue was reported to OpenSSL by Michal Zalewski (Google).
7488 * DoS via reachable assert in SSLv2 servers fix
7490 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7491 servers that both support SSLv2 and enable export cipher suites by sending
7492 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7494 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7495 (OpenSSL development team).
7500 * Use After Free following d2i_ECPrivatekey error fix
7502 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7503 could cause a use after free condition. This, in turn, could cause a double
7504 free in several private key parsing functions (such as d2i_PrivateKey
7505 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7506 for applications that receive EC private keys from untrusted
7507 sources. This scenario is considered rare.
7509 This issue was discovered by the BoringSSL project and fixed in their
7515 * X509_to_X509_REQ NULL pointer deref fix
7517 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7518 the certificate key is invalid. This function is rarely used in practice.
7520 This issue was discovered by Brian Carpenter.
7525 * Removed the export ciphers from the DEFAULT ciphers
7529 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7531 * Build fixes for the Windows and OpenVMS platforms
7533 *Matt Caswell and Richard Levitte*
7535 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7537 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7538 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7539 dereference. This could lead to a Denial Of Service attack. Thanks to
7540 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7545 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7546 dtls1_buffer_record function under certain conditions. In particular this
7547 could occur if an attacker sent repeated DTLS records with the same
7548 sequence number but for the next epoch. The memory leak could be exploited
7549 by an attacker in a Denial of Service attack through memory exhaustion.
7550 Thanks to Chris Mueller for reporting this issue.
7555 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7556 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7557 method would be set to NULL which could later result in a NULL pointer
7558 dereference. Thanks to Frank Schmirler for reporting this issue.
7563 * Abort handshake if server key exchange message is omitted for ephemeral
7566 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7567 reporting this issue.
7572 * Remove non-export ephemeral RSA code on client and server. This code
7573 violated the TLS standard by allowing the use of temporary RSA keys in
7574 non-export ciphersuites and could be used by a server to effectively
7575 downgrade the RSA key length used to a value smaller than the server
7576 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7577 INRIA or reporting this issue.
7582 * Fixed issue where DH client certificates are accepted without verification.
7583 An OpenSSL server will accept a DH certificate for client authentication
7584 without the certificate verify message. This effectively allows a client to
7585 authenticate without the use of a private key. This only affects servers
7586 which trust a client certificate authority which issues certificates
7587 containing DH keys: these are extremely rare and hardly ever encountered.
7588 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7594 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7595 results on some platforms, including x86_64. This bug occurs at random
7596 with a very low probability, and is not known to be exploitable in any
7597 way, though its exact impact is difficult to determine. Thanks to Pieter
7598 Wuille (Blockstream) who reported this issue and also suggested an initial
7599 fix. Further analysis was conducted by the OpenSSL development team and
7600 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7601 the OpenSSL core team.
7606 * Fix various certificate fingerprint issues.
7608 By using non-DER or invalid encodings outside the signed portion of a
7609 certificate the fingerprint can be changed without breaking the signature.
7610 Although no details of the signed portion of the certificate can be changed
7611 this can cause problems with some applications: e.g. those using the
7612 certificate fingerprint for blacklists.
7614 1. Reject signatures with non zero unused bits.
7616 If the BIT STRING containing the signature has non zero unused bits reject
7617 the signature. All current signature algorithms require zero unused bits.
7619 2. Check certificate algorithm consistency.
7621 Check the AlgorithmIdentifier inside TBS matches the one in the
7622 certificate signature. NB: this will result in signature failure
7623 errors for some broken certificates.
7625 Thanks to Konrad Kraszewski from Google for reporting this issue.
7627 3. Check DSA/ECDSA signatures use DER.
7629 Reencode DSA/ECDSA signatures and compare with the original received
7630 signature. Return an error if there is a mismatch.
7632 This will reject various cases including garbage after signature
7633 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7634 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7635 (negative or with leading zeroes).
7637 Further analysis was conducted and fixes were developed by Stephen Henson
7638 of the OpenSSL core team.
7644 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7646 * Session Ticket Memory Leak.
7648 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7649 integrity of that ticket is first verified. In the event of a session
7650 ticket integrity check failing, OpenSSL will fail to free memory
7651 causing a memory leak. By sending a large number of invalid session
7652 tickets an attacker could exploit this issue in a Denial Of Service
7658 * Build option no-ssl3 is incomplete.
7660 When OpenSSL is configured with "no-ssl3" as a build option, servers
7661 could accept and complete a SSL 3.0 handshake, and clients could be
7662 configured to send them.
7665 *Akamai and the OpenSSL team*
7667 * Add support for TLS_FALLBACK_SCSV.
7668 Client applications doing fallback retries should call
7669 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7672 *Adam Langley, Bodo Moeller*
7674 * Add additional DigestInfo checks.
7676 Reencode DigestInto in DER and check against the original when
7677 verifying RSA signature: this will reject any improperly encoded
7678 DigestInfo structures.
7680 Note: this is a precautionary measure and no attacks are currently known.
7684 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7686 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7687 to a denial of service attack. A malicious server can crash the client
7688 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7689 ciphersuite and sending carefully crafted handshake messages.
7691 Thanks to Felix Gröbert (Google) for discovering and researching this
7697 * By sending carefully crafted DTLS packets an attacker could cause openssl
7698 to leak memory. This can be exploited through a Denial of Service attack.
7699 Thanks to Adam Langley for discovering and researching this issue.
7704 * An attacker can force openssl to consume large amounts of memory whilst
7705 processing DTLS handshake messages. This can be exploited through a
7706 Denial of Service attack.
7707 Thanks to Adam Langley for discovering and researching this issue.
7712 * An attacker can force an error condition which causes openssl to crash
7713 whilst processing DTLS packets due to memory being freed twice. This
7714 can be exploited through a Denial of Service attack.
7715 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7721 * If a multithreaded client connects to a malicious server using a resumed
7722 session and the server sends an ec point format extension it could write
7723 up to 255 bytes to freed memory.
7725 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7731 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7732 X509_name_oneline, X509_name_print_ex et al. to leak some information
7733 from the stack. Applications may be affected if they echo pretty printing
7734 output to the attacker.
7736 Thanks to Ivan Fratric (Google) for discovering this issue.
7739 *Emilia Käsper, and Steve Henson*
7741 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7742 for corner cases. (Certain input points at infinity could lead to
7743 bogus results, with non-infinity inputs mapped to infinity too.)
7747 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7749 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7750 handshake can force the use of weak keying material in OpenSSL
7751 SSL/TLS clients and servers.
7753 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7754 researching this issue. ([CVE-2014-0224])
7756 *KIKUCHI Masashi, Steve Henson*
7758 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7759 OpenSSL DTLS client the code can be made to recurse eventually crashing
7762 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7765 *Imre Rad, Steve Henson*
7767 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7768 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7769 client or server. This is potentially exploitable to run arbitrary
7770 code on a vulnerable client or server.
7772 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7774 *Jüri Aedla, Steve Henson*
7776 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7777 are subject to a denial of service attack.
7779 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7780 this issue. ([CVE-2014-3470])
7782 *Felix Gröbert, Ivan Fratric, Steve Henson*
7784 * Harmonize version and its documentation. -f flag is used to display
7787 *mancha <mancha1@zoho.com>*
7789 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7790 in i2d_ECPrivateKey.
7792 *mancha <mancha1@zoho.com>*
7794 * Fix some double frees. These are not thought to be exploitable.
7796 *mancha <mancha1@zoho.com>*
7798 * Fix for the attack described in the paper "Recovering OpenSSL
7799 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7800 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7801 <http://eprint.iacr.org/2014/140>
7803 Thanks to Yuval Yarom and Naomi Benger for discovering this
7804 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7806 *Yuval Yarom and Naomi Benger*
7808 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7810 * Keep original DTLS digest and encryption contexts in retransmission
7811 structures so we can use the previous session parameters if they need
7812 to be resent. ([CVE-2013-6450])
7816 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7817 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7818 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7819 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7820 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7821 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7823 *Rob Stradling, Adam Langley*
7825 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7827 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7829 This addresses the flaw in CBC record processing discovered by
7830 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7831 at: <http://www.isg.rhul.ac.uk/tls/>
7833 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7834 Security Group at Royal Holloway, University of London
7835 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7836 Emilia Käsper for the initial patch.
7839 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7841 * Return an error when checking OCSP signatures when key is NULL.
7842 This fixes a DoS attack. ([CVE-2013-0166])
7846 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7847 the right response is stapled. Also change SSL_get_certificate()
7848 so it returns the certificate actually sent.
7849 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7850 (This is a backport)
7852 *Rob Stradling <rob.stradling@comodo.com>*
7854 * Fix possible deadlock when decoding public keys.
7858 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7860 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7863 * Sanity check record length before skipping explicit IV in DTLS
7866 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7867 fuzzing as a service testing platform.
7872 * Initialise tkeylen properly when encrypting CMS messages.
7873 Thanks to Solar Designer of Openwall for reporting this issue.
7877 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7879 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7880 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7881 in CRYPTO_realloc_clean.
7883 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7884 issue and to Adam Langley <agl@chromium.org> for fixing it.
7887 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7889 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7891 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7892 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7893 content decryption and always return the same error. Note: this attack
7894 needs on average 2^20 messages so it only affects automated senders. The
7895 old behaviour can be re-enabled in the CMS code by setting the
7896 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7897 an MMA defence is not necessary.
7898 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7899 this issue. ([CVE-2012-0884])
7903 * Fix CVE-2011-4619: make sure we really are receiving a
7904 client hello before rejecting multiple SGC restarts. Thanks to
7905 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7909 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7911 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7912 Thanks to Antonio Martin, Enterprise Secure Access Research and
7913 Development, Cisco Systems, Inc. for discovering this bug and
7914 preparing a fix. ([CVE-2012-0050])
7918 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7920 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7921 of the Vaudenay padding oracle attack on CBC mode encryption
7922 which enables an efficient plaintext recovery attack against
7923 the OpenSSL implementation of DTLS. Their attack exploits timing
7924 differences arising during decryption processing. A research
7925 paper describing this attack can be found at:
7926 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7927 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7928 Security Group at Royal Holloway, University of London
7929 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7930 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7931 for preparing the fix. ([CVE-2011-4108])
7933 *Robin Seggelmann, Michael Tuexen*
7935 * Clear bytes used for block padding of SSL 3.0 records.
7938 *Adam Langley (Google)*
7940 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7941 Kadianakis <desnacked@gmail.com> for discovering this issue and
7942 Adam Langley for preparing the fix. ([CVE-2011-4619])
7944 *Adam Langley (Google)*
7946 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7948 *Andrey Kulikov <amdeich@gmail.com>*
7950 * Prevent malformed RFC3779 data triggering an assertion failure.
7951 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7952 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7954 *Rob Austein <sra@hactrn.net>*
7956 * Improved PRNG seeding for VOS.
7958 *Paul Green <Paul.Green@stratus.com>*
7960 * Fix ssl_ciph.c set-up race.
7962 *Adam Langley (Google)*
7964 * Fix spurious failures in ecdsatest.c.
7966 *Emilia Käsper (Google)*
7968 * Fix the BIO_f_buffer() implementation (which was mixing different
7969 interpretations of the `..._len` fields).
7971 *Adam Langley (Google)*
7973 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7974 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7975 threads won't reuse the same blinding coefficients.
7977 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7978 lock to call BN_BLINDING_invert_ex, and avoids one use of
7979 BN_BLINDING_update for each BN_BLINDING structure (previously,
7980 the last update always remained unused).
7982 *Emilia Käsper (Google)*
7984 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7986 *Bob Buckholz (Google)*
7988 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7990 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7991 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7993 *Kaspar Brand <ossl@velox.ch>*
7995 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7996 for multi-threaded use of ECDH. ([CVE-2011-3210])
7998 *Adam Langley (Google)*
8000 * Fix x509_name_ex_d2i memory leak on bad inputs.
8004 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8005 signature public key algorithm by using OID xref utilities instead.
8006 Before this you could only use some ECC ciphersuites with SHA1 only.
8010 * Add protection against ECDSA timing attacks as mentioned in the paper
8011 by Billy Bob Brumley and Nicola Tuveri, see:
8012 <http://eprint.iacr.org/2011/232.pdf>
8014 *Billy Bob Brumley and Nicola Tuveri*
8016 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8018 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8020 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8022 * Fix bug in string printing code: if *any* escaping is enabled we must
8023 escape the escape character (backslash) or the resulting string is
8028 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8030 * Disable code workaround for ancient and obsolete Netscape browsers
8031 and servers: an attacker can use it in a ciphersuite downgrade attack.
8032 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8036 * Fixed J-PAKE implementation error, originally discovered by
8037 Sebastien Martini, further info and confirmation from Stefan
8038 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8042 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8044 * Fix extension code to avoid race conditions which can result in a buffer
8045 overrun vulnerability: resumed sessions must not be modified as they can
8046 be shared by multiple threads. CVE-2010-3864
8050 * Fix WIN32 build system to correctly link an ENGINE directory into
8055 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8057 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8060 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8062 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8064 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8065 context. The operation can be customised via the ctrl mechanism in
8066 case ENGINEs want to include additional functionality.
8070 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8074 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8075 output hashes compatible with older versions of OpenSSL.
8077 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8079 * Fix compression algorithm handling: if resuming a session use the
8080 compression algorithm of the resumed session instead of determining
8081 it from client hello again. Don't allow server to change algorithm.
8085 * Add load_crls() function to commands tidying load_certs() too. Add option
8086 to verify utility to allow additional CRLs to be included.
8090 * Update OCSP request code to permit adding custom headers to the request:
8091 some responders need this.
8095 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8098 *Julia Lawall <julia@diku.dk>*
8100 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8101 needlessly dereferenced structures, used obsolete functions and
8102 didn't handle all updated verify codes correctly.
8106 * Disable MD2 in the default configuration.
8110 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8111 indicate the initial BIO being pushed or popped. This makes it possible
8112 to determine whether the BIO is the one explicitly called or as a result
8113 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8114 it handles reference counts correctly and doesn't zero out the I/O bio
8115 when it is not being explicitly popped. WARNING: applications which
8116 included workarounds for the old buggy behaviour will need to be modified
8117 or they could free up already freed BIOs.
8121 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8122 renaming to all platforms (within the 0.9.8 branch, this was
8123 done conditionally on Netware platforms to avoid a name clash).
8125 *Guenter <lists@gknw.net>*
8127 * Add ECDHE and PSK support to DTLS.
8129 *Michael Tuexen <tuexen@fh-muenster.de>*
8131 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8136 * Add "missing" function EVP_MD_flags() (without this the only way to
8137 retrieve a digest flags is by accessing the structure directly. Update
8138 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8139 or cipher is registered as in the "from" argument. Print out all
8140 registered digests in the dgst usage message instead of manually
8141 attempting to work them out.
8145 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8146 this allows the use of compression and extensions. Change default cipher
8147 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8148 by default unless an application cipher string requests it.
8152 * Alter match criteria in PKCS12_parse(). It used to try to use local
8153 key ids to find matching certificates and keys but some PKCS#12 files
8154 don't follow the (somewhat unwritten) rules and this strategy fails.
8155 Now just gather all certificates together and the first private key
8156 then look for the first certificate that matches the key.
8160 * Support use of registered digest and cipher names for dgst and cipher
8161 commands instead of having to add each one as a special case. So now
8168 openssl dgst -sha256 foo
8170 and this works for ENGINE based algorithms too.
8174 * Update Gost ENGINE to support parameter files.
8176 *Victor B. Wagner <vitus@cryptocom.ru>*
8178 * Support GeneralizedTime in ca utility.
8180 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8182 * Enhance the hash format used for certificate directory links. The new
8183 form uses the canonical encoding (meaning equivalent names will work
8184 even if they aren't identical) and uses SHA1 instead of MD5. This form
8185 is incompatible with the older format and as a result c_rehash should
8186 be used to rebuild symbolic links.
8190 * Make PKCS#8 the default write format for private keys, replacing the
8191 traditional format. This form is standardised, more secure and doesn't
8192 include an implicit MD5 dependency.
8196 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8197 committed to OpenSSL should pass this lot as a minimum.
8201 * Add session ticket override functionality for use by EAP-FAST.
8203 *Jouni Malinen <j@w1.fi>*
8205 * Modify HMAC functions to return a value. Since these can be implemented
8206 in an ENGINE errors can occur.
8210 * Type-checked OBJ_bsearch_ex.
8214 * Type-checked OBJ_bsearch. Also some constification necessitated
8215 by type-checking. Still to come: TXT_DB, bsearch(?),
8216 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8221 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8222 seconds to a tm structure directly, instead of going through OS
8223 specific date routines. This avoids any issues with OS routines such
8224 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8225 and X509_time_adj_ex() to cover the extended range. The existing
8226 X509_time_adj() is still usable and will no longer have any date issues.
8230 * Delta CRL support. New use deltas option which will attempt to locate
8231 and search any appropriate delta CRLs available.
8233 This work was sponsored by Google.
8237 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8238 code and add additional score elements. Validate alternate CRL paths
8239 as part of the CRL checking and indicate a new error "CRL path validation
8240 error" in this case. Applications wanting additional details can use
8241 the verify callback and check the new "parent" field. If this is not
8242 NULL CRL path validation is taking place. Existing applications won't
8243 see this because it requires extended CRL support which is off by
8246 This work was sponsored by Google.
8250 * Support for freshest CRL extension.
8252 This work was sponsored by Google.
8256 * Initial indirect CRL support. Currently only supported in the CRLs
8257 passed directly and not via lookup. Process certificate issuer
8258 CRL entry extension and lookup CRL entries by bother issuer name
8259 and serial number. Check and process CRL issuer entry in IDP extension.
8261 This work was sponsored by Google.
8265 * Add support for distinct certificate and CRL paths. The CRL issuer
8266 certificate is validated separately in this case. Only enabled if
8267 an extended CRL support flag is set: this flag will enable additional
8268 CRL functionality in future.
8270 This work was sponsored by Google.
8274 * Add support for policy mappings extension.
8276 This work was sponsored by Google.
8280 * Fixes to pathlength constraint, self issued certificate handling,
8281 policy processing to align with RFC3280 and PKITS tests.
8283 This work was sponsored by Google.
8287 * Support for name constraints certificate extension. DN, email, DNS
8288 and URI types are currently supported.
8290 This work was sponsored by Google.
8294 * To cater for systems that provide a pointer-based thread ID rather
8295 than numeric, deprecate the current numeric thread ID mechanism and
8296 replace it with a structure and associated callback type. This
8297 mechanism allows a numeric "hash" to be extracted from a thread ID in
8298 either case, and on platforms where pointers are larger than 'long',
8299 mixing is done to help ensure the numeric 'hash' is usable even if it
8300 can't be guaranteed unique. The default mechanism is to use "&errno"
8301 as a pointer-based thread ID to distinguish between threads.
8303 Applications that want to provide their own thread IDs should now use
8304 CRYPTO_THREADID_set_callback() to register a callback that will call
8305 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8307 Note that ERR_remove_state() is now deprecated, because it is tied
8308 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8309 to free the current thread's error state should be replaced by
8310 ERR_remove_thread_state(NULL).
8312 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8313 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8314 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8315 application was previously providing a numeric thread callback that
8316 was inappropriate for distinguishing threads, then uniqueness might
8317 have been obtained with &errno that happened immediately in the
8318 intermediate development versions of OpenSSL; this is no longer the
8319 case, the numeric thread callback will now override the automatic use
8322 *Geoff Thorpe, with help from Bodo Moeller*
8324 * Initial support for different CRL issuing certificates. This covers a
8325 simple case where the self issued certificates in the chain exist and
8326 the real CRL issuer is higher in the existing chain.
8328 This work was sponsored by Google.
8332 * Removed effectively defunct crypto/store from the build.
8336 * Revamp of STACK to provide stronger type-checking. Still to come:
8337 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8338 ASN1_STRING, CONF_VALUE.
8342 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8343 RAM on SSL connections. This option can save about 34k per idle SSL.
8347 * Revamp of LHASH to provide stronger type-checking. Still to come:
8348 STACK, TXT_DB, bsearch, qsort.
8352 * Initial support for Cryptographic Message Syntax (aka CMS) based
8353 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8354 support for data, signedData, compressedData, digestedData and
8355 encryptedData, envelopedData types included. Scripts to check against
8356 RFC4134 examples draft and interop and consistency checks of many
8357 content types and variants.
8361 * Add options to enc utility to support use of zlib compression BIO.
8365 * Extend mk1mf to support importing of options and assembly language
8366 files from Configure script, currently only included in VC-WIN32.
8367 The assembly language rules can now optionally generate the source
8368 files from the associated perl scripts.
8372 * Implement remaining functionality needed to support GOST ciphersuites.
8373 Interop testing has been performed using CryptoPro implementations.
8375 *Victor B. Wagner <vitus@cryptocom.ru>*
8377 * s390x assembler pack.
8381 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8386 * Implement Opaque PRF Input TLS extension as specified in
8387 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8388 official specification yet and no extension type assignment by
8389 IANA exists, this extension (for now) will have to be explicitly
8390 enabled when building OpenSSL by providing the extension number
8391 to use. For example, specify an option
8393 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8395 to the "config" or "Configure" script to enable the extension,
8396 assuming extension number 0x9527 (which is a completely arbitrary
8397 and unofficial assignment based on the MD5 hash of the Internet
8398 Draft). Note that by doing so, you potentially lose
8399 interoperability with other TLS implementations since these might
8400 be using the same extension number for other purposes.
8402 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8403 opaque PRF input value to use in the handshake. This will create
8404 an internal copy of the length-'len' string at 'src', and will
8405 return non-zero for success.
8407 To get more control and flexibility, provide a callback function
8410 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8411 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8415 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8418 Callback function 'cb' will be called in handshakes, and is
8419 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8420 Argument 'arg' is for application purposes (the value as given to
8421 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8422 be provided to the callback function). The callback function
8423 has to return non-zero to report success: usually 1 to use opaque
8424 PRF input just if possible, or 2 to enforce use of the opaque PRF
8425 input. In the latter case, the library will abort the handshake
8426 if opaque PRF input is not successfully negotiated.
8428 Arguments 'peerinput' and 'len' given to the callback function
8429 will always be NULL and 0 in the case of a client. A server will
8430 see the client's opaque PRF input through these variables if
8431 available (NULL and 0 otherwise). Note that if the server
8432 provides an opaque PRF input, the length must be the same as the
8433 length of the client's opaque PRF input.
8435 Note that the callback function will only be called when creating
8436 a new session (session resumption can resume whatever was
8437 previously negotiated), and will not be called in SSL 2.0
8438 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8439 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8440 for applications that need to enforce opaque PRF input.
8444 * Update ssl code to support digests other than SHA1+MD5 for handshake
8447 *Victor B. Wagner <vitus@cryptocom.ru>*
8449 * Add RFC4507 support to OpenSSL. This includes the corrections in
8450 RFC4507bis. The encrypted ticket format is an encrypted encoded
8451 SSL_SESSION structure, that way new session features are automatically
8454 If a client application caches session in an SSL_SESSION structure
8455 support is transparent because tickets are now stored in the encoded
8458 The SSL_CTX structure automatically generates keys for ticket
8459 protection in servers so again support should be possible
8460 with no application modification.
8462 If a client or server wishes to disable RFC4507 support then the option
8463 SSL_OP_NO_TICKET can be set.
8465 Add a TLS extension debugging callback to allow the contents of any client
8466 or server extensions to be examined.
8468 This work was sponsored by Google.
8472 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8473 OpenSSL should now compile cleanly on gcc 4.2
8475 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8477 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8478 support including streaming MAC support: this is required for GOST
8479 ciphersuite support.
8481 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8483 * Add option -stream to use PKCS#7 streaming in smime utility. New
8484 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8485 to output in BER and PEM format.
8489 * Experimental support for use of HMAC via EVP_PKEY interface. This
8490 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8491 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8492 ENGINE support for HMAC keys which are unextractable. New -mac and
8493 -macopt options to dgst utility.
8497 * New option -sigopt to dgst utility. Update dgst to use
8498 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8499 alternative signing parameters such as X9.31 or PSS in the dgst
8504 * Change ssl_cipher_apply_rule(), the internal function that does
8505 the work each time a ciphersuite string requests enabling
8506 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8507 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8508 the order of disabled ciphersuites such that those ciphersuites
8509 that most recently went from enabled to disabled not only stay
8510 in order with respect to each other, but also have higher priority
8511 than other disabled ciphersuites the next time ciphersuites are
8514 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8515 the same ciphersuites as with "HIGH" alone, but in a specific
8516 order where the PSK ciphersuites come first (since they are the
8517 most recently disabled ciphersuites when "HIGH" is parsed).
8519 Also, change ssl_create_cipher_list() (using this new
8520 functionality) such that between otherwise identical
8521 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8526 * Change ssl_create_cipher_list() so that it automatically
8527 arranges the ciphersuites in reasonable order before starting
8528 to process the rule string. Thus, the definition for "DEFAULT"
8529 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8530 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8531 This makes it much easier to arrive at a reasonable default order
8532 in applications for which anonymous ciphers are OK (meaning
8533 that you can't actually use DEFAULT).
8535 *Bodo Moeller; suggested by Victor Duchovni*
8537 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8538 processing) into multiple integers instead of setting
8539 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8540 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8541 (These masks as well as the individual bit definitions are hidden
8542 away into the non-exported interface ssl/ssl_locl.h, so this
8543 change to the definition of the SSL_CIPHER structure shouldn't
8544 affect applications.) This give us more bits for each of these
8545 categories, so there is no longer a need to coagulate AES128 and
8546 AES256 into a single algorithm bit, and to coagulate Camellia128
8547 and Camellia256 into a single algorithm bit, which has led to all
8550 Thus, among other things, the kludge introduced in 0.9.7m and
8551 0.9.8e for masking out AES256 independently of AES128 or masking
8552 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8554 With the change, we also introduce new ciphersuite aliases that
8555 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8560 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8561 Use the leftmost N bytes of the signature input if the input is
8562 larger than the prime q (with N being the size in bytes of q).
8566 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8567 it yet and it is largely untested.
8571 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8575 * Initial incomplete changes to avoid need for function casts in OpenSSL
8576 some compilers (gcc 4.2 and later) reject their use. Safestack is
8577 reimplemented. Update ASN1 to avoid use of legacy functions.
8581 * Win32/64 targets are linked with Winsock2.
8585 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8586 to external functions. This can be used to increase CRL handling
8587 efficiency especially when CRLs are very large by (for example) storing
8588 the CRL revoked certificates in a database.
8592 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8593 new CRLs added to a directory can be used. New command line option
8594 -verify_return_error to s_client and s_server. This causes real errors
8595 to be returned by the verify callback instead of carrying on no matter
8596 what. This reflects the way a "real world" verify callback would behave.
8600 * GOST engine, supporting several GOST algorithms and public key formats.
8601 Kindly donated by Cryptocom.
8605 * Partial support for Issuing Distribution Point CRL extension. CRLs
8606 partitioned by DP are handled but no indirect CRL or reason partitioning
8607 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8608 selected via a scoring technique which handles IDP and AKID in CRLs.
8612 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8613 will ultimately be used for all verify operations: this will remove the
8614 X509_STORE dependency on certificate verification and allow alternative
8615 lookup methods. X509_STORE based implementations of these two callbacks.
8619 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8620 Modify get_crl() to find a valid (unexpired) CRL if possible.
8624 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8625 this would be called X509_CRL_cmp() but that name is already used by
8626 a function that just compares CRL issuer names. Cache several CRL
8627 extensions in X509_CRL structure and cache CRLDP in X509.
8631 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8632 this maps equivalent X509_NAME structures into a consistent structure.
8633 Name comparison can then be performed rapidly using memcmp().
8637 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8642 * Allow digests to supply their own micalg string for S/MIME type using
8643 the ctrl EVP_MD_CTRL_MICALG.
8647 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8648 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8649 ctrl. It can then customise the structure before and/or after signing
8654 * New function OBJ_add_sigid() to allow application defined signature OIDs
8655 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8656 to free up any added signature OIDs.
8660 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8661 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8662 digest and cipher tables. New options added to openssl utility:
8663 list-message-digest-algorithms and list-cipher-algorithms.
8667 * Change the array representation of binary polynomials: the list
8668 of degrees of non-zero coefficients is now terminated with -1.
8669 Previously it was terminated with 0, which was also part of the
8670 value; thus, the array representation was not applicable to
8671 polynomials where t^0 has coefficient zero. This change makes
8672 the array representation useful in a more general context.
8676 * Various modifications and fixes to SSL/TLS cipher string
8677 handling. For ECC, the code now distinguishes between fixed ECDH
8678 with RSA certificates on the one hand and with ECDSA certificates
8679 on the other hand, since these are separate ciphersuites. The
8680 unused code for Fortezza ciphersuites has been removed.
8682 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8683 (not "ECDHE"). For consistency with the code for DH
8684 certificates, use of ECDH certificates is now considered ECDH
8685 authentication, not RSA or ECDSA authentication (the latter is
8686 merely the CA's signing algorithm and not actively used in the
8689 The temporary ciphersuite alias "ECCdraft" is no longer
8690 available, and ECC ciphersuites are no longer excluded from "ALL"
8691 and "DEFAULT". The following aliases now exist for RFC 4492
8692 ciphersuites, most of these by analogy with the DH case:
8694 kECDHr - ECDH cert, signed with RSA
8695 kECDHe - ECDH cert, signed with ECDSA
8696 kECDH - ECDH cert (signed with either RSA or ECDSA)
8697 kEECDH - ephemeral ECDH
8698 ECDH - ECDH cert or ephemeral ECDH
8704 AECDH - anonymous ECDH
8705 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8709 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8710 Use correct micalg parameters depending on digest(s) in signed message.
8714 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8715 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8719 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8720 an engine to register a method. Add ENGINE lookups for methods and
8721 functional reference processing.
8725 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8726 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8731 * New -resign option to smime utility. This adds one or more signers
8732 to an existing PKCS#7 signedData structure. Also -md option to use an
8733 alternative message digest algorithm for signing.
8737 * Tidy up PKCS#7 routines and add new functions to make it easier to
8738 create PKCS7 structures containing multiple signers. Update smime
8739 application to support multiple signers.
8743 * New -macalg option to pkcs12 utility to allow setting of an alternative
8748 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8749 Reorganize PBE internals to lookup from a static table using NIDs,
8750 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8751 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8752 PRF which will be automatically used with PBES2.
8756 * Replace the algorithm specific calls to generate keys in "req" with the
8761 * Update PKCS#7 enveloped data routines to use new API. This is now
8762 supported by any public key method supporting the encrypt operation. A
8763 ctrl is added to allow the public key algorithm to examine or modify
8764 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8769 * Add a ctrl to asn1 method to allow a public key algorithm to express
8770 a default digest type to use. In most cases this will be SHA1 but some
8771 algorithms (such as GOST) need to specify an alternative digest. The
8772 return value indicates how strong the preference is 1 means optional and
8773 2 is mandatory (that is it is the only supported type). Modify
8774 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8775 use the default md. Update openssl utilities to use the default digest
8776 type for signing if it is not explicitly indicated.
8780 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8781 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8782 signing method from the key type. This effectively removes the link
8783 between digests and public key types.
8787 * Add an OID cross reference table and utility functions. Its purpose is to
8788 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8789 rsaEncryption. This will allow some of the algorithm specific hackery
8790 needed to use the correct OID to be removed.
8794 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8795 structures for PKCS7_sign(). They are now set up by the relevant public
8800 * Add provisional EC pkey method with support for ECDSA and ECDH.
8804 * Add support for key derivation (agreement) in the API, DH method and
8809 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8810 public and private key formats. As a side effect these add additional
8811 command line functionality not previously available: DSA signatures can be
8812 generated and verified using pkeyutl and DH key support and generation in
8819 *Oliver Tappe <zooey@hirschkaefer.de>*
8821 * New make target "install_html_docs" installs HTML renditions of the
8824 *Oliver Tappe <zooey@hirschkaefer.de>*
8826 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8827 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8828 support key and parameter generation and add initial key generation
8829 functionality for RSA.
8833 * Add functions for main EVP_PKEY_method operations. The undocumented
8834 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8835 `EVP_PKEY_{encrypt,decrypt}_old`.
8839 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8840 key API, doesn't do much yet.
8844 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8845 public key algorithms. New option to openssl utility:
8846 "list-public-key-algorithms" to print out info.
8850 * Implement the Supported Elliptic Curves Extension for
8851 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8855 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8856 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8860 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8861 utilities such as rsa, dsa, dsaparam etc except they process any key
8866 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8867 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8868 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8873 * Initial support for pluggable public key ASN1.
8874 De-spaghettify the public key ASN1 handling. Move public and private
8875 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8876 algorithm specific handling to a single module within the relevant
8877 algorithm directory. Add functions to allow (near) opaque processing
8878 of public and private key structures.
8882 * Implement the Supported Point Formats Extension for
8883 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8887 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8888 for the psk identity [hint] and the psk callback functions to the
8889 SSL_SESSION, SSL and SSL_CTX structure.
8892 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8896 SSL_CTX_use_psk_identity_hint
8897 SSL_get_psk_identity_hint
8898 SSL_get_psk_identity
8899 SSL_use_psk_identity_hint
8901 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8903 * Add RFC 3161 compliant time stamp request creation, response generation
8904 and response verification functionality.
8906 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8908 * Add initial support for TLS extensions, specifically for the server_name
8909 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8910 have new members for a host name. The SSL data structure has an
8911 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8912 stored in that context to allow for session resumption, even after the
8913 SSL has been switched to a new SSL_CTX in reaction to a client's
8914 server_name extension.
8916 New functions (subject to change):
8918 SSL_get_servername()
8919 SSL_get_servername_type()
8922 New CTRL codes and macros (subject to change):
8924 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8925 - SSL_CTX_set_tlsext_servername_callback()
8926 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8927 - SSL_CTX_set_tlsext_servername_arg()
8928 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8930 openssl s_client has a new '-servername ...' option.
8932 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8933 '-key2 ...', '-servername_fatal' (subject to change). This allows
8934 testing the HostName extension for a specific single host name ('-cert'
8935 and '-key' remain fallbacks for handshakes without HostName
8936 negotiation). If the unrecognized_name alert has to be sent, this by
8937 default is a warning; it becomes fatal with the '-servername_fatal'
8940 *Peter Sylvester, Remy Allais, Christophe Renou*
8942 * Whirlpool hash implementation is added.
8946 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8947 bn(64,32). Because of instruction set limitations it doesn't have
8948 any negative impact on performance. This was done mostly in order
8949 to make it possible to share assembler modules, such as bn_mul_mont
8950 implementations, between 32- and 64-bit builds without hassle.
8954 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8955 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8960 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8961 dedicated Montgomery multiplication procedure, is introduced.
8962 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8963 "64-bit" performance on certain 32-bit targets.
8967 * New option SSL_OP_NO_COMP to disable use of compression selectively
8968 in SSL structures. New SSL ctrl to set maximum send fragment size.
8969 Save memory by setting the I/O buffer sizes dynamically instead of
8970 using the maximum available value.
8974 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8975 in addition to the text details.
8979 * Very, very preliminary EXPERIMENTAL support for printing of general
8980 ASN1 structures. This currently produces rather ugly output and doesn't
8981 handle several customised structures at all.
8985 * Integrated support for PVK file format and some related formats such
8986 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8987 these in the 'rsa' and 'dsa' utilities.
8991 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8995 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8996 place for the (very old) "NETSCAPE" format certificates which are now
8997 handled using new ASN1 code equivalents.
9001 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9002 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9003 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9007 * Modify CRL distribution points extension code to print out previously
9008 unsupported fields. Enhance extension setting code to allow setting of
9013 * Add print and set support for Issuing Distribution Point CRL extension.
9017 * Change 'Configure' script to enable Camellia by default.
9024 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9026 * When rejecting SSL/TLS records due to an incorrect version number, never
9027 update s->server with a new major version number. As of
9028 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9029 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9030 the previous behavior could result in a read attempt at NULL when
9031 receiving specific incorrect SSL/TLS records once record payload
9032 protection is active. ([CVE-2010-0740])
9034 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9036 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9037 could be crashed if the relevant tables were not present (e.g. chrooted).
9039 *Tomas Hoger <thoger@redhat.com>*
9041 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9043 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9045 *Martin Olsson, Neel Mehta*
9047 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9048 accommodate for stack sorting, always a write lock!).
9052 * On some versions of WIN32 Heap32Next is very slow. This can cause
9053 excessive delays in the RAND_poll(): over a minute. As a workaround
9054 include a time check in the inner Heap32Next loop too.
9058 * The code that handled flushing of data in SSL/TLS originally used the
9059 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9060 the problem outlined in PR#1949. The fix suggested there however can
9061 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9062 of Apache). So instead simplify the code to flush unconditionally.
9063 This should be fine since flushing with no data to flush is a no op.
9067 * Handle TLS versions 2.0 and later properly and correctly use the
9068 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9069 off ancient servers have a habit of sticking around for a while...
9073 * Modify compression code so it frees up structures without using the
9074 ex_data callbacks. This works around a problem where some applications
9075 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9076 restarting) then use compression (e.g. SSL with compression) later.
9077 This results in significant per-connection memory leaks and
9078 has caused some security issues including CVE-2008-1678 and
9083 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9084 change when encrypting or decrypting.
9088 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9089 connect and renegotiate with servers which do not support RI.
9090 Until RI is more widely deployed this option is enabled by default.
9094 * Add "missing" ssl ctrls to clear options and mode.
9098 * If client attempts to renegotiate and doesn't support RI respond with
9099 a no_renegotiation alert as required by RFC5746. Some renegotiating
9100 TLS clients will continue a connection gracefully when they receive
9101 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9102 waiting for a server hello which it will never receive. Now we treat a
9103 received no_renegotiation alert as a fatal error. This is because
9104 applications requesting a renegotiation might well expect it to succeed
9105 and would have no code in place to handle the server denying it so the
9106 only safe thing to do is to terminate the connection.
9110 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9111 peer supports secure renegotiation and 0 otherwise. Print out peer
9112 renegotiation support in s_client/s_server.
9116 * Replace the highly broken and deprecated SPKAC certification method with
9117 the updated NID creation version. This should correctly handle UTF8.
9121 * Implement RFC5746. Re-enable renegotiation but require the extension
9122 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9123 turns out to be a bad idea. It has been replaced by
9124 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9125 SSL_CTX_set_options(). This is really not recommended unless you
9126 know what you are doing.
9128 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9130 * Fixes to stateless session resumption handling. Use initial_ctx when
9131 issuing and attempting to decrypt tickets in case it has changed during
9132 servername handling. Use a non-zero length session ID when attempting
9133 stateless session resumption: this makes it possible to determine if
9134 a resumption has occurred immediately after receiving server hello
9135 (several places in OpenSSL subtly assume this) instead of later in
9140 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9141 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9142 fixes for a few places where the return code is not checked
9145 *Julia Lawall <julia@diku.dk>*
9147 * Add --strict-warnings option to Configure script to include devteam
9148 warnings in other configurations.
9152 * Add support for --libdir option and LIBDIR variable in makefiles. This
9153 makes it possible to install openssl libraries in locations which
9154 have names other than "lib", for example "/usr/lib64" which some
9157 *Steve Henson, based on patch from Jeremy Utley*
9159 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9160 X690 8.9.12 and can produce some misleading textual output of OIDs.
9162 *Steve Henson, reported by Dan Kaminsky*
9164 * Delete MD2 from algorithm tables. This follows the recommendation in
9165 several standards that it is not used in new applications due to
9166 several cryptographic weaknesses. For binary compatibility reasons
9167 the MD2 API is still compiled in by default.
9171 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9176 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9177 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9180 *Guenter <lists@gknw.net>*
9182 * Fix the server certificate chain building code to use X509_verify_cert(),
9183 it used to have an ad-hoc builder which was unable to cope with anything
9184 other than a simple chain.
9186 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9188 * Don't check self signed certificate signatures in X509_verify_cert()
9189 by default (a flag can override this): it just wastes time without
9190 adding any security. As a useful side effect self signed root CAs
9191 with non-FIPS digests are now usable in FIPS mode.
9195 * In dtls1_process_out_of_seq_message() the check if the current message
9196 is already buffered was missing. For every new message was memory
9197 allocated, allowing an attacker to perform an denial of service attack
9198 with sending out of seq handshake messages until there is no memory
9199 left. Additionally every future message was buffered, even if the
9200 sequence number made no sense and would be part of another handshake.
9201 So only messages with sequence numbers less than 10 in advance will be
9202 buffered. ([CVE-2009-1378])
9204 *Robin Seggelmann, discovered by Daniel Mentz*
9206 * Records are buffered if they arrive with a future epoch to be
9207 processed after finishing the corresponding handshake. There is
9208 currently no limitation to this buffer allowing an attacker to perform
9209 a DOS attack with sending records with future epochs until there is no
9210 memory left. This patch adds the pqueue_size() function to determine
9211 the size of a buffer and limits the record buffer to 100 entries.
9214 *Robin Seggelmann, discovered by Daniel Mentz*
9216 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9217 parent structure is freed. ([CVE-2009-1379])
9221 * Handle non-blocking I/O properly in SSL_shutdown() call.
9223 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9225 * Add `2.5.4.*` OIDs
9227 *Ilya O. <vrghost@gmail.com>*
9229 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9231 * Disable renegotiation completely - this fixes a severe security
9232 problem ([CVE-2009-3555]) at the cost of breaking all
9233 renegotiation. Renegotiation can be re-enabled by setting
9234 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9235 run-time. This is really not recommended unless you know what
9240 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9242 * Don't set val to NULL when freeing up structures, it is freed up by
9243 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9244 zeroing past the valid field. ([CVE-2009-0789])
9246 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9248 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9249 checked correctly. This would allow some invalid signed attributes to
9250 appear to verify correctly. ([CVE-2009-0591])
9252 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9254 * Reject UniversalString and BMPString types with invalid lengths. This
9255 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9256 a legal length. ([CVE-2009-0590])
9260 * Set S/MIME signing as the default purpose rather than setting it
9261 unconditionally. This allows applications to override it at the store
9266 * Permit restricted recursion of ASN1 strings. This is needed in practice
9267 to handle some structures.
9271 * Improve efficiency of mem_gets: don't search whole buffer each time
9274 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9276 * New -hex option for openssl rand.
9280 * Print out UTF8String and NumericString when parsing ASN1.
9284 * Support NumericString type for name components.
9288 * Allow CC in the environment to override the automatically chosen
9289 compiler. Note that nothing is done to ensure flags work with the
9294 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9296 * Properly check EVP_VerifyFinal() and similar return values
9299 *Ben Laurie, Bodo Moeller, Google Security Team*
9301 * Enable TLS extensions by default.
9305 * Allow the CHIL engine to be loaded, whether the application is
9306 multithreaded or not. (This does not release the developer from the
9307 obligation to set up the dynamic locking callbacks.)
9309 *Sander Temme <sander@temme.net>*
9311 * Use correct exit code if there is an error in dgst command.
9313 *Steve Henson; problem pointed out by Roland Dirlewanger*
9315 * Tweak Configure so that you need to say "experimental-jpake" to enable
9316 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9320 * Add experimental JPAKE support, including demo authentication in
9321 s_client and s_server.
9325 * Set the comparison function in v3_addr_canonize().
9327 *Rob Austein <sra@hactrn.net>*
9329 * Add support for XMPP STARTTLS in s_client.
9331 *Philip Paeps <philip@freebsd.org>*
9333 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9334 to ensure that even with this option, only ciphersuites in the
9335 server's preference list will be accepted. (Note that the option
9336 applies only when resuming a session, so the earlier behavior was
9337 just about the algorithm choice for symmetric cryptography.)
9341 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9343 * Fix NULL pointer dereference if a DTLS server received
9344 ChangeCipherSpec as first record ([CVE-2009-1386]).
9348 * Fix a state transition in s3_srvr.c and d1_srvr.c
9349 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9353 * The fix in 0.9.8c that supposedly got rid of unsafe
9354 double-checked locking was incomplete for RSA blinding,
9355 addressing just one layer of what turns out to have been
9356 doubly unsafe triple-checked locking.
9358 So now fix this for real by retiring the MONT_HELPER macro
9359 in crypto/rsa/rsa_eay.c.
9361 *Bodo Moeller; problem pointed out by Marius Schilder*
9363 * Various precautionary measures:
9365 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9367 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9368 (NB: This would require knowledge of the secret session ticket key
9369 to exploit, in which case you'd be SOL either way.)
9371 - Change bn_nist.c so that it will properly handle input BIGNUMs
9372 outside the expected range.
9374 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9377 *Neel Mehta, Bodo Moeller*
9379 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9380 the load fails. Useful for distros.
9382 *Ben Laurie and the FreeBSD team*
9384 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9388 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9392 * Expand ENGINE to support engine supplied SSL client certificate functions.
9394 This work was sponsored by Logica.
9398 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9399 keystores. Support for SSL/TLS client authentication too.
9400 Not compiled unless enable-capieng specified to Configure.
9402 This work was sponsored by Logica.
9406 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9407 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9408 attribute creation routines such as certificate requests and PKCS#12
9413 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9415 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9416 handshake which could lead to a client crash as found using the
9417 Codenomicon TLS test suite ([CVE-2008-1672])
9419 *Steve Henson, Mark Cox*
9421 * Fix double free in TLS server name extensions which could lead to
9422 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9426 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9428 Clear the error queue to ensure that error entries left from
9429 older function calls do not interfere with the correct operation.
9431 *Lutz Jaenicke, Erik de Castro Lopo*
9433 * Remove root CA certificates of commercial CAs:
9435 The OpenSSL project does not recommend any specific CA and does not
9436 have any policy with respect to including or excluding any CA.
9437 Therefore it does not make any sense to ship an arbitrary selection
9438 of root CA certificates with the OpenSSL software.
9442 * RSA OAEP patches to fix two separate invalid memory reads.
9443 The first one involves inputs when 'lzero' is greater than
9444 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9445 before the beginning of from). The second one involves inputs where
9446 the 'db' section contains nothing but zeroes (there is a one-byte
9447 invalid read after the end of 'db').
9449 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9451 * Partial backport from 0.9.9-dev:
9453 Introduce bn_mul_mont (dedicated Montgomery multiplication
9454 procedure) as a candidate for BIGNUM assembler implementation.
9455 While 0.9.9-dev uses assembler for various architectures, only
9456 x86_64 is available by default here in the 0.9.8 branch, and
9457 32-bit x86 is available through a compile-time setting.
9459 To try the 32-bit x86 assembler implementation, use Configure
9460 option "enable-montasm" (which exists only for this backport).
9462 As "enable-montasm" for 32-bit x86 disclaims code stability
9463 anyway, in this constellation we activate additional code
9464 backported from 0.9.9-dev for further performance improvements,
9465 namely BN_from_montgomery_word. (To enable this otherwise,
9466 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9468 *Andy Polyakov (backport partially by Bodo Moeller)*
9470 * Add TLS session ticket callback. This allows an application to set
9471 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9472 values. This is useful for key rollover for example where several key
9473 sets may exist with different names.
9477 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9478 This was broken until now in 0.9.8 releases, such that the only way
9479 a registered ENGINE could be used (assuming it initialises
9480 successfully on the host) was to explicitly set it as the default
9481 for the relevant algorithms. This is in contradiction with 0.9.7
9482 behaviour and the documentation. With this fix, when an ENGINE is
9483 registered into a given algorithm's table of implementations, the
9484 'uptodate' flag is reset so that auto-discovery will be used next
9485 time a new context for that algorithm attempts to select an
9488 *Ian Lister (tweaked by Geoff Thorpe)*
9490 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9491 implementation in the following ways:
9493 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9496 Lack of BER streaming support means one pass streaming processing is
9497 only supported if data is detached: setting the streaming flag is
9498 ignored for embedded content.
9500 CMS support is disabled by default and must be explicitly enabled
9501 with the enable-cms configuration option.
9505 * Update the GMP engine glue to do direct copies between BIGNUM and
9506 mpz_t when openssl and GMP use the same limb size. Otherwise the
9507 existing "conversion via a text string export" trick is still used.
9509 *Paul Sheer <paulsheer@gmail.com>*
9511 * Zlib compression BIO. This is a filter BIO which compressed and
9512 uncompresses any data passed through it.
9516 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9517 RFC3394 compatible AES key wrapping.
9521 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9522 sets string data without copying. X509_ALGOR_set0() and
9523 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9524 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9525 from an X509_ATTRIBUTE structure optionally checking it occurs only
9526 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9531 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9532 to get the expected BN_FLG_CONSTTIME behavior.
9534 *Bodo Moeller (Google)*
9538 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9539 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9540 - added some more tests to do_tests.pl
9541 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9542 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9543 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9544 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9545 - various changes to netware.pl to enable gcc-cross builds on Win32
9547 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9548 - various changes to fix missing prototype warnings
9549 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9550 - added AES, WHIRLPOOL and CPUID assembler code to build files
9551 - added missing AES assembler make rules to mk1mf.pl
9552 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9554 *Guenter Knauf <eflash@gmx.net>*
9556 * Implement certificate status request TLS extension defined in RFC3546.
9557 A client can set the appropriate parameters and receive the encoded
9558 OCSP response via a callback. A server can query the supplied parameters
9559 and set the encoded OCSP response in the callback. Add simplified examples
9560 to s_client and s_server.
9564 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9567 + Binary incompatibility of ssl_ctx_st structure
9568 + DTLS interoperation with non-compliant servers
9569 + Don't call get_session_cb() without proposed session
9570 + Fix ia64 assembler code
9572 *Andy Polyakov, Steve Henson*
9574 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9576 * DTLS Handshake overhaul. There were longstanding issues with
9577 OpenSSL DTLS implementation, which were making it impossible for
9578 RFC 4347 compliant client to communicate with OpenSSL server.
9579 Unfortunately just fixing these incompatibilities would "cut off"
9580 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9581 server keeps tolerating non RFC compliant syntax. The opposite is
9582 not true, 0.9.8f client can not communicate with earlier server.
9583 This update even addresses CVE-2007-4995.
9587 * Changes to avoid need for function casts in OpenSSL: some compilers
9588 (gcc 4.2 and later) reject their use.
9589 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9592 * Add RFC4507 support to OpenSSL. This includes the corrections in
9593 RFC4507bis. The encrypted ticket format is an encrypted encoded
9594 SSL_SESSION structure, that way new session features are automatically
9597 If a client application caches session in an SSL_SESSION structure
9598 support is transparent because tickets are now stored in the encoded
9601 The SSL_CTX structure automatically generates keys for ticket
9602 protection in servers so again support should be possible
9603 with no application modification.
9605 If a client or server wishes to disable RFC4507 support then the option
9606 SSL_OP_NO_TICKET can be set.
9608 Add a TLS extension debugging callback to allow the contents of any client
9609 or server extensions to be examined.
9611 This work was sponsored by Google.
9615 * Add initial support for TLS extensions, specifically for the server_name
9616 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9617 have new members for a host name. The SSL data structure has an
9618 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9619 stored in that context to allow for session resumption, even after the
9620 SSL has been switched to a new SSL_CTX in reaction to a client's
9621 server_name extension.
9623 New functions (subject to change):
9625 SSL_get_servername()
9626 SSL_get_servername_type()
9629 New CTRL codes and macros (subject to change):
9631 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9632 - SSL_CTX_set_tlsext_servername_callback()
9633 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9634 - SSL_CTX_set_tlsext_servername_arg()
9635 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9637 openssl s_client has a new '-servername ...' option.
9639 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9640 '-key2 ...', '-servername_fatal' (subject to change). This allows
9641 testing the HostName extension for a specific single host name ('-cert'
9642 and '-key' remain fallbacks for handshakes without HostName
9643 negotiation). If the unrecognized_name alert has to be sent, this by
9644 default is a warning; it becomes fatal with the '-servername_fatal'
9647 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9649 * Add AES and SSE2 assembly language support to VC++ build.
9653 * Mitigate attack on final subtraction in Montgomery reduction.
9657 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9658 (which previously caused an internal error).
9662 * Squeeze another 10% out of IGE mode when in != out.
9666 * AES IGE mode speedup.
9668 *Dean Gaudet (Google)*
9670 * Add the Korean symmetric 128-bit cipher SEED (see
9671 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9672 add SEED ciphersuites from RFC 4162:
9674 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9675 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9676 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9677 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9679 To minimize changes between patchlevels in the OpenSSL 0.9.8
9680 series, SEED remains excluded from compilation unless OpenSSL
9681 is configured with 'enable-seed'.
9683 *KISA, Bodo Moeller*
9685 * Mitigate branch prediction attacks, which can be practical if a
9686 single processor is shared, allowing a spy process to extract
9687 information. For detailed background information, see
9688 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9689 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9690 and Necessary Software Countermeasures"). The core of the change
9691 are new versions BN_div_no_branch() and
9692 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9693 respectively, which are slower, but avoid the security-relevant
9694 conditional branches. These are automatically called by BN_div()
9695 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9696 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9697 remove a conditional branch.
9699 BN_FLG_CONSTTIME is the new name for the previous
9700 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9701 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9702 in the exponent causes BN_mod_exp_mont() to use the alternative
9703 implementation in BN_mod_exp_mont_consttime().) The old name
9704 remains as a deprecated alias.
9706 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9707 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9708 constant-time implementations for more than just exponentiation.
9709 Here too the old name is kept as a deprecated alias.
9711 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9712 the BN_BLINDING structure gets an independent copy of the
9713 modulus. This means that the previous `BIGNUM *m` argument to
9714 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9715 essentially becomes `const BIGNUM *m`, although we can't actually
9716 change this in the header file before 0.9.9. It allows
9717 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9718 enable BN_FLG_CONSTTIME.
9720 *Matthew D Wood (Intel Corp)*
9722 * In the SSL/TLS server implementation, be strict about session ID
9723 context matching (which matters if an application uses a single
9724 external cache for different purposes). Previously,
9725 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9726 set. This did ensure strict client verification, but meant that,
9727 with applications using a single external cache for quite
9728 different requirements, clients could circumvent ciphersuite
9729 restrictions for a given session ID context by starting a session
9730 in a different context.
9734 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9735 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9736 authentication-only ciphersuites.
9740 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9741 not complete and could lead to a possible single byte overflow
9742 ([CVE-2007-5135]) [Ben Laurie]
9744 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9746 * Since AES128 and AES256 (and similarly Camellia128 and
9747 Camellia256) share a single mask bit in the logic of
9748 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9749 kludge to work properly if AES128 is available and AES256 isn't
9750 (or if Camellia128 is available and Camellia256 isn't).
9754 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9755 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9756 When a point or a seed is encoded in a BIT STRING, we need to
9757 prevent the removal of trailing zero bits to get the proper DER
9758 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9759 of a NamedBitList, for which trailing 0 bits need to be removed.)
9763 * Have SSL/TLS server implementation tolerate "mismatched" record
9764 protocol version while receiving ClientHello even if the
9765 ClientHello is fragmented. (The server can't insist on the
9766 particular protocol version it has chosen before the ServerHello
9767 message has informed the client about his choice.)
9771 * Add RFC 3779 support.
9773 *Rob Austein for ARIN, Ben Laurie*
9775 * Load error codes if they are not already present instead of using a
9776 static variable. This allows them to be cleanly unloaded and reloaded.
9777 Improve header file function name parsing.
9781 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9782 or CAPABILITY handshake as required by RFCs.
9786 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9788 * Introduce limits to prevent malicious keys being able to
9789 cause a denial of service. ([CVE-2006-2940])
9791 *Steve Henson, Bodo Moeller*
9793 * Fix ASN.1 parsing of certain invalid structures that can result
9794 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9796 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9797 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9799 * Fix SSL client code which could crash if connecting to a
9800 malicious SSLv2 server. ([CVE-2006-4343])
9802 *Tavis Ormandy and Will Drewry, Google Security Team*
9804 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9805 match only those. Before that, "AES256-SHA" would be interpreted
9806 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9807 the same strength classification in 0.9.7h) as we currently only
9808 have a single AES bit in the ciphersuite description bitmap.
9809 That change, however, also applied to ciphersuite strings such as
9810 "RC4-MD5" that intentionally matched multiple ciphersuites --
9811 namely, SSL 2.0 ciphersuites in addition to the more common ones
9812 from SSL 3.0/TLS 1.0.
9814 So we change the selection algorithm again: Naming an explicit
9815 ciphersuite selects this one ciphersuite, and any other similar
9816 ciphersuite (same bitmap) from *other* protocol versions.
9817 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9818 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9820 Since SSL 2.0 does not have any ciphersuites for which the
9821 128/256 bit distinction would be relevant, this works for now.
9822 The proper fix will be to use different bits for AES128 and
9823 AES256, which would have avoided the problems from the beginning;
9824 however, bits are scarce, so we can only do this in a new release
9825 (not just a patchlevel) when we can change the SSL_CIPHER
9826 definition to split the single 'unsigned long mask' bitmap into
9827 multiple values to extend the available space.
9831 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9833 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9834 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9836 * Add AES IGE and biIGE modes.
9840 * Change the Unix randomness entropy gathering to use poll() when
9841 possible instead of select(), since the latter has some
9842 undesirable limitations.
9844 *Darryl Miles via Richard Levitte and Bodo Moeller*
9846 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9847 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9848 cannot be implicitly activated as part of, e.g., the "AES" alias.
9849 However, please upgrade to OpenSSL 0.9.9[-dev] for
9850 non-experimental use of the ECC ciphersuites to get TLS extension
9851 support, which is required for curve and point format negotiation
9852 to avoid potential handshake problems.
9856 * Disable rogue ciphersuites:
9858 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9859 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9860 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9862 The latter two were purportedly from
9863 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9866 Also deactivate the remaining ciphersuites from
9867 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9868 unofficial, and the ID has long expired.
9872 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9873 dual-core machines) and other potential thread-safety issues.
9877 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9878 versions), which is now available for royalty-free use
9879 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9880 Also, add Camellia TLS ciphersuites from RFC 4132.
9882 To minimize changes between patchlevels in the OpenSSL 0.9.8
9883 series, Camellia remains excluded from compilation unless OpenSSL
9884 is configured with 'enable-camellia'.
9888 * Disable the padding bug check when compression is in use. The padding
9889 bug check assumes the first packet is of even length, this is not
9890 necessarily true if compression is enabled and can result in false
9891 positives causing handshake failure. The actual bug test is ancient
9892 code so it is hoped that implementations will either have fixed it by
9893 now or any which still have the bug do not support compression.
9897 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9899 * When applying a cipher rule check to see if string match is an explicit
9900 cipher suite and only match that one cipher suite if it is.
9904 * Link in manifests for VC++ if needed.
9906 *Austin Ziegler <halostatue@gmail.com>*
9908 * Update support for ECC-based TLS ciphersuites according to
9909 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9910 TLS extensions, which are supported starting with the 0.9.9
9911 branch, not in the OpenSSL 0.9.8 branch).
9915 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9916 opaque EVP_CIPHER_CTX handling.
9920 * Fixes and enhancements to zlib compression code. We now only use
9921 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9922 to conform with the standards mentioned here:
9923 <http://www.zlib.net/DLL_FAQ.txt>
9924 Static zlib linking now works on Windows and the new --with-zlib-include
9925 --with-zlib-lib options to Configure can be used to supply the location
9926 of the headers and library. Gracefully handle case where zlib library
9931 * Several fixes and enhancements to the OID generation code. The old code
9932 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9933 handle numbers larger than ULONG_MAX, truncated printing and had a
9934 non standard OBJ_obj2txt() behaviour.
9938 * Add support for building of engines under engine/ as shared libraries
9939 under VC++ build system.
9943 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9944 Hopefully, we will not see any false combination of paths any more.
9948 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9950 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9951 (part of SSL_OP_ALL). This option used to disable the
9952 countermeasure against man-in-the-middle protocol-version
9953 rollback in the SSL 2.0 server implementation, which is a bad
9954 idea. ([CVE-2005-2969])
9956 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9957 for Information Security, National Institute of Advanced Industrial
9958 Science and Technology [AIST], Japan)*
9960 * Add two function to clear and return the verify parameter flags.
9964 * Keep cipherlists sorted in the source instead of sorting them at
9965 runtime, thus removing the need for a lock.
9969 * Avoid some small subgroup attacks in Diffie-Hellman.
9971 *Nick Mathewson and Ben Laurie*
9973 * Add functions for well-known primes.
9977 * Extended Windows CE support.
9979 *Satoshi Nakamura and Andy Polyakov*
9981 * Initialize SSL_METHOD structures at compile time instead of during
9982 runtime, thus removing the need for a lock.
9986 * Make PKCS7_decrypt() work even if no certificate is supplied by
9987 attempting to decrypt each encrypted key in turn. Add support to
9992 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9994 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9997 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10001 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10002 key into the same file any more.
10006 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10010 * Add -utf8 command line and config file option to 'ca'.
10012 *Stefan <stf@udoma.org*
10014 * Removed the macro des_crypt(), as it seems to conflict with some
10015 libraries. Use DES_crypt().
10019 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10020 involves renaming the source and generated shared-libs for
10021 both. The engines will accept the corrected or legacy ids
10022 ('ncipher' and '4758_cca' respectively) when binding. NB,
10023 this only applies when building 'shared'.
10025 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10027 * Add attribute functions to EVP_PKEY structure. Modify
10028 PKCS12_create() to recognize a CSP name attribute and
10029 use it. Make -CSP option work again in pkcs12 utility.
10033 * Add new functionality to the bn blinding code:
10034 - automatic re-creation of the BN_BLINDING parameters after
10035 a fixed number of uses (currently 32)
10036 - add new function for parameter creation
10037 - introduce flags to control the update behaviour of the
10038 BN_BLINDING parameters
10039 - hide BN_BLINDING structure
10040 Add a second BN_BLINDING slot to the RSA structure to improve
10041 performance when a single RSA object is shared among several
10046 * Add support for DTLS.
10048 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10050 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10051 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10055 * Remove buggy and incomplete DH cert support from
10056 ssl/ssl_rsa.c and ssl/s3_both.c
10060 * Use SHA-1 instead of MD5 as the default digest algorithm for
10061 the `apps/openssl` commands.
10065 * Compile clean with "-Wall -Wmissing-prototypes
10066 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10067 DEBUG_SAFESTACK must also be set.
10071 * Change ./Configure so that certain algorithms can be disabled by default.
10072 The new counterpiece to "no-xxx" is "enable-xxx".
10074 The patented RC5 and MDC2 algorithms will now be disabled unless
10075 "enable-rc5" and "enable-mdc2", respectively, are specified.
10077 (IDEA remains enabled despite being patented. This is because IDEA
10078 is frequently required for interoperability, and there is no license
10079 fee for non-commercial use. As before, "no-idea" can be used to
10080 avoid this algorithm.)
10084 * Add processing of proxy certificates (see RFC 3820). This work was
10085 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10086 EGEE (Enabling Grids for E-science in Europe).
10090 * RC4 performance overhaul on modern architectures/implementations, such
10091 as Intel P4, IA-64 and AMD64.
10095 * New utility extract-section.pl. This can be used specify an alternative
10096 section number in a pod file instead of having to treat each file as
10097 a separate case in Makefile. This can be done by adding two lines to the
10100 =for comment openssl_section:XXX
10102 The blank line is mandatory.
10106 * New arguments -certform, -keyform and -pass for s_client and s_server
10107 to allow alternative format key and certificate files and passphrase
10112 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10113 update associated structures and add various utility functions.
10115 Add new policy related verify parameters, include policy checking in
10116 standard verify code. Enhance 'smime' application with extra parameters
10117 to support policy checking and print out.
10121 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10122 Nehemiah processors. These extensions support AES encryption in hardware
10123 as well as RNG (though RNG support is currently disabled).
10125 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10127 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10131 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10133 *Andy Polyakov and a number of other people*
10135 * Improved PowerPC platform support. Most notably BIGNUM assembler
10136 implementation contributed by IBM.
10138 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10140 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10141 exponent rather than 'unsigned long'. There is a corresponding change to
10142 the new 'rsa_keygen' element of the RSA_METHOD structure.
10144 *Jelte Jansen, Geoff Thorpe*
10146 * Functionality for creating the initial serial number file is now
10147 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10149 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10150 number file to 1, which is bound to cause problems. To avoid
10151 the problems while respecting compatibility between different 0.9.7
10152 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10153 CA.pl for serial number initialization. With the new release 0.9.8,
10154 we can fix the problem directly in the 'ca' utility.)
10158 * Reduced header interdependencies by declaring more opaque objects in
10159 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10160 give fewer recursive includes, which could break lazy source code - so
10161 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10162 developers should define this symbol when building and using openssl to
10163 ensure they track the recommended behaviour, interfaces, [etc], but
10164 backwards-compatible behaviour prevails when this isn't defined.
10168 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10172 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10173 This will generate a random key of the appropriate length based on the
10174 cipher context. The EVP_CIPHER can provide its own random key generation
10175 routine to support keys of a specific form. This is used in the des and
10176 3des routines to generate a key of the correct parity. Update S/MIME
10177 code to use new functions and hence generate correct parity DES keys.
10178 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10179 valid (weak or incorrect parity).
10183 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10184 as looking them up. This is useful when the verified structure may contain
10185 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10186 present unless the new PKCS7_NO_CRL flag is asserted.
10190 * Extend ASN1 oid configuration module. It now additionally accepts the
10193 shortName = some long name, 1.2.3.4
10197 * Reimplemented the BN_CTX implementation. There is now no more static
10198 limitation on the number of variables it can handle nor the depth of the
10199 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10200 information can now expand as required, and rather than having a single
10201 static array of bignums, BN_CTX now uses a linked-list of such arrays
10202 allowing it to expand on demand whilst maintaining the usefulness of
10203 BN_CTX's "bundling".
10207 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10208 to allow all RSA operations to function using a single BN_CTX.
10212 * Preliminary support for certificate policy evaluation and checking. This
10213 is initially intended to pass the tests outlined in "Conformance Testing
10214 of Relying Party Client Certificate Path Processing Logic" v1.07.
10218 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10219 remained unused and not that useful. A variety of other little bignum
10220 tweaks and fixes have also been made continuing on from the audit (see
10225 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10226 associated ASN1, EVP and SSL functions and old ASN1 macros.
10230 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10231 and this should never fail. So the return value from the use of
10232 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10233 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10237 * BN_CTX_get() should return zero-valued bignums, providing the same
10238 initialised value as BN_new().
10240 *Geoff Thorpe, suggested by Ulf Möller*
10242 * Support for inhibitAnyPolicy certificate extension.
10246 * An audit of the BIGNUM code is underway, for which debugging code is
10247 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10248 is considered valid when processing BIGNUMs, and causes execution to
10249 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10250 further steps are taken to deliberately pollute unused data in BIGNUM
10251 structures to try and expose faulty code further on. For now, openssl will
10252 (in its default mode of operation) continue to tolerate the inconsistent
10253 forms that it has tolerated in the past, but authors and packagers should
10254 consider trying openssl and their own applications when compiled with
10255 these debugging symbols defined. It will help highlight potential bugs in
10256 their own code, and will improve the test coverage for OpenSSL itself. At
10257 some point, these tighter rules will become openssl's default to improve
10258 maintainability, though the assert()s and other overheads will remain only
10259 in debugging configurations. See bn.h for more details.
10261 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10263 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10264 that can only be obtained through BN_CTX_new() (which implicitly
10265 initialises it). The presence of this function only made it possible
10266 to overwrite an existing structure (and cause memory leaks).
10270 * Because of the callback-based approach for implementing LHASH as a
10271 template type, lh_insert() adds opaque objects to hash-tables and
10272 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10273 to clean up those corresponding objects before destroying the hash table
10274 (and losing the object pointers). So some over-zealous constifications in
10275 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10276 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10277 prototyped to have "const" restrictions on the object pointers they are
10278 given (and so aren't required to cast them away any more).
10282 * The tmdiff.h API was so ugly and minimal that our own timing utility
10283 (speed) prefers to use its own implementation. The two implementations
10284 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10285 its object type properly exposed (MS_TM) instead of casting to/from
10286 `char *`. This may still change yet if someone realises MS_TM and
10288 aren't necessarily the greatest nomenclatures - but this is what was used
10289 internally to the implementation so I've used that for now.
10293 * Ensure that deprecated functions do not get compiled when
10294 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10295 the self-tests were still using deprecated key-generation functions so
10296 these have been updated also.
10300 * Reorganise PKCS#7 code to separate the digest location functionality
10301 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10302 New function PKCS7_set_digest() to set the digest type for PKCS#7
10303 digestedData type. Add additional code to correctly generate the
10304 digestedData type and add support for this type in PKCS7 initialization
10309 * New function PKCS7_set0_type_other() this initializes a PKCS7
10310 structure of type "other".
10314 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10315 sure the loop does correctly stop and breaking ("division by zero")
10316 modulus operations are not performed. The (pre-generated) prime
10317 table crypto/bn/bn_prime.h was already correct, but it could not be
10318 re-generated on some platforms because of the "division by zero"
10319 situation in the script.
10321 *Ralf S. Engelschall*
10323 * Update support for ECC-based TLS ciphersuites according to
10324 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10325 SHA-1 now is only used for "small" curves (where the
10326 representation of a field element takes up to 24 bytes); for
10327 larger curves, the field element resulting from ECDH is directly
10328 used as premaster secret.
10330 *Douglas Stebila (Sun Microsystems Laboratories)*
10332 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10333 curve secp160r1 to the tests.
10335 *Douglas Stebila (Sun Microsystems Laboratories)*
10337 * Add the possibility to load symbols globally with DSO.
10339 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10341 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10342 control of the error stack.
10346 * Add support for STORE in ENGINE.
10350 * Add the STORE type. The intention is to provide a common interface
10351 to certificate and key stores, be they simple file-based stores, or
10352 HSM-type store, or LDAP stores, or...
10353 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10357 * Add a generic structure called OPENSSL_ITEM. This can be used to
10358 pass a list of arguments to any function as well as provide a way
10359 for a function to pass data back to the caller.
10363 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10364 works like BUF_strdup() but can be used to duplicate a portion of
10365 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10370 * Add the function sk_find_ex() which works like sk_find(), but will
10371 return an index to an element even if an exact match couldn't be
10372 found. The index is guaranteed to point at the element where the
10373 searched-for key would be inserted to preserve sorting order.
10377 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10378 takes an extra flags argument for optional functionality. Currently,
10379 the following flags are defined:
10381 OBJ_BSEARCH_VALUE_ON_NOMATCH
10382 This one gets OBJ_bsearch_ex() to return a pointer to the first
10383 element where the comparing function returns a negative or zero
10386 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10387 This one gets OBJ_bsearch_ex() to return a pointer to the first
10388 element where the comparing function returns zero. This is useful
10389 if there are more than one element where the comparing function
10394 * Make it possible to create self-signed certificates with 'openssl ca'
10395 in such a way that the self-signed certificate becomes part of the
10396 CA database and uses the same mechanisms for serial number generation
10397 as all other certificate signing. The new flag '-selfsign' enables
10398 this functionality. Adapt CA.sh and CA.pl.in.
10402 * Add functionality to check the public key of a certificate request
10403 against a given private. This is useful to check that a certificate
10404 request can be signed by that key (self-signing).
10408 * Make it possible to have multiple active certificates with the same
10409 subject in the CA index file. This is done only if the keyword
10410 'unique_subject' is set to 'no' in the main CA section (default
10411 if 'CA_default') of the configuration file. The value is saved
10412 with the database itself in a separate index attribute file,
10413 named like the index file with '.attr' appended to the name.
10417 * Generate multi-valued AVAs using '+' notation in config files for
10422 * Support for nameConstraints certificate extension.
10426 * Support for policyConstraints certificate extension.
10430 * Support for policyMappings certificate extension.
10434 * Make sure the default DSA_METHOD implementation only uses its
10435 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10436 and change its own handlers to be NULL so as to remove unnecessary
10437 indirection. This lets alternative implementations fallback to the
10438 default implementation more easily.
10442 * Support for directoryName in GeneralName related extensions
10447 * Make it possible to link applications using Makefile.shared.
10448 Make that possible even when linking against static libraries!
10452 * Support for single pass processing for S/MIME signing. This now
10453 means that S/MIME signing can be done from a pipe, in addition
10454 cleartext signing (multipart/signed type) is effectively streaming
10455 and the signed data does not need to be all held in memory.
10457 This is done with a new flag PKCS7_STREAM. When this flag is set
10458 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10459 is done after the data is output (and digests calculated) in
10460 SMIME_write_PKCS7().
10464 * Add full support for -rpath/-R, both in shared libraries and
10465 applications, at least on the platforms where it's known how
10470 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10471 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10472 will now compute a table of multiples of the generator that
10473 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10474 faster (notably in the case of a single point multiplication,
10475 scalar * generator).
10477 *Nils Larsch, Bodo Moeller*
10479 * IPv6 support for certificate extensions. The various extensions
10480 which use the IP:a.b.c.d can now take IPv6 addresses using the
10481 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10486 * Added an ENGINE that implements RSA by performing private key
10487 exponentiations with the GMP library. The conversions to and from
10488 GMP's mpz_t format aren't optimised nor are any montgomery forms
10489 cached, and on x86 it appears OpenSSL's own performance has caught up.
10490 However there are likely to be other architectures where GMP could
10491 provide a boost. This ENGINE is not built in by default, but it can be
10492 specified at Configure time and should be accompanied by the necessary
10493 linker additions, eg;
10494 ./config -DOPENSSL_USE_GMP -lgmp
10498 * "openssl engine" will not display ENGINE/DSO load failure errors when
10499 testing availability of engines with "-t" - the old behaviour is
10500 produced by increasing the feature's verbosity with "-tt".
10504 * ECDSA routines: under certain error conditions uninitialized BN objects
10505 could be freed. Solution: make sure initialization is performed early
10506 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10511 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10512 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10513 software implementations. For DSA and DH, parameter generation can
10514 also be overridden by providing the appropriate method callbacks.
10518 * Change the "progress" mechanism used in key-generation and
10519 primality testing to functions that take a new BN_GENCB pointer in
10520 place of callback/argument pairs. The new API functions have `_ex`
10521 postfixes and the older functions are reimplemented as wrappers for
10522 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10523 declarations of the old functions to help (graceful) attempts to
10524 migrate to the new functions. Also, the new key-generation API
10525 functions operate on a caller-supplied key-structure and return
10526 success/failure rather than returning a key or NULL - this is to
10527 help make "keygen" another member function of RSA_METHOD etc.
10529 Example for using the new callback interface:
10531 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10532 void *my_arg = ...;
10535 BN_GENCB_set(&my_cb, my_callback, my_arg);
10537 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10538 /* For the meaning of a, b in calls to my_callback(), see the
10539 * documentation of the function that calls the callback.
10540 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10541 * my_callback should return 1 if it wants BN_is_prime_ex()
10542 * to continue, or 0 to stop.
10547 * Change the ZLIB compression method to be stateful, and make it
10548 available to TLS with the number defined in
10549 draft-ietf-tls-compression-04.txt.
10553 * Add the ASN.1 structures and functions for CertificatePair, which
10554 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10556 CertificatePair ::= SEQUENCE {
10557 forward [0] Certificate OPTIONAL,
10558 reverse [1] Certificate OPTIONAL,
10559 -- at least one of the pair shall be present -- }
10561 Also implement the PEM functions to read and write certificate
10562 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10564 This needed to be defined, mostly for the sake of the LDAP
10565 attribute crossCertificatePair, but may prove useful elsewhere as
10570 * Make it possible to inhibit symlinking of shared libraries in
10571 Makefile.shared, for Cygwin's sake.
10575 * Extend the BIGNUM API by creating a function
10576 void BN_set_negative(BIGNUM *a, int neg);
10577 and a macro that behave like
10578 int BN_is_negative(const BIGNUM *a);
10580 to avoid the need to access 'a->neg' directly in applications.
10584 * Implement fast modular reduction for pseudo-Mersenne primes
10585 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10586 EC_GROUP_new_curve_GFp() will now automatically use this
10589 *Nils Larsch <nla@trustcenter.de>*
10591 * Add new lock type (CRYPTO_LOCK_BN).
10595 * Change the ENGINE framework to automatically load engines
10596 dynamically from specific directories unless they could be
10597 found to already be built in or loaded. Move all the
10598 current engines except for the cryptodev one to a new
10599 directory engines/.
10600 The engines in engines/ are built as shared libraries if
10601 the "shared" options was given to ./Configure or ./config.
10602 Otherwise, they are inserted in libcrypto.a.
10603 /usr/local/ssl/engines is the default directory for dynamic
10604 engines, but that can be overridden at configure time through
10605 the usual use of --prefix and/or --openssldir, and at run
10606 time with the environment variable OPENSSL_ENGINES.
10608 *Geoff Thorpe and Richard Levitte*
10610 * Add Makefile.shared, a helper makefile to build shared
10611 libraries. Adapt Makefile.org.
10615 * Add version info to Win32 DLLs.
10617 *Peter 'Luna' Runestig" <peter@runestig.com>*
10619 * Add new 'medium level' PKCS#12 API. Certificates and keys
10620 can be added using this API to created arbitrary PKCS#12
10621 files while avoiding the low level API.
10623 New options to PKCS12_create(), key or cert can be NULL and
10624 will then be omitted from the output file. The encryption
10625 algorithm NIDs can be set to -1 for no encryption, the mac
10626 iteration count can be set to 0 to omit the mac.
10628 Enhance pkcs12 utility by making the -nokeys and -nocerts
10629 options work when creating a PKCS#12 file. New option -nomac
10630 to omit the mac, NONE can be set for an encryption algorithm.
10631 New code is modified to use the enhanced PKCS12_create()
10632 instead of the low level API.
10636 * Extend ASN1 encoder to support indefinite length constructed
10637 encoding. This can output sequences tags and octet strings in
10638 this form. Modify pk7_asn1.c to support indefinite length
10639 encoding. This is experimental and needs additional code to
10640 be useful, such as an ASN1 bio and some enhanced streaming
10643 Extend template encode functionality so that tagging is passed
10644 down to the template encoder.
10648 * Let 'openssl req' fail if an argument to '-newkey' is not
10649 recognized instead of using RSA as a default.
10653 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10654 As these are not official, they are not included in "ALL";
10655 the "ECCdraft" ciphersuite group alias can be used to select them.
10657 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10659 * Add ECDH engine support.
10661 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10663 * Add ECDH in new directory crypto/ecdh/.
10665 *Douglas Stebila (Sun Microsystems Laboratories)*
10667 * Let BN_rand_range() abort with an error after 100 iterations
10668 without success (which indicates a broken PRNG).
10672 * Change BN_mod_sqrt() so that it verifies that the input value
10673 is really the square of the return value. (Previously,
10674 BN_mod_sqrt would show GIGO behaviour.)
10678 * Add named elliptic curves over binary fields from X9.62, SECG,
10679 and WAP/WTLS; add OIDs that were still missing.
10681 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10683 * Extend the EC library for elliptic curves over binary fields
10684 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10687 EC_GF2m_simple_method
10691 EC_GROUP_new_curve_GF2m
10692 EC_GROUP_set_curve_GF2m
10693 EC_GROUP_get_curve_GF2m
10694 EC_POINT_set_affine_coordinates_GF2m
10695 EC_POINT_get_affine_coordinates_GF2m
10696 EC_POINT_set_compressed_coordinates_GF2m
10698 Point compression for binary fields is disabled by default for
10699 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10702 As binary polynomials are represented as BIGNUMs, various members
10703 of the EC_GROUP and EC_POINT data structures can be shared
10704 between the implementations for prime fields and binary fields;
10705 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10706 are essentially identical to their `..._GFp` counterparts.
10707 (For simplicity, the `..._GFp` prefix has been dropped from
10708 various internal method names.)
10710 An internal 'field_div' method (similar to 'field_mul' and
10711 'field_sqr') has been added; this is used only for binary fields.
10713 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10715 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10716 through methods ('mul', 'precompute_mult').
10718 The generic implementations (now internally called 'ec_wNAF_mul'
10719 and 'ec_wNAF_precomputed_mult') remain the default if these
10720 methods are undefined.
10722 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10724 * New function EC_GROUP_get_degree, which is defined through
10725 EC_METHOD. For curves over prime fields, this returns the bit
10726 length of the modulus.
10728 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10730 * New functions EC_GROUP_dup, EC_POINT_dup.
10731 (These simply call ..._new and ..._copy).
10733 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10735 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10736 Polynomials are represented as BIGNUMs (where the sign bit is not
10737 used) in the following functions [macros]:
10740 BN_GF2m_sub [= BN_GF2m_add]
10741 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10742 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10743 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10745 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10746 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10747 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10748 BN_GF2m_cmp [= BN_ucmp]
10750 (Note that only the 'mod' functions are actually for fields GF(2^m).
10751 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10753 For some functions, an the irreducible polynomial defining a
10754 field can be given as an 'unsigned int[]' with strictly
10755 decreasing elements giving the indices of those bits that are set;
10756 i.e., p[] represents the polynomial
10757 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10759 p[0] > p[1] > ... > p[k] = 0.
10760 This applies to the following functions:
10763 BN_GF2m_mod_mul_arr
10764 BN_GF2m_mod_sqr_arr
10765 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10766 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10767 BN_GF2m_mod_exp_arr
10768 BN_GF2m_mod_sqrt_arr
10769 BN_GF2m_mod_solve_quad_arr
10773 Conversion can be performed by the following functions:
10778 bntest.c has additional tests for binary polynomial arithmetic.
10780 Two implementations for BN_GF2m_mod_div() are available.
10781 The default algorithm simply uses BN_GF2m_mod_inv() and
10782 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10783 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10784 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10786 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10788 * Add new error code 'ERR_R_DISABLED' that can be used when some
10789 functionality is disabled at compile-time.
10791 *Douglas Stebila <douglas.stebila@sun.com>*
10793 * Change default behaviour of 'openssl asn1parse' so that more
10794 information is visible when viewing, e.g., a certificate:
10796 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10797 mode the content of non-printable OCTET STRINGs is output in a
10798 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10799 avoid the appearance of a printable string.
10801 *Nils Larsch <nla@trustcenter.de>*
10803 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10805 EC_GROUP_set_asn1_flag()
10806 EC_GROUP_get_asn1_flag()
10807 EC_GROUP_set_point_conversion_form()
10808 EC_GROUP_get_point_conversion_form()
10809 These control ASN1 encoding details:
10810 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10811 has been set to OPENSSL_EC_NAMED_CURVE.
10812 - Points are encoded in uncompressed form by default; options for
10813 asn1_for are as for point2oct, namely
10814 POINT_CONVERSION_COMPRESSED
10815 POINT_CONVERSION_UNCOMPRESSED
10816 POINT_CONVERSION_HYBRID
10818 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10820 EC_GROUP_set_seed()
10821 EC_GROUP_get0_seed()
10822 EC_GROUP_get_seed_len()
10823 This is used only for ASN1 purposes (so far).
10825 *Nils Larsch <nla@trustcenter.de>*
10827 * Add 'field_type' member to EC_METHOD, which holds the NID
10828 of the appropriate field type OID. The new function
10829 EC_METHOD_get_field_type() returns this value.
10831 *Nils Larsch <nla@trustcenter.de>*
10834 EC_POINT_point2bn()
10835 EC_POINT_bn2point()
10836 EC_POINT_point2hex()
10837 EC_POINT_hex2point()
10838 providing useful interfaces to EC_POINT_point2oct() and
10839 EC_POINT_oct2point().
10841 *Nils Larsch <nla@trustcenter.de>*
10843 * Change internals of the EC library so that the functions
10844 EC_GROUP_set_generator()
10845 EC_GROUP_get_generator()
10846 EC_GROUP_get_order()
10847 EC_GROUP_get_cofactor()
10848 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10849 to methods, which would lead to unnecessary code duplication when
10850 adding different types of curves.
10852 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10854 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10855 arithmetic, and such that modified wNAFs are generated
10856 (which avoid length expansion in many cases).
10860 * Add a function EC_GROUP_check_discriminant() (defined via
10861 EC_METHOD) that verifies that the curve discriminant is non-zero.
10863 Add a function EC_GROUP_check() that makes some sanity tests
10864 on a EC_GROUP, its generator and order. This includes
10865 EC_GROUP_check_discriminant().
10867 *Nils Larsch <nla@trustcenter.de>*
10869 * Add ECDSA in new directory crypto/ecdsa/.
10871 Add applications 'openssl ecparam' and 'openssl ecdsa'
10872 (these are based on 'openssl dsaparam' and 'openssl dsa').
10874 ECDSA support is also included in various other files across the
10875 library. Most notably,
10876 - 'openssl req' now has a '-newkey ecdsa:file' option;
10877 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10878 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10879 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10880 them suitable for ECDSA where domain parameters must be
10881 extracted before the specific public key;
10882 - ECDSA engine support has been added.
10884 *Nils Larsch <nla@trustcenter.de>*
10886 * Include some named elliptic curves, and add OIDs from X9.62,
10887 SECG, and WAP/WTLS. Each curve can be obtained from the new
10889 EC_GROUP_new_by_curve_name(),
10890 and the list of available named curves can be obtained with
10891 EC_get_builtin_curves().
10892 Also add a 'curve_name' member to EC_GROUP objects, which can be
10894 EC_GROUP_set_curve_name()
10895 EC_GROUP_get_curve_name()
10897 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10899 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10900 was actually never needed) and in BN_mul(). The removal in BN_mul()
10901 required a small change in bn_mul_part_recursive() and the addition
10902 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10903 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10904 bn_sub_words() and bn_add_words() except they take arrays with
10909 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10911 * Cleanse PEM buffers before freeing them since they may contain
10914 *Benjamin Bennett <ben@psc.edu>*
10916 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10917 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10918 authentication-only ciphersuites.
10922 * Since AES128 and AES256 share a single mask bit in the logic of
10923 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10924 kludge to work properly if AES128 is available and AES256 isn't.
10928 * Expand security boundary to match 1.1.1 module.
10932 * Remove redundant features: hash file source, editing of test vectors
10933 modify fipsld to use external fips_premain.c signature.
10937 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10938 run algorithm test programs.
10942 * Make algorithm test programs more tolerant of whitespace.
10946 * Have SSL/TLS server implementation tolerate "mismatched" record
10947 protocol version while receiving ClientHello even if the
10948 ClientHello is fragmented. (The server can't insist on the
10949 particular protocol version it has chosen before the ServerHello
10950 message has informed the client about his choice.)
10954 * Load error codes if they are not already present instead of using a
10955 static variable. This allows them to be cleanly unloaded and reloaded.
10959 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10961 * Introduce limits to prevent malicious keys being able to
10962 cause a denial of service. ([CVE-2006-2940])
10964 *Steve Henson, Bodo Moeller*
10966 * Fix ASN.1 parsing of certain invalid structures that can result
10967 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10969 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10970 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10972 * Fix SSL client code which could crash if connecting to a
10973 malicious SSLv2 server. ([CVE-2006-4343])
10975 *Tavis Ormandy and Will Drewry, Google Security Team*
10977 * Change ciphersuite string processing so that an explicit
10978 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10979 will no longer include "AES128-SHA"), and any other similar
10980 ciphersuite (same bitmap) from *other* protocol versions (so that
10981 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10982 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10983 changes from 0.9.8b and 0.9.8d.
10987 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10989 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10990 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10992 * Change the Unix randomness entropy gathering to use poll() when
10993 possible instead of select(), since the latter has some
10994 undesirable limitations.
10996 *Darryl Miles via Richard Levitte and Bodo Moeller*
10998 * Disable rogue ciphersuites:
11000 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11001 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11002 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11004 The latter two were purportedly from
11005 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11008 Also deactivate the remaining ciphersuites from
11009 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11010 unofficial, and the ID has long expired.
11014 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11015 dual-core machines) and other potential thread-safety issues.
11019 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11021 * Adapt fipsld and the build system to link against the validated FIPS
11022 module in FIPS mode.
11026 * Fixes for VC++ 2005 build under Windows.
11030 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11031 from a Windows bash shell such as MSYS. It is autodetected from the
11032 "config" script when run from a VC++ environment. Modify standard VC++
11033 build to use fipscanister.o from the GNU make build.
11037 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11039 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11040 The value now differs depending on if you build for FIPS or not.
11041 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11042 safely run with a non-FIPSed libcrypto, as it may crash because of
11043 the difference induced by this change.
11047 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11049 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11050 (part of SSL_OP_ALL). This option used to disable the
11051 countermeasure against man-in-the-middle protocol-version
11052 rollback in the SSL 2.0 server implementation, which is a bad
11053 idea. ([CVE-2005-2969])
11055 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11056 for Information Security, National Institute of Advanced Industrial
11057 Science and Technology [AIST, Japan)]*
11059 * Minimal support for X9.31 signatures and PSS padding modes. This is
11060 mainly for FIPS compliance and not fully integrated at this stage.
11064 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11065 the exponentiation using a fixed-length exponent. (Otherwise,
11066 the information leaked through timing could expose the secret key
11067 after many signatures; cf. Bleichenbacher's attack on DSA with
11072 * Make a new fixed-window mod_exp implementation the default for
11073 RSA, DSA, and DH private-key operations so that the sequence of
11074 squares and multiplies and the memory access pattern are
11075 independent of the particular secret key. This will mitigate
11076 cache-timing and potential related attacks.
11078 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11079 and this is automatically used by BN_mod_exp_mont() if the new flag
11080 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11081 will use this BN flag for private exponents unless the flag
11082 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11083 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11085 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11087 * Change the client implementation for SSLv23_method() and
11088 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11089 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11090 (Previously, the SSL 2.0 backwards compatible Client Hello
11091 message format would be used even with SSL_OP_NO_SSLv2.)
11095 * Add support for smime-type MIME parameter in S/MIME messages which some
11100 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11101 a threadsafe manner. Modify rsa code to use new function and add calls
11102 to dsa and dh code (which had race conditions before).
11106 * Include the fixed error library code in the C error file definitions
11107 instead of fixing them up at runtime. This keeps the error code
11108 structures constant.
11112 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11114 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11117 * Fixes for newer kerberos headers. NB: the casts are needed because
11118 the 'length' field is signed on one version and unsigned on another
11119 with no (?) obvious way to tell the difference, without these VC++
11120 complains. Also the "definition" of FAR (blank) is no longer included
11121 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11122 some needed definitions.
11126 * Undo Cygwin change.
11130 * Added support for proxy certificates according to RFC 3820.
11131 Because they may be a security thread to unaware applications,
11132 they must be explicitly allowed in run-time. See
11133 docs/HOWTO/proxy_certificates.txt for further information.
11137 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11139 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11140 server and client random values. Previously
11141 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11142 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11144 This change has negligible security impact because:
11146 1. Server and client random values still have 24 bytes of pseudo random
11149 2. Server and client random values are sent in the clear in the initial
11152 3. The master secret is derived using the premaster secret (48 bytes in
11153 size for static RSA ciphersuites) as well as client server and random
11156 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11159 *Stephen Henson, reported by UK NISCC*
11161 * Use Windows randomness collection on Cygwin.
11165 * Fix hang in EGD/PRNGD query when communication socket is closed
11166 prematurely by EGD/PRNGD.
11168 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11170 * Prompt for pass phrases when appropriate for PKCS12 input format.
11174 * Back-port of selected performance improvements from development
11175 branch, as well as improved support for PowerPC platforms.
11179 * Add lots of checks for memory allocation failure, error codes to indicate
11180 failure and freeing up memory if a failure occurs.
11182 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11184 * Add new -passin argument to dgst.
11188 * Perform some character comparisons of different types in X509_NAME_cmp:
11189 this is needed for some certificates that re-encode DNs into UTF8Strings
11190 (in violation of RFC3280) and can't or won't issue name rollover
11195 * Make an explicit check during certificate validation to see that
11196 the CA setting in each certificate on the chain is correct. As a
11197 side effect always do the following basic checks on extensions,
11198 not just when there's an associated purpose to the check:
11200 - if there is an unhandled critical extension (unless the user
11201 has chosen to ignore this fault)
11202 - if the path length has been exceeded (if one is set at all)
11203 - that certain extensions fit the associated purpose (if one has
11208 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11210 * Avoid a race condition when CRLs are checked in a multi threaded
11211 environment. This would happen due to the reordering of the revoked
11212 entries during signature checking and serial number lookup. Now the
11213 encoding is cached and the serial number sort performed under a lock.
11214 Add new STACK function sk_is_sorted().
11218 * Add Delta CRL to the extension code.
11222 * Various fixes to s3_pkt.c so alerts are sent properly.
11224 *David Holmes <d.holmes@f5.com>*
11226 * Reduce the chances of duplicate issuer name and serial numbers (in
11227 violation of RFC3280) using the OpenSSL certificate creation utilities.
11228 This is done by creating a random 64 bit value for the initial serial
11229 number when a serial number file is created or when a self signed
11230 certificate is created using 'openssl req -x509'. The initial serial
11231 number file is created using 'openssl x509 -next_serial' in CA.pl
11232 rather than being initialized to 1.
11236 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11238 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11239 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11241 *Joe Orton, Steve Henson*
11243 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11246 *Joe Orton, Steve Henson*
11248 * Make it possible to have multiple active certificates with the same
11249 subject in the CA index file. This is done only if the keyword
11250 'unique_subject' is set to 'no' in the main CA section (default
11251 if 'CA_default') of the configuration file. The value is saved
11252 with the database itself in a separate index attribute file,
11253 named like the index file with '.attr' appended to the name.
11257 * X509 verify fixes. Disable broken certificate workarounds when
11258 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11259 keyUsage extension present. Don't accept CRLs with unhandled critical
11260 extensions: since verify currently doesn't process CRL extensions this
11261 rejects a CRL with *any* critical extensions. Add new verify error codes
11266 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11267 A clarification of RFC2560 will require the use of OCTET STRINGs and
11268 some implementations cannot handle the current raw format. Since OpenSSL
11269 copies and compares OCSP nonces as opaque blobs without any attempt at
11270 parsing them this should not create any compatibility issues.
11274 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11275 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11276 this HMAC (and other) operations are several times slower than OpenSSL
11281 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11283 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11285 * Use the correct content when signing type "other".
11289 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11291 * Fix various bugs revealed by running the NISCC test suite:
11293 Stop out of bounds reads in the ASN1 code when presented with
11294 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11296 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11298 If verify callback ignores invalid public key errors don't try to check
11299 certificate signature with the NULL public key.
11303 * New -ignore_err option in ocsp application to stop the server
11304 exiting on the first error in a request.
11308 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11309 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11314 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11315 extra data after the compression methods not only for TLS 1.0
11316 but also for SSL 3.0 (as required by the specification).
11318 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11320 * Change X509_certificate_type() to mark the key as exported/exportable
11321 when it's 512 *bits* long, not 512 bytes.
11325 * Change AES_cbc_encrypt() so it outputs exact multiple of
11326 blocks during encryption.
11330 * Various fixes to base64 BIO and non blocking I/O. On write
11331 flushes were not handled properly if the BIO retried. On read
11332 data was not being buffered properly and had various logic bugs.
11333 This also affects blocking I/O when the data being decoded is a
11338 * Various S/MIME bugfixes and compatibility changes:
11339 output correct application/pkcs7 MIME type if
11340 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11341 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11342 of files as .eml work). Correctly handle very long lines in MIME
11347 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11349 * Countermeasure against the Klima-Pokorny-Rosa extension of
11350 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11351 a protocol version number mismatch like a decryption error
11352 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11356 * Turn on RSA blinding by default in the default implementation
11357 to avoid a timing attack. Applications that don't want it can call
11358 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11359 They would be ill-advised to do so in most cases.
11361 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11363 * Change RSA blinding code so that it works when the PRNG is not
11364 seeded (in this case, the secret RSA exponent is abused as
11365 an unpredictable seed -- if it is not unpredictable, there
11366 is no point in blinding anyway). Make RSA blinding thread-safe
11367 by remembering the creator's thread ID in rsa->blinding and
11368 having all other threads use local one-time blinding factors
11369 (this requires more computation than sharing rsa->blinding, but
11370 avoids excessive locking; and if an RSA object is not shared
11371 between threads, blinding will still be very fast).
11375 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11376 ENGINE as defaults for all supported algorithms irrespective of
11377 the 'flags' parameter. 'flags' is now honoured, so applications
11378 should make sure they are passing it correctly.
11382 * Target "mingw" now allows native Windows code to be generated in
11383 the Cygwin environment as well as with the MinGW compiler.
11387 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11389 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11390 via timing by performing a MAC computation even if incorrect
11391 block cipher padding has been found. This is a countermeasure
11392 against active attacks where the attacker has to distinguish
11393 between bad padding and a MAC verification error. ([CVE-2003-0078])
11395 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11396 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11397 Martin Vuagnoux (EPFL, Ilion)*
11399 * Make the no-err option work as intended. The intention with no-err
11400 is not to have the whole error stack handling routines removed from
11401 libcrypto, it's only intended to remove all the function name and
11402 reason texts, thereby removing some of the footprint that may not
11403 be interesting if those errors aren't displayed anyway.
11405 NOTE: it's still possible for any application or module to have its
11406 own set of error texts inserted. The routines are there, just not
11407 used by default when no-err is given.
11411 * Add support for FreeBSD on IA64.
11413 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11415 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11416 Kerberos function mit_des_cbc_cksum(). Before this change,
11417 the value returned by DES_cbc_cksum() was like the one from
11418 mit_des_cbc_cksum(), except the bytes were swapped.
11420 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11422 * Allow an application to disable the automatic SSL chain building.
11423 Before this a rather primitive chain build was always performed in
11424 ssl3_output_cert_chain(): an application had no way to send the
11425 correct chain if the automatic operation produced an incorrect result.
11427 Now the chain builder is disabled if either:
11429 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11431 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11433 The reasoning behind this is that an application would not want the
11434 auto chain building to take place if extra chain certificates are
11435 present and it might also want a means of sending no additional
11436 certificates (for example the chain has two certificates and the
11441 * Add the possibility to build without the ENGINE framework.
11443 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11445 * Under Win32 gmtime() can return NULL: check return value in
11446 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11450 * DSA routines: under certain error conditions uninitialized BN objects
11451 could be freed. Solution: make sure initialization is performed early
11452 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11453 Nils Larsch <nla@trustcenter.de> via PR#459)
11457 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11458 checked on reconnect on the client side, therefore session resumption
11459 could still fail with a "ssl session id is different" error. This
11460 behaviour is masked when SSL_OP_ALL is used due to
11461 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11462 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11463 followup to PR #377.
11467 * IA-32 assembler support enhancements: unified ELF targets, support
11468 for SCO/Caldera platforms, fix for Cygwin shared build.
11472 * Add support for FreeBSD on sparc64. As a consequence, support for
11473 FreeBSD on non-x86 processors is separate from x86 processors on
11474 the config script, much like the NetBSD support.
11476 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11478 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11480 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11483 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11484 code (06) was taken as the first octet of the session ID and the last
11485 octet was ignored consequently. As a result SSLv2 client side session
11486 caching could not have worked due to the session ID mismatch between
11488 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11493 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11494 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11499 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11500 seems that in spite of existing for more than a year, many application
11501 author have done nothing to provide the necessary callbacks, which
11502 means that this particular engine will not work properly anywhere.
11503 This is a very unfortunate situation which forces us, in the name
11504 of usability, to give the hw_ncipher.c a static lock, which is part
11506 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11507 appear in 0.9.8 or later. We EXPECT application authors to have
11508 dealt properly with this when 0.9.8 is released (unless we actually
11509 make such changes in the libcrypto locking code that changes will
11510 have to be made anyway).
11514 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11515 octets have been read, EOF or an error occurs. Without this change
11516 some truncated ASN1 structures will not produce an error.
11520 * Disable Heimdal support, since it hasn't been fully implemented.
11521 Still give the possibility to force the use of Heimdal, but with
11522 warnings and a request that patches get sent to openssl-dev.
11526 * Add the VC-CE target, introduce the WINCE sysname, and add
11527 INSTALL.WCE and appropriate conditionals to make it build.
11529 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11531 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11532 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11533 edit numbers of the version.
11535 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11537 * Introduce safe string copy and catenation functions
11538 (BUF_strlcpy() and BUF_strlcat()).
11540 *Ben Laurie (CHATS) and Richard Levitte*
11542 * Avoid using fixed-size buffers for one-line DNs.
11544 *Ben Laurie (CHATS)*
11546 * Add BUF_MEM_grow_clean() to avoid information leakage when
11547 resizing buffers containing secrets, and use where appropriate.
11549 *Ben Laurie (CHATS)*
11551 * Avoid using fixed size buffers for configuration file location.
11553 *Ben Laurie (CHATS)*
11555 * Avoid filename truncation for various CA files.
11557 *Ben Laurie (CHATS)*
11559 * Use sizeof in preference to magic numbers.
11561 *Ben Laurie (CHATS)*
11563 * Avoid filename truncation in cert requests.
11565 *Ben Laurie (CHATS)*
11567 * Add assertions to check for (supposedly impossible) buffer
11570 *Ben Laurie (CHATS)*
11572 * Don't cache truncated DNS entries in the local cache (this could
11573 potentially lead to a spoofing attack).
11575 *Ben Laurie (CHATS)*
11577 * Fix various buffers to be large enough for hex/decimal
11578 representations in a platform independent manner.
11580 *Ben Laurie (CHATS)*
11582 * Add CRYPTO_realloc_clean() to avoid information leakage when
11583 resizing buffers containing secrets, and use where appropriate.
11585 *Ben Laurie (CHATS)*
11587 * Add BIO_indent() to avoid much slightly worrying code to do
11590 *Ben Laurie (CHATS)*
11592 * Convert sprintf()/BIO_puts() to BIO_printf().
11594 *Ben Laurie (CHATS)*
11596 * buffer_gets() could terminate with the buffer only half
11599 *Ben Laurie (CHATS)*
11601 * Add assertions to prevent user-supplied crypto functions from
11602 overflowing internal buffers by having large block sizes, etc.
11604 *Ben Laurie (CHATS)*
11606 * New OPENSSL_assert() macro (similar to assert(), but enabled
11609 *Ben Laurie (CHATS)*
11611 * Eliminate unused copy of key in RC4.
11613 *Ben Laurie (CHATS)*
11615 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11617 *Ben Laurie (CHATS)*
11619 * Fix off-by-one error in EGD path.
11621 *Ben Laurie (CHATS)*
11623 * If RANDFILE path is too long, ignore instead of truncating.
11625 *Ben Laurie (CHATS)*
11627 * Eliminate unused and incorrectly sized X.509 structure
11630 *Ben Laurie (CHATS)*
11632 * Eliminate unused and dangerous function knumber().
11634 *Ben Laurie (CHATS)*
11636 * Eliminate unused and dangerous structure, KSSL_ERR.
11638 *Ben Laurie (CHATS)*
11640 * Protect against overlong session ID context length in an encoded
11641 session object. Since these are local, this does not appear to be
11644 *Ben Laurie (CHATS)*
11646 * Change from security patch (see 0.9.6e below) that did not affect
11647 the 0.9.6 release series:
11649 Remote buffer overflow in SSL3 protocol - an attacker could
11650 supply an oversized master key in Kerberos-enabled versions.
11653 *Ben Laurie (CHATS)*
11655 * Change the SSL kerb5 codes to match RFC 2712.
11659 * Make -nameopt work fully for req and add -reqopt switch.
11661 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11663 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11665 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11667 * Make sure tests can be performed even if the corresponding algorithms
11668 have been removed entirely. This was also the last step to make
11669 OpenSSL compilable with DJGPP under all reasonable conditions.
11671 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11673 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11674 to allow version independent disabling of normally unselected ciphers,
11675 which may be activated as a side-effect of selecting a single cipher.
11677 (E.g., cipher list string "RSA" enables ciphersuites that are left
11678 out of "ALL" because they do not provide symmetric encryption.
11679 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11681 *Lutz Jaenicke, Bodo Moeller*
11683 * Add appropriate support for separate platform-dependent build
11684 directories. The recommended way to make a platform-dependent
11685 build directory is the following (tested on Linux), maybe with
11688 # Place yourself outside of the OpenSSL source tree. In
11689 # this example, the environment variable OPENSSL_SOURCE
11690 # is assumed to contain the absolute OpenSSL source directory.
11691 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11692 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11693 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11694 mkdir -p `dirname $F`
11695 ln -s $OPENSSL_SOURCE/$F $F
11698 To be absolutely sure not to disturb the source tree, a "make clean"
11699 is a good thing. If it isn't successful, don't worry about it,
11700 it probably means the source directory is very clean.
11704 * Make sure any ENGINE control commands make local copies of string
11705 pointers passed to them whenever necessary. Otherwise it is possible
11706 the caller may have overwritten (or deallocated) the original string
11707 data when a later ENGINE operation tries to use the stored values.
11709 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11711 * Improve diagnostics in file reading and command-line digests.
11713 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11715 * Add AES modes CFB and OFB to the object database. Correct an
11716 error in AES-CFB decryption.
11720 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11721 allows existing EVP_CIPHER_CTX structures to be reused after
11722 calling `EVP_*Final()`. This behaviour is used by encryption
11723 BIOs and some applications. This has the side effect that
11724 applications must explicitly clean up cipher contexts with
11725 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11729 * Check the values of dna and dnb in bn_mul_recursive before calling
11730 bn_mul_comba (a non zero value means the a or b arrays do not contain
11731 n2 elements) and fallback to bn_mul_normal if either is not zero.
11735 * Fix escaping of non-ASCII characters when using the -subj option
11736 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11740 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11741 form for "surname", serialNumber has no short form.
11742 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11743 therefore remove "mail" short name for "internet 7".
11744 The OID for unique identifiers in X509 certificates is
11745 x500UniqueIdentifier, not uniqueIdentifier.
11746 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11750 * Add an "init" command to the ENGINE config module and auto initialize
11751 ENGINEs. Without any "init" command the ENGINE will be initialized
11752 after all ctrl commands have been executed on it. If init=1 the
11753 ENGINE is initialized at that point (ctrls before that point are run
11754 on the uninitialized ENGINE and after on the initialized one). If
11755 init=0 then the ENGINE will not be initialized at all.
11759 * Fix the 'app_verify_callback' interface so that the user-defined
11760 argument is actually passed to the callback: In the
11761 SSL_CTX_set_cert_verify_callback() prototype, the callback
11762 declaration has been changed from
11765 int (*cb)(X509_STORE_CTX *,void *);
11766 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11767 i=s->ctx->app_verify_callback(&ctx)
11768 has been changed into
11769 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11771 To update applications using SSL_CTX_set_cert_verify_callback(),
11772 a dummy argument can be added to their callback functions.
11774 *D. K. Smetters <smetters@parc.xerox.com>*
11776 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11778 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11780 * Add and OPENSSL_LOAD_CONF define which will cause
11781 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11782 This allows older applications to transparently support certain
11783 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11784 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11785 load the config file and OPENSSL_add_all_algorithms_conf() which will
11786 always load it have also been added.
11790 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11791 Adjust NIDs and EVP layer.
11793 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11795 * Config modules support in openssl utility.
11797 Most commands now load modules from the config file,
11798 though in a few (such as version) this isn't done
11799 because it couldn't be used for anything.
11801 In the case of ca and req the config file used is
11802 the same as the utility itself: that is the -config
11803 command line option can be used to specify an
11808 * Move default behaviour from OPENSSL_config(). If appname is NULL
11809 use "openssl_conf" if filename is NULL use default openssl config file.
11813 * Add an argument to OPENSSL_config() to allow the use of an alternative
11814 config section name. Add a new flag to tolerate a missing config file
11815 and move code to CONF_modules_load_file().
11819 * Support for crypto accelerator cards from Accelerated Encryption
11820 Processing, www.aep.ie. (Use engine 'aep')
11821 The support was copied from 0.9.6c [engine] and adapted/corrected
11822 to work with the new engine framework.
11824 *AEP Inc. and Richard Levitte*
11826 * Support for SureWare crypto accelerator cards from Baltimore
11827 Technologies. (Use engine 'sureware')
11828 The support was copied from 0.9.6c [engine] and adapted
11829 to work with the new engine framework.
11833 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11834 make the newer ENGINE framework commands for the CHIL engine work.
11836 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11838 * Make it possible to produce shared libraries on ReliantUNIX.
11840 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11842 * Add the configuration target debug-linux-ppro.
11843 Make 'openssl rsa' use the general key loading routines
11844 implemented in `apps.c`, and make those routines able to
11845 handle the key format FORMAT_NETSCAPE and the variant
11848 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11850 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11852 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11854 * Add -keyform to rsautl, and document -engine.
11856 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11858 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11859 BIO_R_NO_SUCH_FILE error code rather than the generic
11860 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11864 * Add new functions
11865 ERR_peek_last_error
11866 ERR_peek_last_error_line
11867 ERR_peek_last_error_line_data.
11868 These are similar to
11870 ERR_peek_error_line
11871 ERR_peek_error_line_data,
11872 but report on the latest error recorded rather than the first one
11873 still in the error queue.
11875 *Ben Laurie, Bodo Moeller*
11877 * default_algorithms option in ENGINE config module. This allows things
11879 default_algorithms = ALL
11880 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11884 * Preliminary ENGINE config module.
11888 * New experimental application configuration code.
11892 * Change the AES code to follow the same name structure as all other
11893 symmetric ciphers, and behave the same way. Move everything to
11894 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11896 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11898 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11900 *Ben Laurie and Theo de Raadt*
11902 * Add option to output public keys in req command.
11904 *Massimiliano Pala madwolf@openca.org*
11906 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11907 (up to about 10% better than before for P-192 and P-224).
11911 * New functions/macros
11913 SSL_CTX_set_msg_callback(ctx, cb)
11914 SSL_CTX_set_msg_callback_arg(ctx, arg)
11915 SSL_set_msg_callback(ssl, cb)
11916 SSL_set_msg_callback_arg(ssl, arg)
11918 to request calling a callback function
11920 void cb(int write_p, int version, int content_type,
11921 const void *buf, size_t len, SSL *ssl, void *arg)
11923 whenever a protocol message has been completely received
11924 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11925 protocol version according to which the SSL library interprets
11926 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11927 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11928 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11929 specification (change_cipher_spec(20), alert(21), handshake(22)).
11930 'buf' and 'len' point to the actual message, 'ssl' to the
11931 SSL object, and 'arg' is the application-defined value set by
11932 SSL[_CTX]_set_msg_callback_arg().
11934 'openssl s_client' and 'openssl s_server' have new '-msg' options
11935 to enable a callback that displays all protocol messages.
11939 * Change the shared library support so shared libraries are built as
11940 soon as the corresponding static library is finished, and thereby get
11941 openssl and the test programs linked against the shared library.
11942 This still only happens when the keyword "shard" has been given to
11943 the configuration scripts.
11945 NOTE: shared library support is still an experimental thing, and
11946 backward binary compatibility is still not guaranteed.
11948 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11950 * Add support for Subject Information Access extension.
11952 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11954 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11955 additional bytes when new memory had to be allocated, not just
11956 when reusing an existing buffer.
11960 * New command line and configuration option 'utf8' for the req command.
11961 This allows field values to be specified as UTF8 strings.
11965 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11966 runs for the former and machine-readable output for the latter.
11970 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11971 of the e-mail address in the DN (i.e., it will go into a certificate
11972 extension only). The new configuration file option 'email_in_dn = no'
11973 has the same effect.
11975 *Massimiliano Pala madwolf@openca.org*
11977 * Change all functions with names starting with `des_` to be starting
11978 with `DES_` instead. Add wrappers that are compatible with libdes,
11979 but are named `_ossl_old_des_*`. Finally, add macros that map the
11980 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11981 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11982 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11985 Since we provide two compatibility mappings, the user needs to
11986 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11987 compatibility is desired. The default (i.e., when that macro
11988 isn't defined) is OpenSSL 0.9.6c compatibility.
11990 There are also macros that enable and disable the support of old
11991 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11992 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11993 are defined, the default will apply: to support the old des routines.
11995 In either case, one must include openssl/des.h to get the correct
11996 definitions. Do not try to just include openssl/des_old.h, that
11999 NOTE: This is a major break of an old API into a new one. Software
12000 authors are encouraged to switch to the `DES_` style functions. Some
12001 time in the future, des_old.h and the libdes compatibility functions
12002 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12003 default), and then completely removed.
12007 * Test for certificates which contain unsupported critical extensions.
12008 If such a certificate is found during a verify operation it is
12009 rejected by default: this behaviour can be overridden by either
12010 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12011 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12012 X509_supported_extension() has also been added which returns 1 if a
12013 particular extension is supported.
12017 * Modify the behaviour of EVP cipher functions in similar way to digests
12018 to retain compatibility with existing code.
12022 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12023 compatibility with existing code. In particular the 'ctx' parameter does
12024 not have to be to be initialized before the call to EVP_DigestInit() and
12025 it is tidied up after a call to EVP_DigestFinal(). New function
12026 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12027 EVP_MD_CTX_copy() changed to not require the destination to be
12028 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12029 requires the destination to be valid.
12031 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12032 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12036 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12037 so that complete 'Handshake' protocol structures are kept in memory
12038 instead of overwriting 'msg_type' and 'length' with 'body' data.
12042 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12044 *Massimo Santin via Richard Levitte*
12046 * Major restructuring to the underlying ENGINE code. This includes
12047 reduction of linker bloat, separation of pure "ENGINE" manipulation
12048 (initialisation, etc) from functionality dealing with implementations
12049 of specific crypto interfaces. This change also introduces integrated
12050 support for symmetric ciphers and digest implementations - so ENGINEs
12051 can now accelerate these by providing EVP_CIPHER and EVP_MD
12052 implementations of their own. This is detailed in
12053 [crypto/engine/README.md](crypto/engine/README.md)
12054 as it couldn't be adequately described here. However, there are a few
12055 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12056 were changed in the original introduction of ENGINE code have now
12057 reverted back - the hooking from this code to ENGINE is now a good
12058 deal more passive and at run-time, operations deal directly with
12059 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12060 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12061 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12062 they were not being used by the framework as there is no concept of a
12063 BIGNUM_METHOD and they could not be generalised to the new
12064 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12065 ENGINE_cpy() has been removed as it cannot be consistently defined in
12070 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12074 * Change mkdef.pl to sort symbols that get the same entry number,
12075 and make sure the automatically generated functions `ERR_load_*`
12076 become part of libeay.num as well.
12080 * New function SSL_renegotiate_pending(). This returns true once
12081 renegotiation has been requested (either SSL_renegotiate() call
12082 or HelloRequest/ClientHello received from the peer) and becomes
12083 false once a handshake has been completed.
12084 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12085 sends a HelloRequest, but does not ensure that a handshake takes
12086 place. SSL_renegotiate_pending() is useful for checking if the
12087 client has followed the request.)
12091 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12092 By default, clients may request session resumption even during
12093 renegotiation (if session ID contexts permit); with this option,
12094 session resumption is possible only in the first handshake.
12096 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12097 more bits available for options that should not be part of
12098 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12102 * Add some demos for certificate and certificate request creation.
12106 * Make maximum certificate chain size accepted from the peer application
12107 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12108 "Douglas E. Engert" <deengert@anl.gov>.
12112 * Add support for shared libraries for Unixware-7
12113 (Boyd Lynn Gerber <gerberb@zenez.com>).
12117 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12118 be done prior to destruction. Use this to unload error strings from
12119 ENGINEs that load their own error strings. NB: This adds two new API
12120 functions to "get" and "set" this destroy handler in an ENGINE.
12124 * Alter all existing ENGINE implementations (except "openssl" and
12125 "openbsd") to dynamically instantiate their own error strings. This
12126 makes them more flexible to be built both as statically-linked ENGINEs
12127 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12128 Also, add stub code to each that makes building them as self-contained
12129 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12133 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12134 implementations into applications that are completely implemented in
12135 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12136 commands that can be used to configure what shared-library to load and
12137 to control aspects of the way it is handled. Also, made an update to
12138 the [README-Engine.md](README-Engine.md) file
12139 that brings its information up-to-date and
12140 provides some information and instructions on the "dynamic" ENGINE
12141 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12145 * Make it possible to unload ranges of ERR strings with a new
12146 "ERR_unload_strings" function.
12150 * Add a copy() function to EVP_MD.
12154 * Make EVP_MD routines take a context pointer instead of just the
12155 md_data void pointer.
12159 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12160 that the digest can only process a single chunk of data
12161 (typically because it is provided by a piece of
12162 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12163 is only going to provide a single chunk of data, and hence the
12164 framework needn't accumulate the data for oneshot drivers.
12168 * As with "ERR", make it possible to replace the underlying "ex_data"
12169 functions. This change also alters the storage and management of global
12170 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12171 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12172 index counters. The API functions that use this state have been changed
12173 to take a "class_index" rather than pointers to the class's local STACK
12174 and counter, and there is now an API function to dynamically create new
12175 classes. This centralisation allows us to (a) plug a lot of the
12176 thread-safety problems that existed, and (b) makes it possible to clean
12177 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12178 such data would previously have always leaked in application code and
12179 workarounds were in place to make the memory debugging turn a blind eye
12180 to it. Application code that doesn't use this new function will still
12181 leak as before, but their memory debugging output will announce it now
12182 rather than letting it slide.
12184 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12185 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12186 has a return value to indicate success or failure.
12190 * Make it possible to replace the underlying "ERR" functions such that the
12191 global state (2 LHASH tables and 2 locks) is only used by the "default"
12192 implementation. This change also adds two functions to "get" and "set"
12193 the implementation prior to it being automatically set the first time
12194 any other ERR function takes place. Ie. an application can call "get",
12195 pass the return value to a module it has just loaded, and that module
12196 can call its own "set" function using that value. This means the
12197 module's "ERR" operations will use (and modify) the error state in the
12198 application and not in its own statically linked copy of OpenSSL code.
12202 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12203 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12204 the operation, and provides a more encapsulated way for external code
12205 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12206 to use these functions rather than manually incrementing the counts.
12208 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12212 * Add EVP test program.
12216 * Add symmetric cipher support to ENGINE. Expect the API to change!
12220 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12221 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12222 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12223 These allow a CRL to be built without having to access X509_CRL fields
12224 directly. Modify 'ca' application to use new functions.
12228 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12229 bug workarounds. Rollback attack detection is a security feature.
12230 The problem will only arise on OpenSSL servers when TLSv1 is not
12231 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12232 Software authors not wanting to support TLSv1 will have special reasons
12233 for their choice and can explicitly enable this option.
12235 *Bodo Moeller, Lutz Jaenicke*
12237 * Rationalise EVP so it can be extended: don't include a union of
12238 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12239 (similar to those existing for EVP_CIPHER_CTX).
12244 EVP_MD_CTX_init(&md); /* new function call */
12245 EVP_DigestInit(&md, EVP_sha1());
12246 EVP_DigestUpdate(&md, in, len);
12247 EVP_DigestFinal(&md, out, NULL);
12248 EVP_MD_CTX_cleanup(&md); /* new function call */
12252 * Make DES key schedule conform to the usual scheme, as well as
12253 correcting its structure. This means that calls to DES functions
12254 now have to pass a pointer to a des_key_schedule instead of a
12255 plain des_key_schedule (which was actually always a pointer
12258 des_key_schedule ks;
12260 des_set_key_checked(..., &ks);
12261 des_ncbc_encrypt(..., &ks, ...);
12263 (Note that a later change renames 'des_...' into 'DES_...'.)
12267 * Initial reduction of linker bloat: the use of some functions, such as
12268 PEM causes large amounts of unused functions to be linked in due to
12269 poor organisation. For example pem_all.c contains every PEM function
12270 which has a knock on effect of linking in large amounts of (unused)
12271 ASN1 code. Grouping together similar functions and splitting unrelated
12272 functions prevents this.
12276 * Cleanup of EVP macros.
12280 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12281 correct `_ecb suffix`.
12285 * Add initial OCSP responder support to ocsp application. The
12286 revocation information is handled using the text based index
12287 use by the ca application. The responder can either handle
12288 requests generated internally, supplied in files (for example
12289 via a CGI script) or using an internal minimal server.
12293 * Add configuration choices to get zlib compression for TLS.
12297 * Changes to Kerberos SSL for RFC 2712 compliance:
12298 1. Implemented real KerberosWrapper, instead of just using
12299 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12300 2. Implemented optional authenticator field of KerberosWrapper.
12302 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12303 and authenticator structs; see crypto/krb5/.
12305 Generalized Kerberos calls to support multiple Kerberos libraries.
12306 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12307 via Richard Levitte*
12309 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12310 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12311 values for each of the key sizes rather than having just
12312 parameters (and 'speed' generating keys each time).
12316 * Speed up EVP routines.
12319 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12320 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12321 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12322 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12324 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12325 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12326 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12329 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12331 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12335 * Added the OS2-EMX target.
12337 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12339 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12340 New functions to support `NCONF `routines in extension code.
12341 New function `CONF_set_nconf()`
12342 to allow functions which take an `NCONF` to also handle the old `LHASH`
12343 structure: this means that the old `CONF` compatible routines can be
12344 retained (in particular w.rt. extensions) without having to duplicate the
12345 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12349 * Enhance the general user interface with mechanisms for inner control
12350 and with possibilities to have yes/no kind of prompts.
12354 * Change all calls to low level digest routines in the library and
12355 applications to use EVP. Add missing calls to HMAC_cleanup() and
12356 don't assume HMAC_CTX can be copied using memcpy().
12358 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12360 * Add the possibility to control engines through control names but with
12361 arbitrary arguments instead of just a string.
12362 Change the key loaders to take a UI_METHOD instead of a callback
12363 function pointer. NOTE: this breaks binary compatibility with earlier
12364 versions of OpenSSL [engine].
12365 Adapt the nCipher code for these new conditions and add a card insertion
12370 * Enhance the general user interface with mechanisms to better support
12371 dialog box interfaces, application-defined prompts, the possibility
12372 to use defaults (for example default passwords from somewhere else)
12373 and interrupts/cancellations.
12377 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12378 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12382 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12383 tidy up some unnecessarily weird code in 'sk_new()').
12385 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12387 * Change the key loading routines for ENGINEs to use the same kind
12388 callback (pem_password_cb) as all other routines that need this
12393 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12394 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12395 than this minimum value is recommended.
12399 * New random seeder for OpenVMS, using the system process statistics
12400 that are easily reachable.
12404 * Windows apparently can't transparently handle global
12405 variables defined in DLLs. Initialisations such as:
12407 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12409 won't compile. This is used by the any applications that need to
12410 declare their own ASN1 modules. This was fixed by adding the option
12411 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12412 needed for static libraries under Win32.
12416 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12417 setting of purpose and trust fields. New X509_STORE trust and
12418 purpose functions and tidy up setting in other SSL functions.
12422 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12423 structure. These are inherited by X509_STORE_CTX when it is
12424 initialised. This allows various defaults to be set in the
12425 X509_STORE structure (such as flags for CRL checking and custom
12426 purpose or trust settings) for functions which only use X509_STORE_CTX
12427 internally such as S/MIME.
12429 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12430 trust settings if they are not set in X509_STORE. This allows X509_STORE
12431 purposes and trust (in S/MIME for example) to override any set by default.
12433 Add command line options for CRL checking to smime, s_client and s_server
12438 * Initial CRL based revocation checking. If the CRL checking flag(s)
12439 are set then the CRL is looked up in the X509_STORE structure and
12440 its validity and signature checked, then if the certificate is found
12441 in the CRL the verify fails with a revoked error.
12443 Various new CRL related callbacks added to X509_STORE_CTX structure.
12445 Command line options added to 'verify' application to support this.
12447 This needs some additional work, such as being able to handle multiple
12448 CRLs with different times, extension based lookup (rather than just
12449 by subject name) and ultimately more complete V2 CRL extension
12454 * Add a general user interface API (crypto/ui/). This is designed
12455 to replace things like des_read_password and friends (backward
12456 compatibility functions using this new API are provided).
12457 The purpose is to remove prompting functions from the DES code
12458 section as well as provide for prompting through dialog boxes in
12459 a window system and the like.
12463 * Add "ex_data" support to ENGINE so implementations can add state at a
12464 per-structure level rather than having to store it globally.
12468 * Make it possible for ENGINE structures to be copied when retrieved by
12469 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12470 This causes the "original" ENGINE structure to act like a template,
12471 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12472 operational state can be localised to each ENGINE structure, despite the
12473 fact they all share the same "methods". New ENGINE structures returned in
12474 this case have no functional references and the return value is the single
12475 structural reference. This matches the single structural reference returned
12476 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12481 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12482 needs to match any other type at all we need to manually clear the
12487 * Changes to the "openssl engine" utility to include;
12488 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12489 about an ENGINE's available control commands.
12490 - executing control commands from command line arguments using the
12491 '-pre' and '-post' switches. '-post' is only used if '-t' is
12492 specified and the ENGINE is successfully initialised. The syntax for
12493 the individual commands are colon-separated, for example;
12494 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12498 * New dynamic control command support for ENGINEs. ENGINEs can now
12499 declare their own commands (numbers), names (strings), descriptions,
12500 and input types for run-time discovery by calling applications. A
12501 subset of these commands are implicitly classed as "executable"
12502 depending on their input type, and only these can be invoked through
12503 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12504 can be based on user input, config files, etc). The distinction is
12505 that "executable" commands cannot return anything other than a boolean
12506 result and can only support numeric or string input, whereas some
12507 discoverable commands may only be for direct use through
12508 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12509 pointers, or other custom uses. The "executable" commands are to
12510 support parameterisations of ENGINE behaviour that can be
12511 unambiguously defined by ENGINEs and used consistently across any
12512 OpenSSL-based application. Commands have been added to all the
12513 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12514 control over shared-library paths without source code alterations.
12518 * Changed all ENGINE implementations to dynamically allocate their
12519 ENGINEs rather than declaring them statically. Apart from this being
12520 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12521 this also allows the implementations to compile without using the
12522 internal engine_int.h header.
12526 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12527 'const' value. Any code that should be able to modify a RAND_METHOD
12528 should already have non-const pointers to it (ie. they should only
12529 modify their own ones).
12533 * Made a variety of little tweaks to the ENGINE code.
12534 - "atalla" and "ubsec" string definitions were moved from header files
12535 to C code. "nuron" string definitions were placed in variables
12536 rather than hard-coded - allowing parameterisation of these values
12537 later on via ctrl() commands.
12538 - Removed unused "#if 0"'d code.
12539 - Fixed engine list iteration code so it uses ENGINE_free() to release
12540 structural references.
12541 - Constified the RAND_METHOD element of ENGINE structures.
12542 - Constified various get/set functions as appropriate and added
12543 missing functions (including a catch-all ENGINE_cpy that duplicates
12544 all ENGINE values onto a new ENGINE except reference counts/state).
12545 - Removed NULL parameter checks in get/set functions. Setting a method
12546 or function to NULL is a way of cancelling out a previously set
12547 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12548 and doesn't justify the extra error symbols and code.
12549 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12550 flags from engine_int.h to engine.h.
12551 - Changed prototypes for ENGINE handler functions (init(), finish(),
12552 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12556 * Implement binary inversion algorithm for BN_mod_inverse in addition
12557 to the algorithm using long division. The binary algorithm can be
12558 used only if the modulus is odd. On 32-bit systems, it is faster
12559 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12560 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12561 up to 450 bits. In 64-bit environments, the binary algorithm
12562 appears to be advantageous for much longer moduli; here we use it
12563 for moduli up to 2048 bits.
12567 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12568 could not support the combine flag in choice fields.
12572 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12573 extensions from a certificate request to the certificate.
12577 * Allow multiple 'certopt' and 'nameopt' options to be separated
12578 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12579 file: this allows the display of the certificate about to be
12580 signed to be customised, to allow certain fields to be included
12581 or excluded and extension details. The old system didn't display
12582 multicharacter strings properly, omitted fields not in the policy
12583 and couldn't display additional details such as extensions.
12587 * Function EC_POINTs_mul for multiple scalar multiplication
12588 of an arbitrary number of elliptic curve points
12589 \sum scalars[i]*points[i],
12590 optionally including the generator defined for the EC_GROUP:
12591 scalar*generator + \sum scalars[i]*points[i].
12593 EC_POINT_mul is a simple wrapper function for the typical case
12594 that the point list has just one item (besides the optional
12599 * First EC_METHODs for curves over GF(p):
12601 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12602 operations and provides various method functions that can also
12603 operate with faster implementations of modular arithmetic.
12605 EC_GFp_mont_method() reuses most functions that are part of
12606 EC_GFp_simple_method, but uses Montgomery arithmetic.
12608 *Bodo Moeller; point addition and point doubling
12609 implementation directly derived from source code provided by
12610 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12612 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12613 crypto/ec/ec_lib.c):
12615 Curves are EC_GROUP objects (with an optional group generator)
12616 based on EC_METHODs that are built into the library.
12618 Points are EC_POINT objects based on EC_GROUP objects.
12620 Most of the framework would be able to handle curves over arbitrary
12621 finite fields, but as there are no obvious types for fields other
12622 than GF(p), some functions are limited to that for now.
12626 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12627 that the file contains a complete HTTP response.
12631 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12632 change the def and num file printf format specifier from "%-40sXXX"
12633 to "%-39s XXX". The latter will always guarantee a space after the
12634 field while the former will cause them to run together if the field
12635 is 40 of more characters long.
12639 * Constify the cipher and digest 'method' functions and structures
12640 and modify related functions to take constant EVP_MD and EVP_CIPHER
12645 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12646 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12650 * Modify `EVP_Digest*()` routines so they now return values. Although the
12651 internal software routines can never fail additional hardware versions
12656 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12658 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12659 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12662 ERR_R_NESTED_ASN1_ERROR
12664 ERR_R_MISSING_ASN1_EOS
12665 were 4 .. 9, conflicting with
12666 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12668 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12669 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12671 Add new error code 'ERR_R_INTERNAL_ERROR'.
12675 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12680 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12681 sets the subject name for a new request or supersedes the
12682 subject name in a given request. Formats that can be parsed are
12683 'CN=Some Name, OU=myOU, C=IT'
12685 'CN=Some Name/OU=myOU/C=IT'.
12687 Add options '-batch' and '-verbose' to 'openssl req'.
12689 *Massimiliano Pala <madwolf@hackmasters.net>*
12691 * Introduce the possibility to access global variables through
12692 functions on platform were that's the best way to handle exporting
12693 global variables in shared libraries. To enable this functionality,
12694 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12695 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12696 is normally done by Configure or something similar).
12698 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12699 in the source file (foo.c) like this:
12701 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12702 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12704 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12705 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12707 OPENSSL_DECLARE_GLOBAL(int,foo);
12708 #define foo OPENSSL_GLOBAL_REF(foo)
12709 OPENSSL_DECLARE_GLOBAL(double,bar);
12710 #define bar OPENSSL_GLOBAL_REF(bar)
12712 The #defines are very important, and therefore so is including the
12713 header file everywhere where the defined globals are used.
12715 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12716 of ASN.1 items, but that structure is a bit different.
12718 The largest change is in util/mkdef.pl which has been enhanced with
12719 better and easier to understand logic to choose which symbols should
12720 go into the Windows .def files as well as a number of fixes and code
12721 cleanup (among others, algorithm keywords are now sorted
12722 lexicographically to avoid constant rewrites).
12726 * In BN_div() keep a copy of the sign of 'num' before writing the
12727 result to 'rm' because if rm==num the value will be overwritten
12728 and produce the wrong result if 'num' is negative: this caused
12729 problems with BN_mod() and BN_nnmod().
12733 * Function OCSP_request_verify(). This checks the signature on an
12734 OCSP request and verifies the signer certificate. The signer
12735 certificate is just checked for a generic purpose and OCSP request
12740 * Add OCSP_check_validity() function to check the validity of OCSP
12741 responses. OCSP responses are prepared in real time and may only
12742 be a few seconds old. Simply checking that the current time lies
12743 between thisUpdate and nextUpdate max reject otherwise valid responses
12744 caused by either OCSP responder or client clock inaccuracy. Instead
12745 we allow thisUpdate and nextUpdate to fall within a certain period of
12746 the current time. The age of the response can also optionally be
12747 checked. Two new options -validity_period and -status_age added to
12752 * If signature or public key algorithm is unrecognized print out its
12753 OID rather that just UNKNOWN.
12757 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12758 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12759 ID to be generated from the issuer certificate alone which can then be
12760 passed to OCSP_id_issuer_cmp().
12764 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12765 ASN1 modules to export functions returning ASN1_ITEM pointers
12766 instead of the ASN1_ITEM structures themselves. This adds several
12767 new macros which allow the underlying ASN1 function/structure to
12768 be accessed transparently. As a result code should not use ASN1_ITEM
12769 references directly (such as &X509_it) but instead use the relevant
12770 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12771 use of the new ASN1 code on platforms where exporting structures
12772 is problematical (for example in shared libraries) but exporting
12773 functions returning pointers to structures is not.
12777 * Add support for overriding the generation of SSL/TLS session IDs.
12778 These callbacks can be registered either in an SSL_CTX or per SSL.
12779 The purpose of this is to allow applications to control, if they wish,
12780 the arbitrary values chosen for use as session IDs, particularly as it
12781 can be useful for session caching in multiple-server environments. A
12782 command-line switch for testing this (and any client code that wishes
12783 to use such a feature) has been added to "s_server".
12785 *Geoff Thorpe, Lutz Jaenicke*
12787 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12788 of the form `#if defined(...) || defined(...) || ...` and
12789 `#if !defined(...) && !defined(...) && ...`. This also avoids
12790 the growing number of special cases it was previously handling.
12794 * Make all configuration macros available for application by making
12795 sure they are available in opensslconf.h, by giving them names starting
12796 with `OPENSSL_` to avoid conflicts with other packages and by making
12797 sure e_os2.h will cover all platform-specific cases together with
12799 Additionally, it is now possible to define configuration/platform-
12800 specific names (called "system identities"). In the C code, these
12801 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12802 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12803 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12808 * New option -set_serial to 'req' and 'x509' this allows the serial
12809 number to use to be specified on the command line. Previously self
12810 signed certificates were hard coded with serial number 0 and the
12811 CA options of 'x509' had to use a serial number in a file which was
12816 * New options to 'ca' utility to support V2 CRL entry extensions.
12817 Currently CRL reason, invalidity date and hold instruction are
12818 supported. Add new CRL extensions to V3 code and some new objects.
12822 * New function EVP_CIPHER_CTX_set_padding() this is used to
12823 disable standard block padding (aka PKCS#5 padding) in the EVP
12824 API, which was previously mandatory. This means that the data is
12825 not padded in any way and so the total length much be a multiple
12826 of the block size, otherwise an error occurs.
12830 * Initial (incomplete) OCSP SSL support.
12834 * New function OCSP_parse_url(). This splits up a URL into its host,
12835 port and path components: primarily to parse OCSP URLs. New -url
12836 option to ocsp utility.
12840 * New nonce behavior. The return value of OCSP_check_nonce() now
12841 reflects the various checks performed. Applications can decide
12842 whether to tolerate certain situations such as an absent nonce
12843 in a response when one was present in a request: the ocsp application
12844 just prints out a warning. New function OCSP_add1_basic_nonce()
12845 this is to allow responders to include a nonce in a response even if
12846 the request is nonce-less.
12850 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12851 skipped when using openssl x509 multiple times on a single input file,
12852 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12856 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12857 set string type: to handle setting ASN1_TIME structures. Fix ca
12858 utility to correctly initialize revocation date of CRLs.
12862 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12863 the clients preferred ciphersuites and rather use its own preferences.
12864 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12865 Internet Explorer by ensuring unchanged hash method during stepup.
12866 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12870 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12871 to aes and add a new 'exist' option to print out symbols that don't
12876 * Additional options to ocsp utility to allow flags to be set and
12877 additional certificates supplied.
12881 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12882 OCSP client a number of certificate to only verify the response
12887 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12888 handle the new API. Currently only ECB, CBC modes supported. Add new
12891 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12892 Encryption Standard (AES) Ciphersuites for Transport Layer
12893 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12894 not enabled by default and were not part of the "ALL" ciphersuite
12895 alias because they were not yet official; they could be
12896 explicitly requested by specifying the "AESdraft" ciphersuite
12897 group alias. In the final release of OpenSSL 0.9.7, the group
12898 alias is called "AES" and is part of "ALL".)
12900 *Ben Laurie, Steve Henson, Bodo Moeller*
12902 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12903 request to response.
12907 * Functions for OCSP responders. OCSP_request_onereq_count(),
12908 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12909 extract information from a certificate request. OCSP_response_create()
12910 creates a response and optionally adds a basic response structure.
12911 OCSP_basic_add1_status() adds a complete single response to a basic
12912 response and returns the OCSP_SINGLERESP structure just added (to allow
12913 extensions to be included for example). OCSP_basic_add1_cert() adds a
12914 certificate to a basic response and OCSP_basic_sign() signs a basic
12915 response with various flags. New helper functions ASN1_TIME_check()
12916 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12917 (converts ASN1_TIME to GeneralizedTime).
12921 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12922 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12923 structure from a certificate. X509_pubkey_digest() digests the public_key
12924 contents: this is used in various key identifiers.
12928 * Make sk_sort() tolerate a NULL argument.
12930 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12932 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12933 passed by the function are trusted implicitly. If any of them signed the
12934 response then it is assumed to be valid and is not verified.
12938 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12939 to data. This was previously part of the PKCS7 ASN1 code. This
12940 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12941 *Steve Henson, reported by Kenneth R. Robinette
12942 <support@securenetterm.com>*
12944 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12945 routines: without these tracing memory leaks is very painful.
12946 Fix leaks in PKCS12 and PKCS7 routines.
12950 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12951 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12952 effectively meant GeneralizedTime would never be used. Now it
12953 is initialised to -1 but X509_time_adj() now has to check the value
12954 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12955 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12956 *Steve Henson, reported by Kenneth R. Robinette
12957 <support@securenetterm.com>*
12959 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12960 result in a zero length in the ASN1_INTEGER structure which was
12961 not consistent with the structure when d2i_ASN1_INTEGER() was used
12962 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12963 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12964 where it did not print out a minus for negative ASN1_INTEGER.
12968 * Add summary printout to ocsp utility. The various functions which
12969 convert status values to strings have been renamed to:
12970 OCSP_response_status_str(), OCSP_cert_status_str() and
12971 OCSP_crl_reason_str() and are no longer static. New options
12972 to verify nonce values and to disable verification. OCSP response
12973 printout format cleaned up.
12977 * Add additional OCSP certificate checks. These are those specified
12978 in RFC2560. This consists of two separate checks: the CA of the
12979 certificate being checked must either be the OCSP signer certificate
12980 or the issuer of the OCSP signer certificate. In the latter case the
12981 OCSP signer certificate must contain the OCSP signing extended key
12982 usage. This check is performed by attempting to match the OCSP
12983 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12984 in the OCSP_CERTID structures of the response.
12988 * Initial OCSP certificate verification added to OCSP_basic_verify()
12989 and related routines. This uses the standard OpenSSL certificate
12990 verify routines to perform initial checks (just CA validity) and
12991 to obtain the certificate chain. Then additional checks will be
12992 performed on the chain. Currently the root CA is checked to see
12993 if it is explicitly trusted for OCSP signing. This is used to set
12994 a root CA as a global signing root: that is any certificate that
12995 chains to that CA is an acceptable OCSP signing certificate.
12999 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13000 extensions from a separate configuration file.
13001 As when reading extensions from the main configuration file,
13002 the '-extensions ...' option may be used for specifying the
13005 *Massimiliano Pala <madwolf@comune.modena.it>*
13007 * New OCSP utility. Allows OCSP requests to be generated or
13008 read. The request can be sent to a responder and the output
13009 parsed, outputted or printed in text form. Not complete yet:
13010 still needs to check the OCSP response validity.
13014 * New subcommands for 'openssl ca':
13015 `openssl ca -status <serial>` prints the status of the cert with
13016 the given serial number (according to the index file).
13017 `openssl ca -updatedb` updates the expiry status of certificates
13020 *Massimiliano Pala <madwolf@comune.modena.it>*
13022 * New '-newreq-nodes' command option to CA.pl. This is like
13023 '-newreq', but calls 'openssl req' with the '-nodes' option
13024 so that the resulting key is not encrypted.
13026 *Damien Miller <djm@mindrot.org>*
13028 * New configuration for the GNU Hurd.
13030 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13032 * Initial code to implement OCSP basic response verify. This
13033 is currently incomplete. Currently just finds the signer's
13034 certificate and verifies the signature on the response.
13038 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13039 value of OPENSSLDIR. This is available via the new '-d' option
13040 to 'openssl version', and is also included in 'openssl version -a'.
13044 * Allowing defining memory allocation callbacks that will be given
13045 file name and line number information in additional arguments
13046 (a `const char*` and an int). The basic functionality remains, as
13047 well as the original possibility to just replace malloc(),
13048 realloc() and free() by functions that do not know about these
13049 additional arguments. To register and find out the current
13050 settings for extended allocation functions, the following
13051 functions are provided:
13053 CRYPTO_set_mem_ex_functions
13054 CRYPTO_set_locked_mem_ex_functions
13055 CRYPTO_get_mem_ex_functions
13056 CRYPTO_get_locked_mem_ex_functions
13058 These work the same way as CRYPTO_set_mem_functions and friends.
13059 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13060 extended allocation function is enabled.
13061 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13062 a conventional allocation function is enabled.
13064 *Richard Levitte, Bodo Moeller*
13066 * Finish off removing the remaining LHASH function pointer casts.
13067 There should no longer be any prototype-casting required when using
13068 the LHASH abstraction, and any casts that remain are "bugs". See
13069 the callback types and macros at the head of lhash.h for details
13070 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13074 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13075 If /dev/[u]random devices are not available or do not return enough
13076 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13078 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13079 /etc/entropy will be queried once each in this sequence, querying stops
13080 when enough entropy was collected without querying more sockets.
13084 * Change the Unix RAND_poll() variant to be able to poll several
13085 random devices, as specified by DEVRANDOM, until a sufficient amount
13086 of data has been collected. We spend at most 10 ms on each file
13087 (select timeout) and read in non-blocking mode. DEVRANDOM now
13088 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13089 (previously it was just the string "/dev/urandom"), so on typical
13090 platforms the 10 ms delay will never occur.
13091 Also separate out the Unix variant to its own file, rand_unix.c.
13092 For VMS, there's a currently-empty rand_vms.c.
13096 * Move OCSP client related routines to ocsp_cl.c. These
13097 provide utility functions which an application needing
13098 to issue a request to an OCSP responder and analyse the
13099 response will typically need: as opposed to those which an
13100 OCSP responder itself would need which will be added later.
13102 OCSP_request_sign() signs an OCSP request with an API similar
13103 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13104 response. OCSP_response_get1_basic() extracts basic response
13105 from response. OCSP_resp_find_status(): finds and extracts status
13106 information from an OCSP_CERTID structure (which will be created
13107 when the request structure is built). These are built from lower
13108 level functions which work on OCSP_SINGLERESP structures but
13109 won't normally be used unless the application wishes to examine
13110 extensions in the OCSP response for example.
13112 Replace nonce routines with a pair of functions.
13113 OCSP_request_add1_nonce() adds a nonce value and optionally
13114 generates a random value. OCSP_check_nonce() checks the
13115 validity of the nonce in an OCSP response.
13119 * Change function OCSP_request_add() to OCSP_request_add0_id().
13120 This doesn't copy the supplied OCSP_CERTID and avoids the
13121 need to free up the newly created id. Change return type
13122 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13123 This can then be used to add extensions to the request.
13124 Deleted OCSP_request_new(), since most of its functionality
13125 is now in OCSP_REQUEST_new() (and the case insensitive name
13126 clash) apart from the ability to set the request name which
13127 will be added elsewhere.
13131 * Update OCSP API. Remove obsolete extensions argument from
13132 various functions. Extensions are now handled using the new
13133 OCSP extension code. New simple OCSP HTTP function which
13134 can be used to send requests and parse the response.
13138 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13139 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13140 uses the special reorder version of SET OF to sort the attributes
13141 and reorder them to match the encoded order. This resolves a long
13142 standing problem: a verify on a PKCS7 structure just after signing
13143 it used to fail because the attribute order did not match the
13144 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13145 it uses the received order. This is necessary to tolerate some broken
13146 software that does not order SET OF. This is handled by encoding
13147 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13148 to produce the required SET OF.
13152 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13153 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13154 files to get correct declarations of the ASN.1 item variables.
13158 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13159 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13160 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13161 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13162 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13163 ASN1_ITEM and no wrapper functions.
13167 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13168 replace the old function pointer based I/O routines. Change most of
13169 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13173 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13174 lines, recognize more "algorithms" that can be deselected, and make
13175 it complain about algorithm deselection that isn't recognised.
13179 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13180 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13181 to use new functions. Add NO_ASN1_OLD which can be set to remove
13182 some old style ASN1 functions: this can be used to determine if old
13183 code will still work when these eventually go away.
13187 * New extension functions for OCSP structures, these follow the
13188 same conventions as certificates and CRLs.
13192 * New function X509V3_add1_i2d(). This automatically encodes and
13193 adds an extension. Its behaviour can be customised with various
13194 flags to append, replace or delete. Various wrappers added for
13195 certificates and CRLs.
13199 * Fix to avoid calling the underlying ASN1 print routine when
13200 an extension cannot be parsed. Correct a typo in the
13201 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13205 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13206 entries for variables.
13210 * Add functionality to `apps/openssl.c` for detecting locking
13211 problems: As the program is single-threaded, all we have
13212 to do is register a locking callback using an array for
13213 storing which locks are currently held by the program.
13217 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13218 SSL_get_ex_data_X509_STORE_idx(), which is used in
13219 ssl_verify_cert_chain() and thus can be called at any time
13220 during TLS/SSL handshakes so that thread-safety is essential.
13221 Unfortunately, the ex_data design is not at all suited
13222 for multi-threaded use, so it probably should be abolished.
13226 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13228 *Broadcom, tweaked and integrated by Geoff Thorpe*
13230 * Move common extension printing code to new function
13231 X509V3_print_extensions(). Reorganise OCSP print routines and
13232 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13236 * New function X509_signature_print() to remove duplication in some
13241 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13242 set (this was treated exactly the same as SET OF previously). This
13243 is used to reorder the STACK representing the structure to match the
13244 encoding. This will be used to get round a problem where a PKCS7
13245 structure which was signed could not be verified because the STACK
13246 order did not reflect the encoded order.
13250 * Reimplement the OCSP ASN1 module using the new code.
13254 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13255 for its ASN1 operations. The old style function pointers still exist
13256 for now but they will eventually go away.
13260 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13261 completely replaces the old ASN1 functionality with a table driven
13262 encoder and decoder which interprets an ASN1_ITEM structure describing
13263 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13264 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13265 has also been converted to the new form.
13269 * Change BN_mod_exp_recp so that negative moduli are tolerated
13270 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13271 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13272 for negative moduli.
13276 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13277 of not touching the result's sign bit.
13281 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13286 * Changed the LHASH code to use prototypes for callbacks, and created
13287 macros to declare and implement thin (optionally static) functions
13288 that provide type-safety and avoid function pointer casting for the
13289 type-specific callbacks.
13293 * Added Kerberos Cipher Suites to be used with TLS, as written in
13295 *Veers Staats <staatsvr@asc.hpc.mil>,
13296 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13298 * Reformat the FAQ so the different questions and answers can be divided
13299 in sections depending on the subject.
13303 * Have the zlib compression code load ZLIB.DLL dynamically under
13308 * New function BN_mod_sqrt for computing square roots modulo a prime
13309 (using the probabilistic Tonelli-Shanks algorithm unless
13310 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13311 be handled deterministically).
13313 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13315 * Make BN_mod_inverse faster by explicitly handling small quotients
13316 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13317 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13321 * New function BN_kronecker.
13325 * Fix BN_gcd so that it works on negative inputs; the result is
13326 positive unless both parameters are zero.
13327 Previously something reasonably close to an infinite loop was
13328 possible because numbers could be growing instead of shrinking
13329 in the implementation of Euclid's algorithm.
13333 * Fix BN_is_word() and BN_is_one() macros to take into account the
13334 sign of the number in question.
13336 Fix BN_is_word(a,w) to work correctly for w == 0.
13338 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13339 because its test if the absolute value of 'a' equals 'w'.
13340 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13341 it exists mostly for use in the implementations of BN_is_zero(),
13342 BN_is_one(), and BN_is_word().
13346 * New function BN_swap.
13350 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13351 the exponentiation functions are more likely to produce reasonable
13352 results on negative inputs.
13356 * Change BN_mod_mul so that the result is always non-negative.
13357 Previously, it could be negative if one of the factors was negative;
13358 I don't think anyone really wanted that behaviour.
13362 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13363 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13364 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13365 and add new functions:
13374 BN_mod_lshift1_quick
13376 BN_mod_lshift_quick
13378 These functions always generate non-negative results.
13380 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13381 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13383 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13384 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13385 be reduced modulo `m`.
13387 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13390 The following entry accidentally appeared in the CHANGES file
13391 distributed with OpenSSL 0.9.7. The modifications described in
13392 it do *not* apply to OpenSSL 0.9.7.
13394 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13395 was actually never needed) and in BN_mul(). The removal in BN_mul()
13396 required a small change in bn_mul_part_recursive() and the addition
13397 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13398 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13399 bn_sub_words() and bn_add_words() except they take arrays with
13405 * In 'openssl passwd', verify passwords read from the terminal
13406 unless the '-salt' option is used (which usually means that
13407 verification would just waste user's time since the resulting
13408 hash is going to be compared with some given password hash)
13409 or the new '-noverify' option is used.
13411 This is an incompatible change, but it does not affect
13412 non-interactive use of 'openssl passwd' (passwords on the command
13413 line, '-stdin' option, '-in ...' option) and thus should not
13414 cause any problems.
13418 * Remove all references to RSAref, since there's no more need for it.
13422 * Make DSO load along a path given through an environment variable
13423 (SHLIB_PATH) with shl_load().
13427 * Constify the ENGINE code as a result of BIGNUM constification.
13428 Also constify the RSA code and most things related to it. In a
13429 few places, most notable in the depth of the ASN.1 code, ugly
13430 casts back to non-const were required (to be solved at a later
13435 * Make it so the openssl application has all engines loaded by default.
13439 * Constify the BIGNUM routines a little more.
13443 * Add the following functions:
13445 ENGINE_load_cswift()
13447 ENGINE_load_atalla()
13448 ENGINE_load_nuron()
13449 ENGINE_load_builtin_engines()
13451 That way, an application can itself choose if external engines that
13452 are built-in in OpenSSL shall ever be used or not. The benefit is
13453 that applications won't have to be linked with libdl or other dso
13454 libraries unless it's really needed.
13456 Changed 'openssl engine' to load all engines on demand.
13457 Changed the engine header files to avoid the duplication of some
13458 declarations (they differed!).
13462 * 'openssl engine' can now list capabilities.
13466 * Better error reporting in 'openssl engine'.
13470 * Never call load_dh_param(NULL) in s_server.
13474 * Add engine application. It can currently list engines by name and
13475 identity, and test if they are actually available.
13479 * Improve RPM specification file by forcing symbolic linking and making
13480 sure the installed documentation is also owned by root.root.
13482 *Damien Miller <djm@mindrot.org>*
13484 * Give the OpenSSL applications more possibilities to make use of
13485 keys (public as well as private) handled by engines.
13489 * Add OCSP code that comes from CertCo.
13493 * Add VMS support for the Rijndael code.
13497 * Added untested support for Nuron crypto accelerator.
13501 * Add support for external cryptographic devices. This code was
13502 previously distributed separately as the "engine" branch.
13504 *Geoff Thorpe, Richard Levitte*
13506 * Rework the filename-translation in the DSO code. It is now possible to
13507 have far greater control over how a "name" is turned into a filename
13508 depending on the operating environment and any oddities about the
13509 different shared library filenames on each system.
13513 * Support threads on FreeBSD-elf in Configure.
13517 * Fix for SHA1 assembly problem with MASM: it produces
13518 warnings about corrupt line number information when assembling
13519 with debugging information. This is caused by the overlapping
13522 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13525 NCONF_get_number() has no error checking at all. As a replacement,
13526 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13527 promoted strongly. The old NCONF_get_number is kept around for
13528 binary backward compatibility.
13529 Make it possible for methods to load from something other than a BIO,
13530 by providing a function pointer that is given a name instead of a BIO.
13531 For example, this could be used to load configuration data from an
13536 * Fix for non blocking accept BIOs. Added new I/O special reason
13537 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13538 with non blocking I/O was not possible because no retry code was
13539 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13544 * Added the beginnings of Rijndael support.
13548 * Fix for bug in DirectoryString mask setting. Add support for
13549 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13550 to allow certificate printing to more controllable, additional
13551 'certopt' option to 'x509' to allow new printing options to be
13556 * Clean old EAY MD5 hack from e_os.h.
13560 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13562 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13563 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13565 *Joe Orton, Steve Henson*
13567 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13569 * Fix additional bug revealed by the NISCC test suite:
13571 Stop bug triggering large recursion when presented with
13572 certain ASN.1 tags ([CVE-2003-0851])
13576 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13578 * Fix various bugs revealed by running the NISCC test suite:
13580 Stop out of bounds reads in the ASN1 code when presented with
13581 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13583 If verify callback ignores invalid public key errors don't try to check
13584 certificate signature with the NULL public key.
13588 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13589 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13594 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13595 extra data after the compression methods not only for TLS 1.0
13596 but also for SSL 3.0 (as required by the specification).
13598 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13600 * Change X509_certificate_type() to mark the key as exported/exportable
13601 when it's 512 *bits* long, not 512 bytes.
13605 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13607 * Countermeasure against the Klima-Pokorny-Rosa extension of
13608 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13609 a protocol version number mismatch like a decryption error
13610 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13614 * Turn on RSA blinding by default in the default implementation
13615 to avoid a timing attack. Applications that don't want it can call
13616 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13617 They would be ill-advised to do so in most cases.
13619 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13621 * Change RSA blinding code so that it works when the PRNG is not
13622 seeded (in this case, the secret RSA exponent is abused as
13623 an unpredictable seed -- if it is not unpredictable, there
13624 is no point in blinding anyway). Make RSA blinding thread-safe
13625 by remembering the creator's thread ID in rsa->blinding and
13626 having all other threads use local one-time blinding factors
13627 (this requires more computation than sharing rsa->blinding, but
13628 avoids excessive locking; and if an RSA object is not shared
13629 between threads, blinding will still be very fast).
13633 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13635 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13636 via timing by performing a MAC computation even if incorrect
13637 block cipher padding has been found. This is a countermeasure
13638 against active attacks where the attacker has to distinguish
13639 between bad padding and a MAC verification error. ([CVE-2003-0078])
13641 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13642 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13643 Martin Vuagnoux (EPFL, Ilion)*
13645 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13647 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13648 memory from its contents. This is done with a counter that will
13649 place alternating values in each byte. This can be used to solve
13650 two issues: 1) the removal of calls to memset() by highly optimizing
13651 compilers, and 2) cleansing with other values than 0, since those can
13652 be read through on certain media, for example a swap space on disk.
13656 * Bugfix: client side session caching did not work with external caching,
13657 because the session->cipher setting was not restored when reloading
13658 from the external cache. This problem was masked, when
13659 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13660 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13664 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13665 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13667 *Zeev Lieber <zeev-l@yahoo.com>*
13669 * Undo an undocumented change introduced in 0.9.6e which caused
13670 repeated calls to OpenSSL_add_all_ciphers() and
13671 OpenSSL_add_all_digests() to be ignored, even after calling
13676 * Change the default configuration reader to deal with last line not
13677 being properly terminated.
13681 * Change X509_NAME_cmp() so it applies the special rules on handling
13682 DN values that are of type PrintableString, as well as RDNs of type
13683 emailAddress where the value has the type ia5String.
13685 *stefank@valicert.com via Richard Levitte*
13687 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13688 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13689 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13690 the bitwise-OR of the two for use by the majority of applications
13691 wanting this behaviour, and update the docs. The documented
13692 behaviour and actual behaviour were inconsistent and had been
13693 changing anyway, so this is more a bug-fix than a behavioural
13696 *Geoff Thorpe, diagnosed by Nadav Har'El*
13698 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13699 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13703 * Fix initialization code race conditions in
13704 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13705 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13706 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13707 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13708 ssl2_get_cipher_by_char(),
13709 ssl3_get_cipher_by_char().
13711 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13713 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13714 the cached sessions are flushed, as the remove_cb() might use ex_data
13715 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13716 (see [openssl.org #212]).
13718 *Geoff Thorpe, Lutz Jaenicke*
13720 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13721 length, instead of the encoding length to d2i_ASN1_OBJECT.
13725 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13727 * [In 0.9.6g-engine release:]
13728 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13730 *Lynn Gazis <lgazis@rainbow.com>*
13732 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13734 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13735 and get fix the header length calculation.
13736 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13737 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13739 * Use proper error handling instead of 'assertions' in buffer
13740 overflow checks added in 0.9.6e. This prevents DoS (the
13741 assertions could call abort()).
13743 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13745 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13747 * Add various sanity checks to asn1_get_length() to reject
13748 the ASN1 length bytes if they exceed sizeof(long), will appear
13749 negative or the content length exceeds the length of the
13752 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13754 * Fix cipher selection routines: ciphers without encryption had no flags
13755 for the cipher strength set and where therefore not handled correctly
13756 by the selection routines (PR #130).
13760 * Fix EVP_dsa_sha macro.
13765 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13766 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13767 that was added in OpenSSL 0.9.6d.
13769 As the countermeasure turned out to be incompatible with some
13770 broken SSL implementations, the new option is part of SSL_OP_ALL.
13771 SSL_OP_ALL is usually employed when compatibility with weird SSL
13772 implementations is desired (e.g. '-bugs' option to 's_client' and
13773 's_server'), so the new option is automatically set in many
13778 * Changes in security patch:
13780 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13781 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13782 Air Force Materiel Command, USAF, under agreement number
13785 * Add various sanity checks to asn1_get_length() to reject
13786 the ASN1 length bytes if they exceed sizeof(long), will appear
13787 negative or the content length exceeds the length of the
13788 supplied buffer. ([CVE-2002-0659])
13790 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13792 * Assertions for various potential buffer overflows, not known to
13793 happen in practice.
13795 *Ben Laurie (CHATS)*
13797 * Various temporary buffers to hold ASCII versions of integers were
13798 too small for 64 bit platforms. ([CVE-2002-0655])
13799 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13801 * Remote buffer overflow in SSL3 protocol - an attacker could
13802 supply an oversized session ID to a client. ([CVE-2002-0656])
13804 *Ben Laurie (CHATS)*
13806 * Remote buffer overflow in SSL2 protocol - an attacker could
13807 supply an oversized client master key. ([CVE-2002-0656])
13809 *Ben Laurie (CHATS)*
13811 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13813 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13814 encoded as NULL) with id-dsa-with-sha1.
13816 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13818 * Check various `X509_...()` return values in `apps/req.c`.
13820 *Nils Larsch <nla@trustcenter.de>*
13822 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13823 an end-of-file condition would erroneously be flagged, when the CRLF
13824 was just at the end of a processed block. The bug was discovered when
13825 processing data through a buffering memory BIO handing the data to a
13826 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13827 <ptsekov@syntrex.com> and Nedelcho Stanev.
13831 * Implement a countermeasure against a vulnerability recently found
13832 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13833 before application data chunks to avoid the use of known IVs
13834 with data potentially chosen by the attacker.
13838 * Fix length checks in ssl3_get_client_hello().
13842 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13843 to prevent ssl3_read_internal() from incorrectly assuming that
13844 ssl3_read_bytes() found application data while handshake
13845 processing was enabled when in fact s->s3->in_read_app_data was
13846 merely automatically cleared during the initial handshake.
13848 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13850 * Fix object definitions for Private and Enterprise: they were not
13851 recognized in their shortname (=lowercase) representation. Extend
13852 obj_dat.pl to issue an error when using undefined keywords instead
13853 of silently ignoring the problem (Svenning Sorensen
13854 <sss@sss.dnsalias.net>).
13858 * Fix DH_generate_parameters() so that it works for 'non-standard'
13859 generators, i.e. generators other than 2 and 5. (Previously, the
13860 code did not properly initialise the 'add' and 'rem' values to
13861 BN_generate_prime().)
13863 In the new general case, we do not insist that 'generator' is
13864 actually a primitive root: This requirement is rather pointless;
13865 a generator of the order-q subgroup is just as good, if not
13870 * Map new X509 verification errors to alerts. Discovered and submitted by
13871 Tom Wu <tom@arcot.com>.
13875 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13876 returning non-zero before the data has been completely received
13877 when using non-blocking I/O.
13879 *Bodo Moeller; problem pointed out by John Hughes*
13881 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13883 *Ben Laurie, Lutz Jaenicke*
13885 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13886 Yoram Zahavi <YoramZ@gilian.com>).
13890 * Add information about CygWin 1.3 and on, and preserve proper
13891 configuration for the versions before that.
13893 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13895 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13896 check whether we deal with a copy of a session and do not delete from
13897 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13898 <izhar@checkpoint.com>.
13902 * Do not store session data into the internal session cache, if it
13903 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13904 flag is set). Proposed by Aslam <aslam@funk.com>.
13908 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13913 * [In 0.9.6d-engine release:]
13914 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13916 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13918 * Add the configuration target linux-s390x.
13920 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13922 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13923 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13924 variable as an indication that a ClientHello message has been
13925 received. As the flag value will be lost between multiple
13926 invocations of ssl3_accept when using non-blocking I/O, the
13927 function may not be aware that a handshake has actually taken
13928 place, thus preventing a new session from being added to the
13931 To avoid this problem, we now set s->new_session to 2 instead of
13932 using a local variable.
13934 *Lutz Jaenicke, Bodo Moeller*
13936 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13937 if the SSL_R_LENGTH_MISMATCH error is detected.
13939 *Geoff Thorpe, Bodo Moeller*
13941 * New 'shared_ldflag' column in Configure platform table.
13945 * Fix EVP_CIPHER_mode macro.
13947 *"Dan S. Camper" <dan@bti.net>*
13949 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13950 type, we must throw them away by setting rr->length to 0.
13952 *D P Chang <dpc@qualys.com>*
13954 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13956 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13957 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13958 worked incorrectly for those cases where range = `10..._2` and
13959 `3*range` is two bits longer than range.)
13963 * Only add signing time to PKCS7 structures if it is not already
13968 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13969 OBJ_ld_ce should be OBJ_id_ce.
13970 Also some ip-pda OIDs in crypto/objects/objects.txt were
13971 incorrect (cf. RFC 3039).
13973 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13975 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13976 returns early because it has nothing to do.
13978 *Andy Schneider <andy.schneider@bjss.co.uk>*
13980 * [In 0.9.6c-engine release:]
13981 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13983 *Andy Schneider <andy.schneider@bjss.co.uk>*
13985 * [In 0.9.6c-engine release:]
13986 Add support for Cryptographic Appliance's keyserver technology.
13987 (Use engine 'keyclient')
13989 *Cryptographic Appliances and Geoff Thorpe*
13991 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13992 is called via tools/c89.sh because arguments have to be
13993 rearranged (all '-L' options must appear before the first object
13996 *Richard Shapiro <rshapiro@abinitio.com>*
13998 * [In 0.9.6c-engine release:]
13999 Add support for Broadcom crypto accelerator cards, backported
14002 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14004 * [In 0.9.6c-engine release:]
14005 Add support for SureWare crypto accelerator cards from
14006 Baltimore Technologies. (Use engine 'sureware')
14008 *Baltimore Technologies and Mark Cox*
14010 * [In 0.9.6c-engine release:]
14011 Add support for crypto accelerator cards from Accelerated
14012 Encryption Processing, www.aep.ie. (Use engine 'aep')
14014 *AEP Inc. and Mark Cox*
14016 * Add a configuration entry for gcc on UnixWare.
14018 *Gary Benson <gbenson@redhat.com>*
14020 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14021 messages are stored in a single piece (fixed-length part and
14022 variable-length part combined) and fix various bugs found on the way.
14026 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14027 instead. BIO_gethostbyname() does not know what timeouts are
14028 appropriate, so entries would stay in cache even when they have
14030 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14032 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14033 faced with a pathologically small ClientHello fragment that does
14034 not contain client_version: Instead of aborting with an error,
14035 simply choose the highest available protocol version (i.e.,
14036 TLS 1.0 unless it is disabled). In practice, ClientHello
14037 messages are never sent like this, but this change gives us
14038 strictly correct behaviour at least for TLS.
14042 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14043 never resets s->method to s->ctx->method when called from within
14044 one of the SSL handshake functions.
14046 *Bodo Moeller; problem pointed out by Niko Baric*
14048 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14049 (sent using the client's version number) if client_version is
14050 smaller than the protocol version in use. Also change
14051 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14052 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14053 the client will at least see that alert.
14057 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14062 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14063 client receives HelloRequest while in a handshake.
14065 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14067 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14068 should end in 'break', not 'goto end' which circumvents various
14069 cleanups done in state SSL_ST_OK. But session related stuff
14070 must be disabled for SSL_ST_OK in the case that we just sent a
14073 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14074 before just sending a HelloRequest.
14076 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14078 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14079 reveal whether illegal block cipher padding was found or a MAC
14080 verification error occurred. (Neither SSLerr() codes nor alerts
14081 are directly visible to potential attackers, but the information
14082 may leak via logfiles.)
14084 Similar changes are not required for the SSL 2.0 implementation
14085 because the number of padding bytes is sent in clear for SSL 2.0,
14086 and the extra bytes are just ignored. However ssl/s2_pkt.c
14087 failed to verify that the purported number of padding bytes is in
14092 * Add OpenUNIX-8 support including shared libraries
14093 (Boyd Lynn Gerber <gerberb@zenez.com>).
14097 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14098 'wristwatch attack' using huge encoding parameters (cf.
14099 James H. Manger's CRYPTO 2001 paper). Note that the
14100 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14101 encoding parameters and hence was not vulnerable.
14105 * BN_sqr() bug fix.
14107 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14109 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14110 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14111 followed by modular reduction.
14113 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14115 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14116 equivalent based on BN_pseudo_rand() instead of BN_rand().
14120 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14121 This function was broken, as the check for a new client hello message
14122 to handle SGC did not allow these large messages.
14123 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14127 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14131 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14132 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14136 * Rework the configuration and shared library support for Tru64 Unix.
14137 The configuration part makes use of modern compiler features and
14138 still retains old compiler behavior for those that run older versions
14139 of the OS. The shared library support part includes a variant that
14140 uses the RPATH feature, and is available through the special
14141 configuration target "alpha-cc-rpath", which will never be selected
14144 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14146 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14147 with the same message size as in ssl3_get_certificate_request().
14148 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14149 messages might inadvertently be reject as too long.
14151 *Petr Lampa <lampa@fee.vutbr.cz>*
14153 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14157 * Modified SSL library such that the verify_callback that has been set
14158 specifically for an SSL object with SSL_set_verify() is actually being
14159 used. Before the change, a verify_callback set with this function was
14160 ignored and the verify_callback() set in the SSL_CTX at the time of
14161 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14162 to allow the necessary settings.
14166 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14167 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14168 done automatically (in contradiction to the requirements of the C
14169 standard). This made problems when used from OpenSSH.
14173 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14174 dh->length and always used
14176 BN_rand_range(priv_key, dh->p).
14178 BN_rand_range() is not necessary for Diffie-Hellman, and this
14179 specific range makes Diffie-Hellman unnecessarily inefficient if
14180 dh->length (recommended exponent length) is much smaller than the
14181 length of dh->p. We could use BN_rand_range() if the order of
14182 the subgroup was stored in the DH structure, but we only have
14187 BN_rand(priv_key, l, ...)
14189 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14196 RSA_eay_public_encrypt
14197 RSA_eay_private_decrypt
14198 RSA_eay_private_encrypt (signing)
14199 RSA_eay_public_decrypt (signature verification)
14201 (default implementations for RSA_public_encrypt,
14202 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14203 always reject numbers >= n.
14207 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14208 to synchronize access to 'locking_thread'. This is necessary on
14209 systems where access to 'locking_thread' (an 'unsigned long'
14210 variable) is not atomic.
14214 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14215 *before* setting the 'crypto_lock_rand' flag. The previous code had
14216 a race condition if 0 is a valid thread ID.
14218 *Travis Vitek <vitek@roguewave.com>*
14220 * Add support for shared libraries under Irix.
14222 *Albert Chin-A-Young <china@thewrittenword.com>*
14224 * Add configuration option to build on Linux on both big-endian and
14225 little-endian MIPS.
14227 *Ralf Baechle <ralf@uni-koblenz.de>*
14229 * Add the possibility to create shared libraries on HP-UX.
14233 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14235 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14236 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14237 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14238 PRNG state recovery was possible based on the output of
14239 one PRNG request appropriately sized to gain knowledge on
14240 'md' followed by enough consecutive 1-byte PRNG requests
14241 to traverse all of 'state'.
14243 1. When updating 'md_local' (the current thread's copy of 'md')
14244 during PRNG output generation, hash all of the previous
14245 'md_local' value, not just the half used for PRNG output.
14247 2. Make the number of bytes from 'state' included into the hash
14248 independent from the number of PRNG bytes requested.
14250 The first measure alone would be sufficient to avoid
14251 Markku-Juhani's attack. (Actually it had never occurred
14252 to me that the half of 'md_local' used for chaining was the
14253 half from which PRNG output bytes were taken -- I had always
14254 assumed that the secret half would be used.) The second
14255 measure makes sure that additional data from 'state' is never
14256 mixed into 'md_local' in small portions; this heuristically
14257 further strengthens the PRNG.
14261 * Fix crypto/bn/asm/mips3.s.
14265 * When only the key is given to "enc", the IV is undefined. Print out
14266 an error message in this case.
14270 * Handle special case when X509_NAME is empty in X509 printing routines.
14274 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14275 positive and less than q.
14279 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14280 used: it isn't thread safe and the add_lock_callback should handle
14283 *Paul Rose <Paul.Rose@bridge.com>*
14285 * Verify that incoming data obeys the block size in
14286 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14292 *Ulf Möller, Bodo Möller*
14294 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14295 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14296 when fixing the server behaviour for backwards-compatible 'client
14297 hello' messages. (Note that the attack is impractical against
14298 SSL 3.0 and TLS 1.0 anyway because length and version checking
14299 means that the probability of guessing a valid ciphertext is
14300 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14303 Before 0.9.5, the countermeasure (hide the error by generating a
14304 random 'decryption result') did not work properly because
14305 ERR_clear_error() was missing, meaning that SSL_get_error() would
14306 detect the supposedly ignored error.
14308 Both problems are now fixed.
14312 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14313 (previously it was 1024).
14317 * Fix for compatibility mode trust settings: ignore trust settings
14318 unless some valid trust or reject settings are present.
14322 * Fix for blowfish EVP: its a variable length cipher.
14326 * Fix various bugs related to DSA S/MIME verification. Handle missing
14327 parameters in DSA public key structures and return an error in the
14328 DSA routines if parameters are absent.
14332 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14333 in the current directory if neither $RANDFILE nor $HOME was set.
14334 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14335 caused some confusion to Windows users who haven't defined $HOME.
14336 Thus RAND_file_name() is changed again: e_os.h can define a
14337 DEFAULT_HOME, which will be used if $HOME is not set.
14338 For Windows, we use "C:"; on other platforms, we still require
14339 environment variables.
14341 * Move 'if (!initialized) RAND_poll()' into regions protected by
14342 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14343 having multiple threads call RAND_poll() concurrently.
14347 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14348 combination of a flag and a thread ID variable.
14349 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14350 flag), *other* threads can enter ssleay_add_bytes without obeying
14351 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14352 that they do not hold after the first thread unsets add_do_not_lock).
14356 * Change bctest again: '-x' expressions are not available in all
14357 versions of 'test'.
14361 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14363 * Fix a couple of memory leaks in PKCS7_dataDecode()
14365 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14367 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14368 the default extension for executables, if any. Also, make the perl
14369 scripts that use symlink() to test if it really exists and use "cp"
14370 if it doesn't. All this made OpenSSL compilable and installable in
14375 * Fix for asn1_GetSequence() for indefinite length constructed data.
14376 If SEQUENCE is length is indefinite just set c->slen to the total
14377 amount of data available.
14379 *Steve Henson, reported by shige@FreeBSD.org*
14381 *This change does not apply to 0.9.7.*
14383 * Change bctest to avoid here-documents inside command substitution
14384 (workaround for FreeBSD /bin/sh bug).
14385 For compatibility with Ultrix, avoid shell functions (introduced
14386 in the bctest version that searches along $PATH).
14390 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14391 with des_encrypt() defined on some operating systems, like Solaris
14396 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14397 On the Importance of Eliminating Errors in Cryptographic
14398 Computations, J. Cryptology 14 (2001) 2, 101-119,
14399 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14403 * MIPS assembler BIGNUM division bug fix.
14407 * Disabled incorrect Alpha assembler code.
14411 * Fix PKCS#7 decode routines so they correctly update the length
14412 after reading an EOC for the EXPLICIT tag.
14416 *This change does not apply to 0.9.7.*
14418 * Fix bug in PKCS#12 key generation routines. This was triggered
14419 if a 3DES key was generated with a 0 initial byte. Include
14420 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14421 (but broken) behaviour.
14425 * Enhance bctest to search for a working bc along $PATH and print
14428 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14430 * Fix memory leaks in err.c: free err_data string if necessary;
14431 don't write to the wrong index in ERR_set_error_data.
14435 * Implement ssl23_peek (analogous to ssl23_read), which previously
14440 * Replace rdtsc with `_emit` statements for VC++ version 5.
14442 *Jeremy Cooper <jeremy@baymoo.org>*
14444 * Make it possible to reuse SSLv2 sessions.
14448 * In copy_email() check for >= 0 as a return value for
14449 X509_NAME_get_index_by_NID() since 0 is a valid index.
14451 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14453 * Avoid coredump with unsupported or invalid public keys by checking if
14454 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14455 PKCS7_verify() fails with non detached data.
14459 * Don't use getenv in library functions when run as setuid/setgid.
14460 New function OPENSSL_issetugid().
14464 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14465 due to incorrect handling of multi-threading:
14467 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14469 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14471 3. Count how many times MemCheck_off() has been called so that
14472 nested use can be treated correctly. This also avoids
14473 inband-signalling in the previous code (which relied on the
14474 assumption that thread ID 0 is impossible).
14478 * Add "-rand" option also to s_client and s_server.
14482 * Fix CPU detection on Irix 6.x.
14483 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14484 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14486 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14491 *This change does not apply to 0.9.7.*
14493 * Use the cached encoding of an X509_NAME structure rather than
14494 copying it. This is apparently the reason for the libsafe "errors"
14495 but the code is actually correct.
14499 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14500 Bleichenbacher's DSA attack.
14501 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14502 to be set and top=0 forces the highest bit to be set; top=-1 is new
14503 and leaves the highest bit random.
14505 *Ulf Moeller, Bodo Moeller*
14507 * In the `NCONF_...`-based implementations for `CONF_...` queries
14508 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14509 a temporary CONF structure with the data component set to NULL
14510 (which gives segmentation faults in lh_retrieve).
14511 Instead, use NULL for the CONF pointer in CONF_get_string and
14512 CONF_get_number (which may use environment variables) and directly
14513 return NULL from CONF_get_section.
14517 * Fix potential buffer overrun for EBCDIC.
14521 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14522 keyUsage if basicConstraints absent for a CA.
14526 * Make SMIME_write_PKCS7() write mail header values with a format that
14527 is more generally accepted (no spaces before the semicolon), since
14528 some programs can't parse those values properly otherwise. Also make
14529 sure BIO's that break lines after each write do not create invalid
14534 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14535 macros previously used would not encode an empty SEQUENCE OF
14536 and break the signature.
14540 *This change does not apply to 0.9.7.*
14542 * Zero the premaster secret after deriving the master secret in
14547 * Add some EVP_add_digest_alias registrations (as found in
14548 OpenSSL_add_all_digests()) to SSL_library_init()
14549 aka OpenSSL_add_ssl_algorithms(). This provides improved
14550 compatibility with peers using X.509 certificates
14551 with unconventional AlgorithmIdentifier OIDs.
14555 * Fix for Irix with NO_ASM.
14557 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14559 * ./config script fixes.
14561 *Ulf Moeller, Richard Levitte*
14563 * Fix 'openssl passwd -1'.
14567 * Change PKCS12_key_gen_asc() so it can cope with non null
14568 terminated strings whose length is passed in the passlen
14569 parameter, for example from PEM callbacks. This was done
14570 by adding an extra length parameter to asc2uni().
14572 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14574 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14575 call failed, free the DSA structure.
14579 * Fix to uni2asc() to cope with zero length Unicode strings.
14580 These are present in some PKCS#12 files.
14584 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14585 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14586 when writing a 32767 byte record.
14588 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14590 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14591 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14593 (RSA objects have a reference count access to which is protected
14594 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14595 so they are meant to be shared between threads.)
14596 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14597 "Reddie, Steven" <Steven.Reddie@ca.com>*
14599 * Fix a deadlock in CRYPTO_mem_leaks().
14603 * Use better test patterns in bntest.
14607 * rand_win.c fix for Borland C.
14611 * BN_rshift bugfix for n == 0.
14615 * Add a 'bctest' script that checks for some known 'bc' bugs
14616 so that 'make test' does not abort just because 'bc' is broken.
14620 * Store verify_result within SSL_SESSION also for client side to
14621 avoid potential security hole. (Re-used sessions on the client side
14622 always resulted in verify_result==X509_V_OK, not using the original
14623 result of the server certificate verification.)
14627 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14628 SSL3_RT_APPLICATION_DATA, return 0.
14629 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14634 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14635 releases, have been re-implemented by renaming the previous
14636 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14637 and ssl3_read_internal, respectively, and adding 'peek' parameters
14638 to them. The new ssl[23]_{read,peek} functions are calls to
14639 ssl[23]_read_internal with the 'peek' flag set appropriately.
14640 A 'peek' parameter has also been added to ssl3_read_bytes, which
14641 does the actual work for ssl3_read_internal.
14645 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14646 the method-specific "init()" handler. Also clean up ex_data after
14647 calling the method-specific "finish()" handler. Previously, this was
14648 happening the other way round.
14652 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14653 The previous value, 12, was not always sufficient for BN_mod_exp().
14657 * Make sure that shared libraries get the internal name engine with
14658 the full version number and not just 0. This should mark the
14659 shared libraries as not backward compatible. Of course, this should
14660 be changed again when we can guarantee backward binary compatibility.
14664 * Fix typo in get_cert_by_subject() in by_dir.c
14666 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14668 * Rework the system to generate shared libraries:
14670 - Make note of the expected extension for the shared libraries and
14671 if there is a need for symbolic links from for example libcrypto.so.0
14672 to libcrypto.so.0.9.7. There is extended info in Configure for
14675 - Make as few rebuilds of the shared libraries as possible.
14677 - Still avoid linking the OpenSSL programs with the shared libraries.
14679 - When installing, install the shared libraries separately from the
14684 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14686 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14687 and not in SSL_clear because the latter is also used by the
14688 accept/connect functions; previously, the settings made by
14689 SSL_set_read_ahead would be lost during the handshake.
14691 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14693 * Correct util/mkdef.pl to be selective about disabled algorithms.
14694 Previously, it would create entries for disabled algorithms no
14699 * Added several new manual pages for SSL_* function.
14703 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14705 * In ssl23_get_client_hello, generate an error message when faced
14706 with an initial SSL 3.0/TLS record that is too small to contain the
14707 first two bytes of the ClientHello message, i.e. client_version.
14708 (Note that this is a pathologic case that probably has never happened
14709 in real life.) The previous approach was to use the version number
14710 from the record header as a substitute; but our protocol choice
14711 should not depend on that one because it is not authenticated
14712 by the Finished messages.
14716 * More robust randomness gathering functions for Windows.
14718 *Jeffrey Altman <jaltman@columbia.edu>*
14720 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14721 not set then we don't setup the error code for issuer check errors
14722 to avoid possibly overwriting other errors which the callback does
14723 handle. If an application does set the flag then we assume it knows
14724 what it is doing and can handle the new informational codes
14729 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14730 a general "ANY" type, as such it should be able to decode anything
14731 including tagged types. However it didn't check the class so it would
14732 wrongly interpret tagged types in the same way as their universal
14733 counterpart and unknown types were just rejected. Changed so that the
14734 tagged and unknown types are handled in the same way as a SEQUENCE:
14735 that is the encoding is stored intact. There is also a new type
14736 "V_ASN1_OTHER" which is used when the class is not universal, in this
14737 case we have no idea what the actual type is so we just lump them all
14742 * On VMS, stdout may very well lead to a file that is written to
14743 in a record-oriented fashion. That means that every write() will
14744 write a separate record, which will be read separately by the
14745 programs trying to read from it. This can be very confusing.
14747 The solution is to put a BIO filter in the way that will buffer
14748 text until a linefeed is reached, and then write everything a
14749 line at a time, so every record written will be an actual line,
14750 not chunks of lines and not (usually doesn't happen, but I've
14751 seen it once) several lines in one record. BIO_f_linebuffer() is
14754 Currently, it's a VMS-only method, because that's where it has
14755 been tested well enough.
14759 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14760 it can return incorrect results.
14761 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14762 but it was in 0.9.6-beta[12].)
14766 * Disable the check for content being present when verifying detached
14767 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14768 include zero length content when signing messages.
14772 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14773 BIO_ctrl (for BIO pairs).
14777 * Add DSO method for VMS.
14781 * Bug fix: Montgomery multiplication could produce results with the
14786 * Add RPM specification openssl.spec and modify it to build three
14787 packages. The default package contains applications, application
14788 documentation and run-time libraries. The devel package contains
14789 include files, static libraries and function documentation. The
14790 doc package contains the contents of the doc directory. The original
14791 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14795 * Add a large number of documentation files for many SSL routines.
14797 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14799 * Add a configuration entry for Sony News 4.
14801 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14803 * Don't set the two most significant bits to one when generating a
14804 random number < q in the DSA library.
14808 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14809 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14810 the underlying transport is blocking) if a handshake took place.
14811 (The default behaviour is needed by applications such as s_client
14812 and s_server that use select() to determine when to use SSL_read;
14813 but for applications that know in advance when to expect data, it
14814 just makes things more complicated.)
14818 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14823 * Add a few more EBCDIC conditionals that make `req` and `x509`
14824 work better on such systems.
14826 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14828 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14829 Update PKCS12_parse() so it copies the friendlyName and the
14830 keyid to the certificates aux info.
14834 * Fix bug in PKCS7_verify() which caused an infinite loop
14835 if there was more than one signature.
14837 *Sven Uszpelkat <su@celocom.de>*
14839 * Major change in util/mkdef.pl to include extra information
14840 about each symbol, as well as presenting variables as well
14841 as functions. This change means that there's n more need
14842 to rebuild the .num files when some algorithms are excluded.
14846 * Allow the verify time to be set by an application,
14847 rather than always using the current time.
14851 * Phase 2 verify code reorganisation. The certificate
14852 verify code now looks up an issuer certificate by a
14853 number of criteria: subject name, authority key id
14854 and key usage. It also verifies self signed certificates
14855 by the same criteria. The main comparison function is
14856 X509_check_issued() which performs these checks.
14858 Lot of changes were necessary in order to support this
14859 without completely rewriting the lookup code.
14861 Authority and subject key identifier are now cached.
14863 The LHASH 'certs' is X509_STORE has now been replaced
14864 by a STACK_OF(X509_OBJECT). This is mainly because an
14865 LHASH can't store or retrieve multiple objects with
14866 the same hash value.
14868 As a result various functions (which were all internal
14869 use only) have changed to handle the new X509_STORE
14870 structure. This will break anything that messed round
14871 with X509_STORE internally.
14873 The functions X509_STORE_add_cert() now checks for an
14874 exact match, rather than just subject name.
14876 The X509_STORE API doesn't directly support the retrieval
14877 of multiple certificates matching a given criteria, however
14878 this can be worked round by performing a lookup first
14879 (which will fill the cache with candidate certificates)
14880 and then examining the cache for matches. This is probably
14881 the best we can do without throwing out X509_LOOKUP
14882 entirely (maybe later...).
14884 The X509_VERIFY_CTX structure has been enhanced considerably.
14886 All certificate lookup operations now go via a get_issuer()
14887 callback. Although this currently uses an X509_STORE it
14888 can be replaced by custom lookups. This is a simple way
14889 to bypass the X509_STORE hackery necessary to make this
14890 work and makes it possible to use more efficient techniques
14891 in future. A very simple version which uses a simple
14892 STACK for its trusted certificate store is also provided
14893 using X509_STORE_CTX_trusted_stack().
14895 The verify_cb() and verify() callbacks now have equivalents
14896 in the X509_STORE_CTX structure.
14898 X509_STORE_CTX also has a 'flags' field which can be used
14899 to customise the verify behaviour.
14903 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14904 excludes S/MIME capabilities.
14908 * When a certificate request is read in keep a copy of the
14909 original encoding of the signed data and use it when outputting
14910 again. Signatures then use the original encoding rather than
14911 a decoded, encoded version which may cause problems if the
14912 request is improperly encoded.
14916 * For consistency with other BIO_puts implementations, call
14917 buffer_write(b, ...) directly in buffer_puts instead of calling
14920 In BIO_puts, increment b->num_write as in BIO_write.
14922 *Peter.Sylvester@EdelWeb.fr*
14924 * Fix BN_mul_word for the case where the word is 0. (We have to use
14925 BN_zero, we may not return a BIGNUM with an array consisting of
14926 words set to zero.)
14930 * Avoid calling abort() from within the library when problems are
14931 detected, except if preprocessor symbols have been defined
14932 (such as REF_CHECK, BN_DEBUG etc.).
14936 * New openssl application 'rsautl'. This utility can be
14937 used for low level RSA operations. DER public key
14938 BIO/fp routines also added.
14942 * New Configure entry and patches for compiling on QNX 4.
14944 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14946 * A demo state-machine implementation was sponsored by
14947 Nuron (<http://www.nuron.com/>) and is now available in
14948 demos/state_machine.
14952 * New options added to the 'dgst' utility for signature
14953 generation and verification.
14957 * Unrecognized PKCS#7 content types are now handled via a
14958 catch all ASN1_TYPE structure. This allows unsupported
14959 types to be stored as a "blob" and an application can
14960 encode and decode it manually.
14964 * Fix various signed/unsigned issues to make a_strex.c
14965 compile under VC++.
14967 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14969 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14970 length if passed a buffer. ASN1_INTEGER_to_BN failed
14971 if passed a NULL BN and its argument was negative.
14973 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14975 * Modification to PKCS#7 encoding routines to output definite
14976 length encoding. Since currently the whole structures are in
14977 memory there's not real point in using indefinite length
14978 constructed encoding. However if OpenSSL is compiled with
14979 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14983 * Added BIO_vprintf() and BIO_vsnprintf().
14987 * Added more prefixes to parse for in the strings written
14988 through a logging bio, to cover all the levels that are available
14989 through syslog. The prefixes are now:
14991 PANIC, EMERG, EMR => LOG_EMERG
14992 ALERT, ALR => LOG_ALERT
14993 CRIT, CRI => LOG_CRIT
14994 ERROR, ERR => LOG_ERR
14995 WARNING, WARN, WAR => LOG_WARNING
14996 NOTICE, NOTE, NOT => LOG_NOTICE
14997 INFO, INF => LOG_INFO
14998 DEBUG, DBG => LOG_DEBUG
15000 and as before, if none of those prefixes are present at the
15001 beginning of the string, LOG_ERR is chosen.
15003 On Win32, the `LOG_*` levels are mapped according to this:
15005 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15006 LOG_WARNING => EVENTLOG_WARNING_TYPE
15007 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15011 * Made it possible to reconfigure with just the configuration
15012 argument "reconf" or "reconfigure". The command line arguments
15013 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15014 and are retrieved from there when reconfiguring.
15020 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15022 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15026 * The obj_dat.pl script was messing up the sorting of object
15027 names. The reason was that it compared the quoted version
15028 of strings as a result "OCSP" > "OCSP Signing" because
15029 " > SPACE. Changed script to store unquoted versions of
15030 names and add quotes on output. It was also omitting some
15031 names from the lookup table if they were given a default
15032 value (that is if SN is missing it is given the same
15033 value as LN and vice versa), these are now added on the
15034 grounds that if an object has a name we should be able to
15035 look it up. Finally added warning output when duplicate
15036 short or long names are found.
15040 * Changes needed for Tandem NSK.
15042 *Scott Uroff <scott@xypro.com>*
15044 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15045 RSA_padding_check_SSLv23(), special padding was never detected
15046 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15047 version rollback attacks was not effective.
15049 In s23_clnt.c, don't use special rollback-attack detection padding
15050 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15051 client; similarly, in s23_srvr.c, don't do the rollback check if
15052 SSL 2.0 is the only protocol enabled in the server.
15056 * Make it possible to get hexdumps of unprintable data with 'openssl
15057 asn1parse'. By implication, the functions ASN1_parse_dump() and
15058 BIO_dump_indent() are added.
15062 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15063 these print out strings and name structures based on various
15064 flags including RFC2253 support and proper handling of
15065 multibyte characters. Added options to the 'x509' utility
15066 to allow the various flags to be set.
15070 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15071 Also change the functions X509_cmp_current_time() and
15072 X509_gmtime_adj() work with an ASN1_TIME structure,
15073 this will enable certificates using GeneralizedTime in validity
15074 dates to be checked.
15078 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15079 negative public key encodings) on by default,
15080 NO_NEG_PUBKEY_BUG can be set to disable it.
15084 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15085 content octets. An i2c_ASN1_OBJECT is unnecessary because
15086 the encoding can be trivially obtained from the structure.
15090 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15091 not read locks (`CRYPTO_r_[un]lock`).
15095 * A first attempt at creating official support for shared
15096 libraries through configuration. I've kept it so the
15097 default is static libraries only, and the OpenSSL programs
15098 are always statically linked for now, but there are
15099 preparations for dynamic linking in place.
15100 This has been tested on Linux and Tru64.
15104 * Randomness polling function for Win9x, as described in:
15105 Peter Gutmann, Software Generation of Practically Strong
15110 * Fix so PRNG is seeded in req if using an already existing
15115 * New options to smime application. -inform and -outform
15116 allow alternative formats for the S/MIME message including
15117 PEM and DER. The -content option allows the content to be
15118 specified separately. This should allow things like Netscape
15119 form signing output easier to verify.
15123 * Fix the ASN1 encoding of tags using the 'long form'.
15127 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15128 STRING types. These convert content octets to and from the
15129 underlying type. The actual tag and length octets are
15130 already assumed to have been read in and checked. These
15131 are needed because all other string types have virtually
15132 identical handling apart from the tag. By having versions
15133 of the ASN1 functions that just operate on content octets
15134 IMPLICIT tagging can be handled properly. It also allows
15135 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15136 and ASN1_INTEGER are identical apart from the tag.
15140 * Change the handling of OID objects as follows:
15142 - New object identifiers are inserted in objects.txt, following
15143 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15144 - objects.pl is used to process obj_mac.num and create a new
15146 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15149 This is currently kind of a hack, and the perl code in objects.pl
15150 isn't very elegant, but it works as I intended. The simplest way
15151 to check that it worked correctly is to look in obj_dat.h and
15152 check the array nid_objs and make sure the objects haven't moved
15153 around (this is important!). Additions are OK, as well as
15154 consistent name changes.
15158 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15162 * Addition of the command line parameter '-rand file' to 'openssl req'.
15163 The given file adds to whatever has already been seeded into the
15164 random pool through the RANDFILE configuration file option or
15165 environment variable, or the default random state file.
15169 * mkstack.pl now sorts each macro group into lexical order.
15170 Previously the output order depended on the order the files
15171 appeared in the directory, resulting in needless rewriting
15176 * Patches to make OpenSSL compile under Win32 again. Mostly
15177 work arounds for the VC++ problem that it treats func() as
15178 func(void). Also stripped out the parts of mkdef.pl that
15179 added extra typesafe functions: these no longer exist.
15183 * Reorganisation of the stack code. The macros are now all
15184 collected in safestack.h . Each macro is defined in terms of
15185 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15186 DEBUG_SAFESTACK is now handled in terms of function casts,
15187 this has the advantage of retaining type safety without the
15188 use of additional functions. If DEBUG_SAFESTACK is not defined
15189 then the non typesafe macros are used instead. Also modified the
15190 mkstack.pl script to handle the new form. Needs testing to see
15191 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15192 the default if no major problems. Similar behaviour for ASN1_SET_OF
15193 and PKCS12_STACK_OF.
15197 * When some versions of IIS use the 'NET' form of private key the
15198 key derivation algorithm is different. Normally MD5(password) is
15199 used as a 128 bit RC4 key. In the modified case
15200 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15201 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15202 as the old Netscape_RSA functions except they have an additional
15203 'sgckey' parameter which uses the modified algorithm. Also added
15204 an -sgckey command line option to the rsa utility. Thanks to
15205 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15206 algorithm to openssl-dev.
15210 * The evp_local.h macros were using 'c.##kname' which resulted in
15211 invalid expansion on some systems (SCO 5.0.5 for example).
15212 Corrected to 'c.kname'.
15214 *Phillip Porch <root@theporch.com>*
15216 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15217 a STACK of email addresses from a certificate or request, these look
15218 in the subject name and the subject alternative name extensions and
15219 omit any duplicate addresses.
15223 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15224 This makes DSA verification about 2 % faster.
15228 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15229 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15230 plus overhead for 1024 bit moduli).
15231 This makes exponentiations about 0.5 % faster for 1024 bit
15232 exponents (as measured by "openssl speed rsa2048").
15236 * Rename memory handling macros to avoid conflicts with other
15238 Malloc => OPENSSL_malloc
15239 Malloc_locked => OPENSSL_malloc_locked
15240 Realloc => OPENSSL_realloc
15241 Free => OPENSSL_free
15245 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15246 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15250 * CygWin32 support.
15252 *John Jarvie <jjarvie@newsguy.com>*
15254 * The type-safe stack code has been rejigged. It is now only compiled
15255 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15256 by default all type-specific stack functions are "#define"d back to
15257 standard stack functions. This results in more streamlined output
15258 but retains the type-safety checking possibilities of the original
15263 * The STACK code has been cleaned up, and certain type declarations
15264 that didn't make a lot of sense have been brought in line. This has
15265 also involved a cleanup of sorts in safestack.h to more correctly
15266 map type-safe stack functions onto their plain stack counterparts.
15267 This work has also resulted in a variety of "const"ifications of
15268 lots of the code, especially `_cmp` operations which should normally
15269 be prototyped with "const" parameters anyway.
15273 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15274 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15275 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15276 where all of 'md' is used each time the PRNG is used, but 'state'
15277 is used only indexed by a cyclic counter. As entropy may not be
15278 well distributed from the beginning, 'md' is important as a
15279 chaining variable. However, the output function chains only half
15280 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15281 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15282 in all of 'state' being rewritten, with the new values depending
15283 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15287 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15288 the handshake is continued after ssl_verify_cert_chain();
15289 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15290 can lead to 'unexplainable' connection aborts later.
15292 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15294 * Major EVP API cipher revision.
15295 Add hooks for extra EVP features. This allows various cipher
15296 parameters to be set in the EVP interface. Support added for variable
15297 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15298 setting of RC2 and RC5 parameters.
15300 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15303 Remove lots of duplicated code from the EVP library. For example *every*
15304 cipher init() function handles the 'iv' in the same way according to the
15305 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15306 for CFB and OFB modes they zero ctx->num.
15308 New functionality allows removal of S/MIME code RC2 hack.
15310 Most of the routines have the same form and so can be declared in terms
15313 By shifting this to the top level EVP_CipherInit() it can be removed from
15314 all individual ciphers. If the cipher wants to handle IVs or keys
15315 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15318 Change lots of functions like EVP_EncryptUpdate() to now return a
15319 value: although software versions of the algorithms cannot fail
15320 any installed hardware versions can.
15324 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15325 this option is set, tolerate broken clients that send the negotiated
15326 protocol version number instead of the requested protocol version
15331 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15332 i.e. non-zero for export ciphersuites, zero otherwise.
15333 Previous versions had this flag inverted, inconsistent with
15334 rsa_tmp_cb (..._TMP_RSA_CB).
15336 *Bodo Moeller; problem reported by Amit Chopra*
15338 * Add missing DSA library text string. Work around for some IIS
15339 key files with invalid SEQUENCE encoding.
15343 * Add a document (doc/standards.txt) that list all kinds of standards
15344 and so on that are implemented in OpenSSL.
15348 * Enhance c_rehash script. Old version would mishandle certificates
15349 with the same subject name hash and wouldn't handle CRLs at all.
15350 Added -fingerprint option to crl utility, to support new c_rehash
15355 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15359 * Fix for SSL server purpose checking. Server checking was
15360 rejecting certificates which had extended key usage present
15361 but no ssl client purpose.
15363 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15365 * Make PKCS#12 code work with no password. The PKCS#12 spec
15366 is a little unclear about how a blank password is handled.
15367 Since the password in encoded as a BMPString with terminating
15368 double NULL a zero length password would end up as just the
15369 double NULL. However no password at all is different and is
15370 handled differently in the PKCS#12 key generation code. NS
15371 treats a blank password as zero length. MSIE treats it as no
15372 password on export: but it will try both on import. We now do
15373 the same: PKCS12_parse() tries zero length and no password if
15374 the password is set to "" or NULL (NULL is now a valid password:
15375 it wasn't before) as does the pkcs12 application.
15379 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15380 perror when PEM_read_bio_X509_REQ fails, the error message must
15381 be obtained from the error queue.
15385 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15386 it in ERR_remove_state if appropriate, and change ERR_get_state
15387 accordingly to avoid race conditions (this is necessary because
15388 thread_hash is no longer constant once set).
15392 * Bugfix for linux-elf makefile.one.
15396 * RSA_get_default_method() will now cause a default
15397 RSA_METHOD to be chosen if one doesn't exist already.
15398 Previously this was only set during a call to RSA_new()
15399 or RSA_new_method(NULL) meaning it was possible for
15400 RSA_get_default_method() to return NULL.
15404 * Added native name translation to the existing DSO code
15405 that will convert (if the flag to do so is set) filenames
15406 that are sufficiently small and have no path information
15407 into a canonical native form. Eg. "blah" converted to
15408 "libblah.so" or "blah.dll" etc.
15412 * New function ERR_error_string_n(e, buf, len) which is like
15413 ERR_error_string(e, buf), but writes at most 'len' bytes
15414 including the 0 terminator. For ERR_error_string_n, 'buf'
15417 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15419 * CONF library reworked to become more general. A new CONF
15420 configuration file reader "class" is implemented as well as a
15421 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15422 old `CONF_*` functions are still there, but are reimplemented to
15423 work in terms of the new functions. Also, a set of functions
15424 to handle the internal storage of the configuration data is
15425 provided to make it easier to write new configuration file
15426 reader "classes" (I can definitely see something reading a
15427 configuration file in XML format, for example), called `_CONF_*`,
15428 or "the configuration storage API"...
15430 The new configuration file reading functions are:
15432 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15433 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15435 NCONF_default, NCONF_WIN32
15437 NCONF_dump_fp, NCONF_dump_bio
15439 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15440 NCONF_new creates a new CONF object. This works in the same way
15441 as other interfaces in OpenSSL, like the BIO interface.
15442 `NCONF_dump_*` dump the internal storage of the configuration file,
15443 which is useful for debugging. All other functions take the same
15444 arguments as the old `CONF_*` functions with the exception of the
15445 first that must be a `CONF *` instead of a `LHASH *`.
15447 To make it easier to use the new classes with the old `CONF_*` functions,
15448 the function CONF_set_default_method is provided.
15452 * Add '-tls1' option to 'openssl ciphers', which was already
15453 mentioned in the documentation but had not been implemented.
15454 (This option is not yet really useful because even the additional
15455 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15459 * Initial DSO code added into libcrypto for letting OpenSSL (and
15460 OpenSSL-based applications) load shared libraries and bind to
15461 them in a portable way.
15463 *Geoff Thorpe, with contributions from Richard Levitte*
15465 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15467 * Make sure _lrotl and _lrotr are only used with MSVC.
15469 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15470 (the default implementation of RAND_status).
15472 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15473 to '-clrext' (= clear extensions), as intended and documented.
15474 *Bodo Moeller; inconsistency pointed out by Michael Attili
15475 <attili@amaxo.com>*
15477 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15478 was larger than the MD block size.
15480 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15482 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15483 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15484 using the passed key: if the passed key was a private key the result
15485 of X509_print(), for example, would be to print out all the private key
15490 * des_quad_cksum() byte order bug fix.
15491 *Ulf Möller, using the problem description in krb4-0.9.7, where
15492 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15494 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15497 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15499 * For easily testing in shell scripts whether some command
15500 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15501 returns with exit code 0 iff no command of the given name is available.
15502 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15503 the output goes to stdout and nothing is printed to stderr.
15504 Additional arguments are always ignored.
15506 Since for each cipher there is a command of the same name,
15507 the 'no-cipher' compilation switches can be tested this way.
15509 ('openssl no-XXX' is not able to detect pseudo-commands such
15510 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15514 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15518 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15519 is set; it will be thrown away anyway because each handshake creates
15521 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15522 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15523 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15524 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15528 * New s_client option -ign_eof: EOF at stdin is ignored, and
15529 'Q' and 'R' lose their special meanings (quit/renegotiate).
15530 This is part of what -quiet does; unlike -quiet, -ign_eof
15531 does not suppress any output.
15535 * Add compatibility options to the purpose and trust code. The
15536 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15537 accepts a certificate or CA, this was the previous behaviour,
15538 with all the associated security issues.
15540 X509_TRUST_COMPAT is the old trust behaviour: only and
15541 automatically trust self signed roots in certificate store. A
15542 new trust setting X509_TRUST_DEFAULT is used to specify that
15543 a purpose has no associated trust setting and it should instead
15544 use the value in the default purpose.
15548 * Fix the PKCS#8 DSA private key code so it decodes keys again
15549 and fix a memory leak.
15553 * In util/mkerr.pl (which implements 'make errors'), preserve
15554 reason strings from the previous version of the .c file, as
15555 the default to have only downcase letters (and digits) in
15556 automatically generated reasons codes is not always appropriate.
15560 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15561 using strerror. Previously, ERR_reason_error_string() returned
15562 library names as reason strings for SYSerr; but SYSerr is a special
15563 case where small numbers are errno values, not library numbers.
15567 * Add '-dsaparam' option to 'openssl dhparam' application. This
15568 converts DSA parameters into DH parameters. (When creating parameters,
15569 DSA_generate_parameters is used.)
15573 * Include 'length' (recommended exponent length) in C code generated
15574 by 'openssl dhparam -C'.
15578 * The second argument to set_label in perlasm was already being used
15579 so couldn't be used as a "file scope" flag. Moved to third argument
15584 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15585 instead of RAND_bytes for encryption IVs and salts.
15589 * Include RAND_status() into RAND_METHOD instead of implementing
15590 it only for md_rand.c Otherwise replacing the PRNG by calling
15591 RAND_set_rand_method would be impossible.
15595 * Don't let DSA_generate_key() enter an infinite loop if the random
15596 number generation fails.
15600 * New 'rand' application for creating pseudo-random output.
15604 * Added configuration support for Linux/IA64
15606 *Rolf Haberrecker <rolf@suse.de>*
15608 * Assembler module support for Mingw32.
15612 * Shared library support for HPUX (in shlib/).
15614 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15616 * Shared library support for Solaris gcc.
15618 *Lutz Behnke <behnke@trustcenter.de>*
15620 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15622 * PKCS7_encrypt() was adding text MIME headers twice because they
15623 were added manually and by SMIME_crlf_copy().
15627 * In bntest.c don't call BN_rand with zero bits argument.
15629 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15631 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15632 case was implemented. This caused BN_div_recp() to fail occasionally.
15636 * Add an optional second argument to the set_label() in the perl
15637 assembly language builder. If this argument exists and is set
15638 to 1 it signals that the assembler should use a symbol whose
15639 scope is the entire file, not just the current function. This
15640 is needed with MASM which uses the format label:: for this scope.
15642 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15644 * Change the ASN1 types so they are typedefs by default. Before
15645 almost all types were #define'd to ASN1_STRING which was causing
15646 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15651 * Change names of new functions to the new get1/get0 naming
15652 convention: After 'get1', the caller owns a reference count
15653 and has to call `..._free`; 'get0' returns a pointer to some
15654 data structure without incrementing reference counters.
15655 (Some of the existing 'get' functions increment a reference
15656 counter, some don't.)
15657 Similarly, 'set1' and 'add1' functions increase reference
15658 counters or duplicate objects.
15662 * Allow for the possibility of temp RSA key generation failure:
15663 the code used to assume it always worked and crashed on failure.
15667 * Fix potential buffer overrun problem in BIO_printf().
15668 *Ulf Möller, using public domain code by Patrick Powell; problem
15669 pointed out by David Sacerdote <das33@cornell.edu>*
15671 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15672 RAND_egd() and RAND_status(). In the command line application,
15673 the EGD socket can be specified like a seed file using RANDFILE
15678 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15679 Some CAs (e.g. Verisign) distribute certificates in this form.
15683 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15684 list to exclude them. This means that no special compilation option
15685 is needed to use anonymous DH: it just needs to be included in the
15690 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15691 EVP_MD_type. The old functionality is available in a new macro called
15692 EVP_MD_md(). Change code that uses it and update docs.
15696 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15697 where the `void *` argument is replaced by a function pointer argument.
15698 Previously `void *` was abused to point to functions, which works on
15699 many platforms, but is not correct. As these functions are usually
15700 called by macros defined in OpenSSL header files, most source code
15701 should work without changes.
15705 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15706 sections with information on -D... compiler switches used for
15707 compiling the library so that applications can see them. To enable
15708 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15709 must be defined. E.g.,
15710 #define OPENSSL_ALGORITHM_DEFINES
15711 #include <openssl/opensslconf.h>
15712 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15714 *Richard Levitte, Ulf and Bodo Möller*
15716 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15721 * Change the 'other' type in certificate aux info to a STACK_OF
15722 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15723 the required ASN1 format: arbitrary types determined by an OID.
15727 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15728 argument to 'req'. This is not because the function is newer or
15729 better than others it just uses the work 'NEW' in the certificate
15730 request header lines. Some software needs this.
15734 * Reorganise password command line arguments: now passwords can be
15735 obtained from various sources. Delete the PEM_cb function and make
15736 it the default behaviour: i.e. if the callback is NULL and the
15737 usrdata argument is not NULL interpret it as a null terminated pass
15738 phrase. If usrdata and the callback are NULL then the pass phrase
15739 is prompted for as usual.
15743 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15744 the support is automatically enabled. The resulting binaries will
15745 autodetect the card and use it if present.
15747 *Ben Laurie and Compaq Inc.*
15749 * Work around for Netscape hang bug. This sends certificate request
15750 and server done in one record. Since this is perfectly legal in the
15751 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15752 the bugs/SSLv3 entry for more info.
15756 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15760 * Add -rand argument to smime and pkcs12 applications and read/write
15765 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15769 * Add command line password options to the remaining applications.
15773 * Bug fix for BN_div_recp() for numerators with an even number of
15778 * More tests in bntest.c, and changed test_bn output.
15782 * ./config recognizes MacOS X now.
15786 * Bug fix for BN_div() when the first words of num and divisor are
15787 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15791 * Add support for various broken PKCS#8 formats, and command line
15792 options to produce them.
15796 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15797 get temporary BIGNUMs from a BN_CTX.
15801 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15806 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15807 include a #define from the old name to the new. The original intent
15808 was that statically linked binaries could for example just call
15809 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15810 link with digests. This never worked because SSLeay_add_all_digests()
15811 and SSLeay_add_all_ciphers() were in the same source file so calling
15812 one would link with the other. They are now in separate source files.
15816 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15820 * Use a less unusual form of the Miller-Rabin primality test (it used
15821 a binary algorithm for exponentiation integrated into the Miller-Rabin
15822 loop, our standard modexp algorithms are faster).
15826 * Support for the EBCDIC character set completed.
15828 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15830 * Source code cleanups: use const where appropriate, eliminate casts,
15831 use `void *` instead of `char *` in lhash.
15835 * Bugfix: ssl3_send_server_key_exchange was not restartable
15836 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15837 this the server could overwrite ephemeral keys that the client
15842 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15843 using 50 iterations of the Rabin-Miller test.
15845 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15846 iterations of the Rabin-Miller test as required by the appendix
15847 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15848 As BN_is_prime_fasttest includes trial division, DSA parameter
15849 generation becomes much faster.
15851 This implies a change for the callback functions in DSA_is_prime
15852 and DSA_generate_parameters: The callback function is called once
15853 for each positive witness in the Rabin-Miller test, not just
15854 occasionally in the inner loop; and the parameters to the
15855 callback function now provide an iteration count for the outer
15856 loop rather than for the current invocation of the inner loop.
15857 DSA_generate_parameters additionally can call the callback
15858 function with an 'iteration count' of -1, meaning that a
15859 candidate has passed the trial division test (when q is generated
15860 from an application-provided seed, trial division is skipped).
15864 * New function BN_is_prime_fasttest that optionally does trial
15865 division before starting the Rabin-Miller test and has
15866 an additional BN_CTX * argument (whereas BN_is_prime always
15867 has to allocate at least one BN_CTX).
15868 'callback(1, -1, cb_arg)' is called when a number has passed the
15869 trial division stage.
15873 * Fix for bug in CRL encoding. The validity dates weren't being handled
15878 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15882 * New function BN_pseudo_rand().
15886 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15887 bignum version of BN_from_montgomery() with the working code from
15888 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15893 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15894 made it impossible to use the same SSL_SESSION data structure in
15895 SSL2 clients in multiple threads.
15899 * The return value of RAND_load_file() no longer counts bytes obtained
15900 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15901 to seed the PRNG (previously an explicit byte count was required).
15903 *Ulf Möller, Bodo Möller*
15905 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15906 used `char *` instead of `void *` and had casts all over the place.
15910 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15914 * Retain source code compatibility for BN_prime_checks macro:
15915 BN_is_prime(..., BN_prime_checks, ...) now uses
15916 BN_prime_checks_for_size to determine the appropriate number of
15917 Rabin-Miller iterations.
15921 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15922 DH_CHECK_P_NOT_SAFE_PRIME.
15923 (Check if this is true? OpenPGP calls them "strong".)
15927 * Merge the functionality of "dh" and "gendh" programs into a new program
15928 "dhparam". The old programs are retained for now but will handle DH keys
15929 (instead of parameters) in future.
15933 * Make the ciphers, s_server and s_client programs check the return values
15934 when a new cipher list is set.
15938 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15939 ciphers. Before when the 56bit ciphers were enabled the sorting was
15942 The syntax for the cipher sorting has been extended to support sorting by
15943 cipher-strength (using the strength_bits hard coded in the tables).
15944 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15946 Fix a bug in the cipher-command parser: when supplying a cipher command
15947 string with an "undefined" symbol (neither command nor alphanumeric
15948 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15949 an error is flagged.
15951 Due to the strength-sorting extension, the code of the
15952 ssl_create_cipher_list() function was completely rearranged. I hope that
15953 the readability was also increased :-)
15955 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15957 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15958 for the first serial number and places 2 in the serial number file. This
15959 avoids problems when the root CA is created with serial number zero and
15960 the first user certificate has the same issuer name and serial number
15965 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15966 the new code. Add documentation for this stuff.
15970 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15971 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15972 structures and behave in an analogous way to the X509v3 functions:
15973 they shouldn't be called directly but wrapper functions should be used
15976 So we also now have some wrapper functions that call the X509at functions
15977 when passed certificate requests. (TO DO: similar things can be done with
15978 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15979 things. Some of these need some d2i or i2d and print functionality
15980 because they handle more complex structures.)
15984 * Add missing #ifndefs that caused missing symbols when building libssl
15985 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15986 NO_RSA in `ssl/s2*.c`.
15988 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15990 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15991 has a return value which indicates the quality of the random data
15992 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15993 error queue. New function RAND_pseudo_bytes() generates output that is
15994 guaranteed to be unique but not unpredictable. RAND_add is like
15995 RAND_seed, but takes an extra argument for an entropy estimate
15996 (RAND_seed always assumes full entropy).
16000 * Do more iterations of Rabin-Miller probable prime test (specifically,
16001 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16002 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16003 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16004 false-positive rate of at most 2^-80 for random input.
16008 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16012 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16013 in the 0.9.5 release), this returns the chain
16014 from an X509_CTX structure with a dup of the stack and all
16015 the X509 reference counts upped: so the stack will exist
16016 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16019 Also make SSL_SESSION_print() print out the verify return
16024 * Add manpage for the pkcs12 command. Also change the default
16025 behaviour so MAC iteration counts are used unless the new
16026 -nomaciter option is used. This improves file security and
16027 only older versions of MSIE (4.0 for example) need it.
16031 * Honor the no-xxx Configure options when creating .DEF files.
16035 * Add PKCS#10 attributes to field table: challengePassword,
16036 unstructuredName and unstructuredAddress. These are taken from
16037 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16038 international characters are used.
16040 More changes to X509_ATTRIBUTE code: allow the setting of types
16041 based on strings. Remove the 'loc' parameter when adding
16042 attributes because these will be a SET OF encoding which is sorted
16047 * Initial changes to the 'req' utility to allow request generation
16048 automation. This will allow an application to just generate a template
16049 file containing all the field values and have req construct the
16052 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16053 used all over the place including certificate requests and PKCS#7
16054 structures. They are currently handled manually where necessary with
16055 some primitive wrappers for PKCS#7. The new functions behave in a
16056 manner analogous to the X509 extension functions: they allow
16057 attributes to be looked up by NID and added.
16059 Later something similar to the X509V3 code would be desirable to
16060 automatically handle the encoding, decoding and printing of the
16061 more complex types. The string types like challengePassword can
16062 be handled by the string table functions.
16064 Also modified the multi byte string table handling. Now there is
16065 a 'global mask' which masks out certain types. The table itself
16066 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16067 is useful when for example there is only one permissible type
16068 (as in countryName) and using the mask might result in no valid
16073 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16074 SSL_get_peer_finished to allow applications to obtain the latest
16075 Finished messages sent to the peer or expected from the peer,
16076 respectively. (SSL_get_peer_finished is usually the Finished message
16077 actually received from the peer, otherwise the protocol will be aborted.)
16079 As the Finished message are message digests of the complete handshake
16080 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16081 be used for external authentication procedures when the authentication
16082 provided by SSL/TLS is not desired or is not enough.
16086 * Enhanced support for Alpha Linux is added. Now ./config checks if
16087 the host supports BWX extension and if Compaq C is present on the
16088 $PATH. Just exploiting of the BWX extension results in 20-30%
16089 performance kick for some algorithms, e.g. DES and RC4 to mention
16090 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16095 * Add support for MS "fast SGC". This is arguably a violation of the
16096 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16097 weak crypto and after checking the certificate is SGC a second one
16098 with strong crypto. MS SGC stops the first handshake after receiving
16099 the server certificate message and sends a second client hello. Since
16100 a server will typically do all the time consuming operations before
16101 expecting any further messages from the client (server key exchange
16102 is the most expensive) there is little difference between the two.
16104 To get OpenSSL to support MS SGC we have to permit a second client
16105 hello message after we have sent server done. In addition we have to
16106 reset the MAC if we do get this second client hello.
16110 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16111 if a DER encoded private key is RSA or DSA traditional format. Changed
16112 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16113 format DER encoded private key. Newer code should use PKCS#8 format which
16114 has the key type encoded in the ASN1 structure. Added DER private key
16115 support to pkcs8 application.
16119 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16120 ciphersuites has been selected (as required by the SSL 3/TLS 1
16121 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16122 is set, we interpret this as a request to violate the specification
16123 (the worst that can happen is a handshake failure, and 'correct'
16124 behaviour would result in a handshake failure anyway).
16128 * In SSL_CTX_add_session, take into account that there might be multiple
16129 SSL_SESSION structures with the same session ID (e.g. when two threads
16130 concurrently obtain them from an external cache).
16131 The internal cache can handle only one SSL_SESSION with a given ID,
16132 so if there's a conflict, we now throw out the old one to achieve
16137 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16138 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16139 some routines that use cipher OIDs: some ciphers do not have OIDs
16140 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16145 * Simplify the trust setting structure and code. Now we just have
16146 two sequences of OIDs for trusted and rejected settings. These will
16147 typically have values the same as the extended key usage extension
16148 and any application specific purposes.
16150 The trust checking code now has a default behaviour: it will just
16151 check for an object with the same NID as the passed id. Functions can
16152 be provided to override either the default behaviour or the behaviour
16153 for a given id. SSL client, server and email already have functions
16154 in place for compatibility: they check the NID and also return "trusted"
16155 if the certificate is self signed.
16159 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16160 traditional format into an EVP_PKEY structure.
16164 * Add a password callback function PEM_cb() which either prompts for
16165 a password if usr_data is NULL or otherwise assumes it is a null
16166 terminated password. Allow passwords to be passed on command line
16167 environment or config files in a few more utilities.
16171 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16172 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16173 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16174 Update documentation.
16178 * Support for ASN1 "NULL" type. This could be handled before by using
16179 ASN1_TYPE but there wasn't any function that would try to read a NULL
16180 and produce an error if it couldn't. For compatibility we also have
16181 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16182 don't allocate anything because they don't need to.
16186 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16189 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16191 * Rebuild of the memory allocation routines used by OpenSSL code and
16192 possibly others as well. The purpose is to make an interface that
16193 provide hooks so anyone can build a separate set of allocation and
16194 deallocation routines to be used by OpenSSL, for example memory
16195 pool implementations, or something else, which was previously hard
16196 since Malloc(), Realloc() and Free() were defined as macros having
16197 the values malloc, realloc and free, respectively (except for Win32
16198 compilations). The same is provided for memory debugging code.
16199 OpenSSL already comes with functionality to find memory leaks, but
16200 this gives people a chance to debug other memory problems.
16202 With these changes, a new set of functions and macros have appeared:
16204 CRYPTO_set_mem_debug_functions() [F]
16205 CRYPTO_get_mem_debug_functions() [F]
16206 CRYPTO_dbg_set_options() [F]
16207 CRYPTO_dbg_get_options() [F]
16208 CRYPTO_malloc_debug_init() [M]
16210 The memory debug functions are NULL by default, unless the library
16211 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16212 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16213 gives the standard debugging functions that come with OpenSSL) or
16214 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16215 provided by the library user) must be used. When the standard
16216 debugging functions are used, CRYPTO_dbg_set_options can be used to
16217 request additional information:
16218 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16219 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16221 Also, things like CRYPTO_set_mem_functions will always give the
16222 expected result (the new set of functions is used for allocation
16223 and deallocation) at all times, regardless of platform and compiler
16226 To finish it up, some functions that were never use in any other
16227 way than through macros have a new API and new semantic:
16229 CRYPTO_dbg_malloc()
16230 CRYPTO_dbg_realloc()
16233 All macros of value have retained their old syntax.
16235 *Richard Levitte and Bodo Moeller*
16237 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16238 ordering of SMIMECapabilities wasn't in "strength order" and there
16239 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16244 * Some ASN1 types with illegal zero length encoding (INTEGER,
16245 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16247 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16249 * Merge in my S/MIME library for OpenSSL. This provides a simple
16250 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16251 functionality to handle multipart/signed properly) and a utility
16252 called 'smime' to call all this stuff. This is based on code I
16253 originally wrote for Celo who have kindly allowed it to be
16254 included in OpenSSL.
16258 * Add variants des_set_key_checked and des_set_key_unchecked of
16259 des_set_key (aka des_key_sched). Global variable des_check_key
16260 decides which of these is called by des_set_key; this way
16261 des_check_key behaves as it always did, but applications and
16262 the library itself, which was buggy for des_check_key == 1,
16263 have a cleaner way to pick the version they need.
16267 * New function PKCS12_newpass() which changes the password of a
16272 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16273 dynamic mix. In both cases the ids can be used as an index into the
16274 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16275 functions so they accept a list of the field values and the
16276 application doesn't need to directly manipulate the X509_TRUST
16281 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16286 * Modify the way the V3 extension code looks up extensions. This now
16287 works in a similar way to the object code: we have some "standard"
16288 extensions in a static table which is searched with OBJ_bsearch()
16289 and the application can add dynamic ones if needed. The file
16290 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16291 updated whenever a new extension is added to the core code and kept
16292 in ext_nid order. There is a simple program 'tabtest.c' which checks
16293 this. New extensions are not added too often so this file can readily
16294 be maintained manually.
16296 There are two big advantages in doing things this way. The extensions
16297 can be looked up immediately and no longer need to be "added" using
16298 X509V3_add_standard_extensions(): this function now does nothing.
16299 Side note: I get *lots* of email saying the extension code doesn't
16300 work because people forget to call this function.
16301 Also no dynamic allocation is done unless new extensions are added:
16302 so if we don't add custom extensions there is no need to call
16303 X509V3_EXT_cleanup().
16307 * Modify enc utility's salting as follows: make salting the default. Add a
16308 magic header, so unsalted files fail gracefully instead of just decrypting
16309 to garbage. This is because not salting is a big security hole, so people
16310 should be discouraged from doing it.
16314 * Fixes and enhancements to the 'x509' utility. It allowed a message
16315 digest to be passed on the command line but it only used this
16316 parameter when signing a certificate. Modified so all relevant
16317 operations are affected by the digest parameter including the
16318 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16319 DSA key was used because it didn't fix the digest.
16323 * Initial certificate chain verify code. Currently tests the untrusted
16324 certificates for consistency with the verify purpose (which is set
16325 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16327 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16328 this is because it will reject chains with invalid extensions whereas
16329 every previous version of OpenSSL and SSLeay made no checks at all.
16331 Trust code: checks the root CA for the relevant trust settings. Trust
16332 settings have an initial value consistent with the verify purpose: e.g.
16333 if the verify purpose is for SSL client use it expects the CA to be
16334 trusted for SSL client use. However the default value can be changed to
16335 permit custom trust settings: one example of this would be to only trust
16336 certificates from a specific "secure" set of CAs.
16338 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16339 which should be used for version portability: especially since the
16340 verify structure is likely to change more often now.
16342 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16343 to set them. If not set then assume SSL clients will verify SSL servers
16346 Two new options to the verify program: -untrusted allows a set of
16347 untrusted certificates to be passed in and -purpose which sets the
16348 intended purpose of the certificate. If a purpose is set then the
16349 new chain verify code is used to check extension consistency.
16353 * Support for the authority information access extension.
16357 * Modify RSA and DSA PEM read routines to transparently handle
16358 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16359 public keys in a format compatible with certificate
16360 SubjectPublicKeyInfo structures. Unfortunately there were already
16361 functions called *_PublicKey_* which used various odd formats so
16362 these are retained for compatibility: however the DSA variants were
16363 never in a public release so they have been deleted. Changed dsa/rsa
16364 utilities to handle the new format: note no releases ever handled public
16365 keys so we should be OK.
16367 The primary motivation for this change is to avoid the same fiasco
16368 that dogs private keys: there are several incompatible private key
16369 formats some of which are standard and some OpenSSL specific and
16370 require various evil hacks to allow partial transparent handling and
16371 even then it doesn't work with DER formats. Given the option anything
16372 other than PKCS#8 should be dumped: but the other formats have to
16373 stay in the name of compatibility.
16375 With public keys and the benefit of hindsight one standard format
16376 is used which works with EVP_PKEY, RSA or DSA structures: though
16377 it clearly returns an error if you try to read the wrong kind of key.
16379 Added a -pubkey option to the 'x509' utility to output the public key.
16380 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16381 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16382 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16383 that do the same as the `EVP_PKEY_assign_*()` except they up the
16384 reference count of the added key (they don't "swallow" the
16389 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16390 CRLs would fail if the file contained no certificates or no CRLs:
16391 added a new function to read in both types and return the number
16392 read: this means that if none are read it will be an error. The
16393 DER versions of the certificate and CRL reader would always fail
16394 because it isn't possible to mix certificates and CRLs in DER format
16395 without choking one or the other routine. Changed this to just read
16396 a certificate: this is the best we can do. Also modified the code
16397 in `apps/verify.c` to take notice of return codes: it was previously
16398 attempting to read in certificates from NULL pointers and ignoring
16399 any errors: this is one reason why the cert and CRL reader seemed
16400 to work. It doesn't check return codes from the default certificate
16401 routines: these may well fail if the certificates aren't installed.
16405 * Code to support otherName option in GeneralName.
16409 * First update to verify code. Change the verify utility
16410 so it warns if it is passed a self signed certificate:
16411 for consistency with the normal behaviour. X509_verify
16412 has been modified to it will now verify a self signed
16413 certificate if *exactly* the same certificate appears
16414 in the store: it was previously impossible to trust a
16415 single self signed certificate. This means that:
16416 openssl verify ss.pem
16417 now gives a warning about a self signed certificate but
16418 openssl verify -CAfile ss.pem ss.pem
16423 * For servers, store verify_result in SSL_SESSION data structure
16424 (and add it to external session representation).
16425 This is needed when client certificate verifications fails,
16426 but an application-provided verification callback (set by
16427 SSL_CTX_set_cert_verify_callback) allows accepting the session
16428 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16429 but returns 1): When the session is reused, we have to set
16430 ssl->verify_result to the appropriate error code to avoid
16433 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16435 * Fix a bug in the new PKCS#7 code: it didn't consider the
16436 case in PKCS7_dataInit() where the signed PKCS7 structure
16437 didn't contain any existing data because it was being created.
16439 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16441 * Add a salt to the key derivation routines in enc.c. This
16442 forms the first 8 bytes of the encrypted file. Also add a
16443 -S option to allow a salt to be input on the command line.
16447 * New function X509_cmp(). Oddly enough there wasn't a function
16448 to compare two certificates. We do this by working out the SHA1
16449 hash and comparing that. X509_cmp() will be needed by the trust
16454 * SSL_get1_session() is like SSL_get_session(), but increments
16455 the reference count in the SSL_SESSION returned.
16457 *Geoff Thorpe <geoff@eu.c2.net>*
16459 * Fix for 'req': it was adding a null to request attributes.
16460 Also change the X509_LOOKUP and X509_INFO code to handle
16461 certificate auxiliary information.
16465 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16470 * Add the possibility to add extra information to the memory leak
16471 detecting output, to form tracebacks, showing from where each
16472 allocation was originated: CRYPTO_push_info("constant string") adds
16473 the string plus current file name and line number to a per-thread
16474 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16475 is like calling CYRPTO_pop_info() until the stack is empty.
16476 Also updated memory leak detection code to be multi-thread-safe.
16480 * Add options -text and -noout to pkcs7 utility and delete the
16481 encryption options which never did anything. Update docs.
16485 * Add options to some of the utilities to allow the pass phrase
16486 to be included on either the command line (not recommended on
16487 OSes like Unix) or read from the environment. Update the
16488 manpages and fix a few bugs.
16492 * Add a few manpages for some of the openssl commands.
16496 * Fix the -revoke option in ca. It was freeing up memory twice,
16497 leaking and not finding already revoked certificates.
16501 * Extensive changes to support certificate auxiliary information.
16502 This involves the use of X509_CERT_AUX structure and X509_AUX
16503 functions. An X509_AUX function such as PEM_read_X509_AUX()
16504 can still read in a certificate file in the usual way but it
16505 will also read in any additional "auxiliary information". By
16506 doing things this way a fair degree of compatibility can be
16507 retained: existing certificates can have this information added
16508 using the new 'x509' options.
16510 Current auxiliary information includes an "alias" and some trust
16511 settings. The trust settings will ultimately be used in enhanced
16512 certificate chain verification routines: currently a certificate
16513 can only be trusted if it is self signed and then it is trusted
16518 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16519 The problem was that one of the replacement routines had not been working
16520 since SSLeay releases. For now the offending routine has been replaced
16521 with non-optimised assembler. Even so, this now gives around 95%
16522 performance improvement for 1024 bit RSA signs.
16526 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16527 handling. Most clients have the effective key size in bits equal to
16528 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16529 A few however don't do this and instead use the size of the decrypted key
16530 to determine the RC2 key length and the AlgorithmIdentifier to determine
16531 the effective key length. In this case the effective key length can still
16532 be 40 bits but the key length can be 168 bits for example. This is fixed
16533 by manually forcing an RC2 key into the EVP_PKEY structure because the
16534 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16535 the key length and effective key length are equal.
16539 * Add a bunch of functions that should simplify the creation of
16540 X509_NAME structures. Now you should be able to do:
16541 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16542 and have it automatically work out the correct field type and fill in
16543 the structures. The more adventurous can try:
16544 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16545 and it will (hopefully) work out the correct multibyte encoding.
16549 * Change the 'req' utility to use the new field handling and multibyte
16550 copy routines. Before the DN field creation was handled in an ad hoc
16551 way in req, ca, and x509 which was rather broken and didn't support
16552 BMPStrings or UTF8Strings. Since some software doesn't implement
16553 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16554 using the dirstring_type option. See the new comment in the default
16555 openssl.cnf for more info.
16559 * Make crypto/rand/md_rand.c more robust:
16560 - Assure unique random numbers after fork().
16561 - Make sure that concurrent threads access the global counter and
16562 md serializably so that we never lose entropy in them
16563 or use exactly the same state in multiple threads.
16564 Access to the large state is not always serializable because
16565 the additional locking could be a performance killer, and
16566 md should be large enough anyway.
16570 * New file `apps/app_rand.c` with commonly needed functionality
16571 for handling the random seed file.
16573 Use the random seed file in some applications that previously did not:
16575 dsaparam -genkey (which also ignored its '-rand' option),
16578 x509 (when signing).
16579 Except on systems with /dev/urandom, it is crucial to have a random
16580 seed file at least for key creation, DSA signing, and for DH exchanges;
16581 for RSA signatures we could do without one.
16583 gendh and gendsa (unlike genrsa) used to read only the first byte
16584 of each file listed in the '-rand' option. The function as previously
16585 found in genrsa is now in app_rand.c and is used by all programs
16586 that support '-rand'.
16590 * In RAND_write_file, use mode 0600 for creating files;
16591 don't just chmod when it may be too late.
16595 * Report an error from X509_STORE_load_locations
16596 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16600 * New function ASN1_mbstring_copy() this copies a string in either
16601 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16602 into an ASN1_STRING type. A mask of permissible types is passed
16603 and it chooses the "minimal" type to use or an error if not type
16608 * Add function equivalents to the various macros in asn1.h. The old
16609 macros are retained with an `M_` prefix. Code inside the library can
16610 use the `M_` macros. External code (including the openssl utility)
16611 should *NOT* in order to be "shared library friendly".
16615 * Add various functions that can check a certificate's extensions
16616 to see if it usable for various purposes such as SSL client,
16617 server or S/MIME and CAs of these types. This is currently
16618 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16619 verification. Also added a -purpose flag to x509 utility to
16620 print out all the purposes.
16624 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16629 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16630 for, obtain and decode and extension and obtain its critical flag.
16631 This allows all the necessary extension code to be handled in a
16632 single function call.
16636 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16637 platforms. See crypto/rc4/rc4_enc.c for further details.
16641 * New -noout option to asn1parse. This causes no output to be produced
16642 its main use is when combined with -strparse and -out to extract data
16643 from a file (which may not be in ASN.1 format).
16647 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16648 when producing the local key id.
16650 *Richard Levitte <levitte@stacken.kth.se>*
16652 * New option -dhparam in s_server. This allows a DH parameter file to be
16653 stated explicitly. If it is not stated then it tries the first server
16654 certificate file. The previous behaviour hard coded the filename
16659 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16660 a public key to be input or output. For example:
16661 openssl rsa -in key.pem -pubout -out pubkey.pem
16662 Also added necessary DSA public key functions to handle this.
16666 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16667 in the message. This was handled by allowing
16668 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16670 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16672 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16673 to the end of the strings whereas this didn't. This would cause problems
16674 if strings read with d2i_ASN1_bytes() were later modified.
16676 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16678 * Fix for base64 decode bug. When a base64 bio reads only one line of
16679 data and it contains EOF it will end up returning an error. This is
16680 caused by input 46 bytes long. The cause is due to the way base64
16681 BIOs find the start of base64 encoded data. They do this by trying a
16682 trial decode on each line until they find one that works. When they
16683 do a flag is set and it starts again knowing it can pass all the
16684 data directly through the decoder. Unfortunately it doesn't reset
16685 the context it uses. This means that if EOF is reached an attempt
16686 is made to pass two EOFs through the context and this causes the
16687 resulting error. This can also cause other problems as well. As is
16688 usual with these problems it takes *ages* to find and the fix is
16689 trivial: move one line.
16691 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16693 * Ugly workaround to get s_client and s_server working under Windows. The
16694 old code wouldn't work because it needed to select() on sockets and the
16695 tty (for keypresses and to see if data could be written). Win32 only
16696 supports select() on sockets so we select() with a 1s timeout on the
16697 sockets and then see if any characters are waiting to be read, if none
16698 are present then we retry, we also assume we can always write data to
16699 the tty. This isn't nice because the code then blocks until we've
16700 received a complete line of data and it is effectively polling the
16701 keyboard at 1s intervals: however it's quite a bit better than not
16702 working at all :-) A dedicated Windows application might handle this
16703 with an event loop for example.
16707 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16708 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16709 will be called when RSA_sign() and RSA_verify() are used. This is useful
16710 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16711 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16712 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16713 This necessitated the support of an extra signature type NID_md5_sha1
16714 for SSL signatures and modifications to the SSL library to use it instead
16715 of calling RSA_public_decrypt() and RSA_private_encrypt().
16719 * Add new -verify -CAfile and -CApath options to the crl program, these
16720 will lookup a CRL issuers certificate and verify the signature in a
16721 similar way to the verify program. Tidy up the crl program so it
16722 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16723 less strict. It will now permit CRL extensions even if it is not
16724 a V2 CRL: this will allow it to tolerate some broken CRLs.
16728 * Initialize all non-automatic variables each time one of the openssl
16729 sub-programs is started (this is necessary as they may be started
16730 multiple times from the "OpenSSL>" prompt).
16732 *Lennart Bang, Bodo Moeller*
16734 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16735 removing all other RSA functionality (this is what NO_RSA does). This
16736 is so (for example) those in the US can disable those operations covered
16737 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16742 * Non-copying interface to BIO pairs.
16743 (still largely untested)
16747 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16748 ASCII string. This was handled independently in various places before.
16752 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16753 UTF8 strings a character at a time.
16757 * Use client_version from client hello to select the protocol
16758 (s23_srvr.c) and for RSA client key exchange verification
16759 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16763 * Add various utility functions to handle SPKACs, these were previously
16764 handled by poking round in the structure internals. Added new function
16765 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16766 print, verify and generate SPKACs. Based on an original idea from
16767 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16771 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16775 * Allow the config file extension section to be overwritten on the
16776 command line. Based on an original idea from Massimiliano Pala
16777 <madwolf@comune.modena.it>. The new option is called -extensions
16778 and can be applied to ca, req and x509. Also -reqexts to override
16779 the request extensions in req and -crlexts to override the crl extensions
16784 * Add new feature to the SPKAC handling in ca. Now you can include
16785 the same field multiple times by preceding it by "XXXX." for example:
16788 this is the same syntax as used in the req config file.
16792 * Allow certificate extensions to be added to certificate requests. These
16793 are specified in a 'req_extensions' option of the req section of the
16794 config file. They can be printed out with the -text option to req but
16795 are otherwise ignored at present.
16799 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16800 data read consists of only the final block it would not decrypted because
16801 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16802 A misplaced 'break' also meant the decrypted final block might not be
16803 copied until the next read.
16807 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16808 a few extra parameters to the DH structure: these will be useful if
16809 for example we want the value of 'q' or implement X9.42 DH.
16813 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16814 provides hooks that allow the default DSA functions or functions on a
16815 "per key" basis to be replaced. This allows hardware acceleration and
16816 hardware key storage to be handled without major modification to the
16817 library. Also added low level modexp hooks and CRYPTO_EX structure and
16818 associated functions.
16822 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16823 as "read only": it can't be written to and the buffer it points to will
16824 not be freed. Reading from a read only BIO is much more efficient than
16825 a normal memory BIO. This was added because there are several times when
16826 an area of memory needs to be read from a BIO. The previous method was
16827 to create a memory BIO and write the data to it, this results in two
16828 copies of the data and an O(n^2) reading algorithm. There is a new
16829 function BIO_new_mem_buf() which creates a read only memory BIO from
16830 an area of memory. Also modified the PKCS#7 routines to use read only
16835 * Bugfix: ssl23_get_client_hello did not work properly when called in
16836 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16837 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16838 but a retry condition occurred while trying to read the rest.
16842 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16843 NID_pkcs7_encrypted by default: this was wrong since this should almost
16844 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16845 the encrypted data type: this is a more sensible place to put it and it
16846 allows the PKCS#12 code to be tidied up that duplicated this
16851 * Changed obj_dat.pl script so it takes its input and output files on
16852 the command line. This should avoid shell escape redirection problems
16857 * Initial support for certificate extension requests, these are included
16858 in things like Xenroll certificate requests. Included functions to allow
16859 extensions to be obtained and added.
16863 * -crlf option to s_client and s_server for sending newlines as
16864 CRLF (as required by many protocols).
16868 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16870 * Install libRSAglue.a when OpenSSL is built with RSAref.
16872 *Ralf S. Engelschall*
16874 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16876 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16878 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16883 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16884 DH parameters/keys (q is lost during that conversion, but the resulting
16885 DH parameters contain its length).
16887 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16888 much faster than DH_generate_parameters (which creates parameters
16889 where `p = 2*q + 1`), and also the smaller q makes DH computations
16890 much more efficient (160-bit exponentiation instead of 1024-bit
16891 exponentiation); so this provides a convenient way to support DHE
16892 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16893 utter importance to use
16894 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16896 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16897 when such DH parameters are used, because otherwise small subgroup
16898 attacks may become possible!
16902 * Avoid memory leak in i2d_DHparams.
16906 * Allow the -k option to be used more than once in the enc program:
16907 this allows the same encrypted message to be read by multiple recipients.
16911 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16912 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16913 it will always use the numerical form of the OID, even if it has a short
16918 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16919 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16920 otherwise bn_mod_exp was called. In the case of hardware keys for example
16921 no private key components need be present and it might store extra data
16922 in the RSA structure, which cannot be accessed from bn_mod_exp.
16923 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16924 private key operations.
16928 * Added support for SPARC Linux.
16932 * pem_password_cb function type incompatibly changed from
16933 typedef int pem_password_cb(char *buf, int size, int rwflag);
16935 ....(char *buf, int size, int rwflag, void *userdata);
16936 so that applications can pass data to their callbacks:
16937 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16938 additional void * argument, which is just handed through whenever
16939 the password callback is called.
16941 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16943 New function SSL_CTX_set_default_passwd_cb_userdata.
16945 Compatibility note: As many C implementations push function arguments
16946 onto the stack in reverse order, the new library version is likely to
16947 interoperate with programs that have been compiled with the old
16948 pem_password_cb definition (PEM_whatever takes some data that
16949 happens to be on the stack as its last argument, and the callback
16950 just ignores this garbage); but there is no guarantee whatsoever that
16953 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16954 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16955 problems not only on Windows, but also on some Unix platforms.
16956 To avoid problematic command lines, these definitions are now in an
16957 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16958 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16962 * MIPS III/IV assembler module is reimplemented.
16966 * More DES library cleanups: remove references to srand/rand and
16967 delete an unused file.
16971 * Add support for the free Netwide assembler (NASM) under Win32,
16972 since not many people have MASM (ml) and it can be hard to obtain.
16973 This is currently experimental but it seems to work OK and pass all
16974 the tests. Check out INSTALL.W32 for info.
16978 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16979 without temporary keys kept an extra copy of the server key,
16980 and connections with temporary keys did not free everything in case
16985 * New function RSA_check_key and new openssl rsa option -check
16986 for verifying the consistency of RSA keys.
16988 *Ulf Moeller, Bodo Moeller*
16990 * Various changes to make Win32 compile work:
16991 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16992 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16993 comparison" warnings.
16994 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16998 * Add a debugging option to PKCS#5 v2 key generation function: when
16999 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17000 derived keys are printed to stderr.
17004 * Copy the flags in ASN1_STRING_dup().
17006 *Roman E. Pavlov <pre@mo.msk.ru>*
17008 * The x509 application mishandled signing requests containing DSA
17009 keys when the signing key was also DSA and the parameters didn't match.
17011 It was supposed to omit the parameters when they matched the signing key:
17012 the verifying software was then supposed to automatically use the CA's
17013 parameters if they were absent from the end user certificate.
17015 Omitting parameters is no longer recommended. The test was also
17016 the wrong way round! This was probably due to unusual behaviour in
17017 EVP_cmp_parameters() which returns 1 if the parameters match.
17018 This meant that parameters were omitted when they *didn't* match and
17019 the certificate was useless. Certificates signed with 'ca' didn't have
17022 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17024 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17025 The interface is as follows:
17026 Applications can use
17027 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17028 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17029 "off" is now the default.
17030 The library internally uses
17031 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17032 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17033 to disable memory-checking temporarily.
17035 Some inconsistent states that previously were possible (and were
17036 even the default) are now avoided.
17038 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17039 with each memory chunk allocated; this is occasionally more helpful
17040 than just having a counter.
17042 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17044 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17049 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17050 which largely parallels "options", but is for changing API behaviour,
17051 whereas "options" are about protocol behaviour.
17052 Initial "mode" flags are:
17054 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17055 a single record has been written.
17056 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17057 retries use the same buffer location.
17058 (But all of the contents must be
17063 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17066 * Fix problems with no-hmac etc.
17068 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17070 * New functions RSA_get_default_method(), RSA_set_method() and
17071 RSA_get_method(). These allows replacement of RSA_METHODs without having
17072 to mess around with the internals of an RSA structure.
17076 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17077 Also really enable memory leak checks in openssl.c and in some
17080 *Chad C. Mulligan, Bodo Moeller*
17082 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17083 up the length of negative integers. This has now been simplified to just
17084 store the length when it is first determined and use it later, rather
17085 than trying to keep track of where data is copied and updating it to
17087 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17089 * Add a new function PKCS7_signatureVerify. This allows the verification
17090 of a PKCS#7 signature but with the signing certificate passed to the
17091 function itself. This contrasts with PKCS7_dataVerify which assumes the
17092 certificate is present in the PKCS#7 structure. This isn't always the
17093 case: certificates can be omitted from a PKCS#7 structure and be
17094 distributed by "out of band" means (such as a certificate database).
17098 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17099 function prototypes in pem.h, also change util/mkdef.pl to add the
17100 necessary function names.
17104 * mk1mf.pl (used by Windows builds) did not properly read the
17105 options set by Configure in the top level Makefile, and Configure
17106 was not even able to write more than one option correctly.
17107 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17111 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17112 file to be loaded from a BIO or FILE pointer. The BIO version will
17113 for example allow memory BIOs to contain config info.
17117 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17118 Whoever hopes to achieve shared-library compatibility across versions
17119 must use this, not the compile-time macro.
17120 (Exercise 0.9.4: Which is the minimum library version required by
17122 Note: All this applies only to multi-threaded programs, others don't
17127 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17128 through a BIO pair triggered the default case, i.e.
17129 SSLerr(...,SSL_R_UNKNOWN_STATE).
17133 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17134 can use the SSL library even if none of the specific BIOs is
17139 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17140 for the encoded length.
17142 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17144 * Add initial documentation of the X509V3 functions.
17148 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17149 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17150 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17151 secure PKCS#8 private key format with a high iteration count.
17155 * Fix determination of Perl interpreter: A perl or perl5
17156 *directory* in $PATH was also accepted as the interpreter.
17158 *Ralf S. Engelschall*
17160 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17161 wrong with it but it was very old and did things like calling
17162 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17163 unusual formatting.
17167 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17168 to use the new extension code.
17172 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17173 with macros. This should make it easier to change their form, add extra
17174 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17179 * Add to configuration table a new entry that can specify an alternative
17180 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17181 according to Mark Crispin <MRC@Panda.COM>.
17185 * DES CBC did not update the IV. Weird.
17189 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17190 Changing the behaviour of the former might break existing programs --
17191 where IV updating is needed, des_ncbc_encrypt can be used.
17194 * When bntest is run from "make test" it drives bc to check its
17195 calculations, as well as internally checking them. If an internal check
17196 fails, it needs to cause bc to give a non-zero result or make test carries
17197 on without noticing the failure. Fixed.
17201 * DES library cleanups.
17205 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17206 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17207 ciphers. NOTE: although the key derivation function has been verified
17208 against some published test vectors it has not been extensively tested
17209 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17214 * Instead of "mkdir -p", which is not fully portable, use new
17215 Perl script "util/mkdir-p.pl".
17219 * Rewrite the way password based encryption (PBE) is handled. It used to
17220 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17221 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17222 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17223 the 'parameter' field of the AlgorithmIdentifier is passed to the
17224 underlying key generation function so it must do its own ASN1 parsing.
17225 This has also changed the EVP_PBE_CipherInit() function which now has a
17226 'parameter' argument instead of literal salt and iteration count values
17227 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17231 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17232 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17233 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17234 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17235 value was just used as a "magic string" and not used directly its
17236 value doesn't matter.
17240 * Introduce some semblance of const correctness to BN. Shame C doesn't
17245 * "linux-sparc64" configuration (ultrapenguin).
17247 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17248 "linux-sparc" configuration.
17250 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17252 * config now generates no-xxx options for missing ciphers.
17256 * Support the EBCDIC character set (work in progress).
17257 File ebcdic.c not yet included because it has a different license.
17259 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17261 * Support BS2000/OSD-POSIX.
17263 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17265 * Make callbacks for key generation use `void *` instead of `char *`.
17269 * Make S/MIME samples compile (not yet tested).
17273 * Additional typesafe stacks.
17277 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17281 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17283 * New configuration variant "sco5-gcc".
17285 * Updated some demos.
17287 *Sean O Riordain, Wade Scholine*
17289 * Add missing BIO_free at exit of pkcs12 application.
17293 * Fix memory leak in conf.c.
17297 * Updates for Win32 to assembler version of MD5.
17301 * Set #! path to perl in `apps/der_chop` to where we found it
17302 instead of using a fixed path.
17306 * SHA library changes for irix64-mips4-cc.
17310 * Improvements for VMS support.
17314 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17316 * Bignum library bug fix. IRIX 6 passes "make test" now!
17317 This also avoids the problems with SC4.2 and unpatched SC5.
17319 *Andy Polyakov <appro@fy.chalmers.se>*
17321 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17322 These are required because of the typesafe stack would otherwise break
17323 existing code. If old code used a structure member which used to be STACK
17324 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17325 sk_num or sk_value it would produce an error because the num, data members
17326 are not present in STACK_OF. Now it just produces a warning. sk_set
17327 replaces the old method of assigning a value to sk_value
17328 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17329 that does this will no longer work (and should use sk_set instead) but
17330 this could be regarded as a "questionable" behaviour anyway.
17334 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17335 correctly handle encrypted S/MIME data.
17339 * Change type of various DES function arguments from des_cblock
17340 (which means, in function argument declarations, pointer to char)
17341 to des_cblock * (meaning pointer to array with 8 char elements),
17342 which allows the compiler to do more typechecking; it was like
17343 that back in SSLeay, but with lots of ugly casts.
17345 Introduce new type const_des_cblock.
17349 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17350 problems: find RecipientInfo structure that matches recipient certificate
17351 and initialise the ASN1 structures properly based on passed cipher.
17355 * Belatedly make the BN tests actually check the results.
17359 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17360 to and from BNs: it was completely broken. New compilation option
17361 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17362 key elements as negative integers.
17366 * Reorganize and speed up MD5.
17368 *Andy Polyakov <appro@fy.chalmers.se>*
17372 *Richard Levitte <richard@levitte.org>*
17374 * New option -out to asn1parse to allow the parsed structure to be
17375 output to a file. This is most useful when combined with the -strparse
17376 option to examine the output of things like OCTET STRINGS.
17380 * Make SSL library a little more fool-proof by not requiring any longer
17381 that `SSL_set_{accept,connect}_state` be called before
17382 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17383 in many applications because usually everything *appeared* to work as
17384 intended anyway -- now it really works as intended).
17388 * Move openssl.cnf out of lib/.
17392 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17393 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17394 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17396 *Ralf S. Engelschall*
17398 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17399 handle PKCS#7 enveloped data properly.
17401 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17403 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17404 copying pointers. The cert_st handling is changed by this in
17405 various ways (and thus what used to be known as ctx->default_cert
17406 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17407 any longer when s->cert does not give us what we need).
17408 ssl_cert_instantiate becomes obsolete by this change.
17409 As soon as we've got the new code right (possibly it already is?),
17410 we have solved a couple of bugs of the earlier code where s->cert
17411 was used as if it could not have been shared with other SSL structures.
17413 Note that using the SSL API in certain dirty ways now will result
17414 in different behaviour than observed with earlier library versions:
17415 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17416 does not influence s as it used to.
17418 In order to clean up things more thoroughly, inside SSL_SESSION
17419 we don't use CERT any longer, but a new structure SESS_CERT
17420 that holds per-session data (if available); currently, this is
17421 the peer's certificate chain and, for clients, the server's certificate
17422 and temporary key. CERT holds only those values that can have
17423 meaningful defaults in an SSL_CTX.
17427 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17428 from the internal representation. Various PKCS#7 fixes: remove some
17429 evil casts and set the enc_dig_alg field properly based on the signing
17434 * Allow PKCS#12 password to be set from the command line or the
17435 environment. Let 'ca' get its config file name from the environment
17436 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17441 * Allow certificate policies extension to use an IA5STRING for the
17442 organization field. This is contrary to the PKIX definition but
17443 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17448 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17449 without disallowing inline assembler and the like for non-pedantic builds.
17453 * Support Borland C++ builder.
17455 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17461 * SHA-1 cleanups and performance enhancements.
17463 *Andy Polyakov <appro@fy.chalmers.se>*
17465 * Sparc v8plus assembler for the bignum library.
17467 *Andy Polyakov <appro@fy.chalmers.se>*
17469 * Accept any -xxx and +xxx compiler options in Configure.
17473 * Update HPUX configuration.
17477 * Add missing `sk_<type>_unshift()` function to safestack.h
17479 *Ralf S. Engelschall*
17481 * New function SSL_CTX_use_certificate_chain_file that sets the
17482 "extra_cert"s in addition to the certificate. (This makes sense
17483 only for "PEM" format files, as chains as a whole are not
17488 * Support verify_depth from the SSL API.
17489 x509_vfy.c had what can be considered an off-by-one-error:
17490 Its depth (which was not part of the external interface)
17491 was actually counting the number of certificates in a chain;
17492 now it really counts the depth.
17496 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17497 instead of X509err, which often resulted in confusing error
17498 messages since the error codes are not globally unique
17499 (e.g. an alleged error in ssl3_accept when a certificate
17500 didn't match the private key).
17502 * New function SSL_CTX_set_session_id_context that allows to set a default
17503 value (so that you don't need SSL_set_session_id_context for each
17504 connection using the SSL_CTX).
17508 * OAEP decoding bug fix.
17512 * Support INSTALL_PREFIX for package builders, as proposed by
17517 * New Configure options "threads" and "no-threads". For systems
17518 where the proper compiler options are known (currently Solaris
17519 and Linux), "threads" is the default.
17523 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17527 * Install various scripts to $(OPENSSLDIR)/misc, not to
17528 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17529 such as /usr/local/bin.
17533 * "make linux-shared" to build shared libraries.
17535 *Niels Poppe <niels@netbox.org>*
17537 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17541 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17542 extension adding in x509 utility.
17546 * Remove NOPROTO sections and error code comments.
17550 * Partial rewrite of the DEF file generator to now parse the ANSI
17555 * New Configure options --prefix=DIR and --openssldir=DIR.
17559 * Complete rewrite of the error code script(s). It is all now handled
17560 by one script at the top level which handles error code gathering,
17561 header rewriting and C source file generation. It should be much better
17562 than the old method: it now uses a modified version of Ulf's parser to
17563 read the ANSI prototypes in all header files (thus the old K&R definitions
17564 aren't needed for error creation any more) and do a better job of
17565 translating function codes into names. The old 'ASN1 error code embedded
17566 in a comment' is no longer necessary and it doesn't use .err files which
17567 have now been deleted. Also the error code call doesn't have to appear all
17568 on one line (which resulted in some large lines...).
17572 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17576 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17577 0 (which usually indicates a closed connection), but continue reading.
17581 * Fix some race conditions.
17585 * Add support for CRL distribution points extension. Add Certificate
17586 Policies and CRL distribution points documentation.
17590 * Move the autogenerated header file parts to crypto/opensslconf.h.
17594 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17595 8 of keying material. Merlin has also confirmed interop with this fix
17596 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17598 *Merlin Hughes <merlin@baltimore.ie>*
17600 * Fix lots of warnings.
17602 *Richard Levitte <levitte@stacken.kth.se>*
17604 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17605 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17607 *Richard Levitte <levitte@stacken.kth.se>*
17609 * Fix problems with sizeof(long) == 8.
17611 *Andy Polyakov <appro@fy.chalmers.se>*
17613 * Change functions to ANSI C.
17617 * Fix typos in error codes.
17619 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17621 * Remove defunct assembler files from Configure.
17625 * SPARC v8 assembler BIGNUM implementation.
17627 *Andy Polyakov <appro@fy.chalmers.se>*
17629 * Support for Certificate Policies extension: both print and set.
17630 Various additions to support the r2i method this uses.
17634 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17635 return a const string when you are expecting an allocated buffer.
17639 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17640 types DirectoryString and DisplayText.
17644 * Add code to allow r2i extensions to access the configuration database,
17645 add an LHASH database driver and add several ctx helper functions.
17649 * Fix an evil bug in bn_expand2() which caused various BN functions to
17650 fail when they extended the size of a BIGNUM.
17654 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17655 support typesafe stack.
17659 * Fix typo in SSL_[gs]et_options().
17661 *Nils Frostberg <nils@medcom.se>*
17663 * Delete various functions and files that belonged to the (now obsolete)
17664 old X509V3 handling code.
17668 * New Configure option "rsaref".
17672 * Don't auto-generate pem.h.
17676 * Introduce type-safe ASN.1 SETs.
17680 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17682 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17684 * Introduce type-safe STACKs. This will almost certainly break lots of code
17685 that links with OpenSSL (well at least cause lots of warnings), but fear
17686 not: the conversion is trivial, and it eliminates loads of evil casts. A
17687 few STACKed things have been converted already. Feel free to convert more.
17688 In the fullness of time, I'll do away with the STACK type altogether.
17692 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17693 specified in `<certfile>` by updating the entry in the index.txt file.
17694 This way one no longer has to edit the index.txt file manually for
17695 revoking a certificate. The -revoke option does the gory details now.
17697 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17699 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17700 `-text` option at all and this way the `-noout -text` combination was
17701 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17703 *Ralf S. Engelschall*
17705 * Make sure a corresponding plain text error message exists for the
17706 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17707 verify callback function determined that a certificate was revoked.
17709 *Ralf S. Engelschall*
17711 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17712 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17713 all available ciphers including rc5, which was forgotten until now.
17714 In order to let the testing shell script know which algorithms
17715 are available, a new (up to now undocumented) command
17716 `openssl list-cipher-commands` is used.
17720 * Bugfix: s_client occasionally would sleep in select() when
17721 it should have checked SSL_pending() first.
17725 * New functions DSA_do_sign and DSA_do_verify to provide access to
17726 the raw DSA values prior to ASN.1 encoding.
17730 * Tweaks to Configure
17732 *Niels Poppe <niels@netbox.org>*
17734 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17739 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17743 * New config option to avoid instructions that are illegal on the 80386.
17744 The default code is faster, but requires at least a 486.
17748 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17749 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17750 same as SSL2_VERSION anyway.
17754 * New "-showcerts" option for s_client.
17758 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17759 application. Various cleanups and fixes.
17763 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17764 modify error routines to work internally. Add error codes and PBE init
17765 to library startup routines.
17769 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17770 packing functions to asn1 and evp. Changed function names and error
17771 codes along the way.
17775 * PKCS12 integration: and so it begins... First of several patches to
17776 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17777 objects to objects.h
17781 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17782 and display support for Thawte strong extranet extension.
17786 * Add LinuxPPC support.
17788 *Jeff Dubrule <igor@pobox.org>*
17790 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17791 bn_div_words in alpha.s.
17793 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17795 * Make sure the RSA OAEP test is skipped under -DRSAref because
17796 OAEP isn't supported when OpenSSL is built with RSAref.
17798 *Ulf Moeller <ulf@fitug.de>*
17800 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17801 so they no longer are missing under -DNOPROTO.
17803 *Soren S. Jorvang <soren@t.dk>*
17805 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17807 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17808 doesn't work when the session is reused. Coming soon!
17812 * Fix a security hole, that allows sessions to be reused in the wrong
17813 context thus bypassing client cert protection! All software that uses
17814 client certs and session caches in multiple contexts NEEDS PATCHING to
17815 allow session reuse! A fuller solution is in the works.
17817 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17819 * Some more source tree cleanups (removed obsolete files
17820 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17821 permission on "config" script to be executable) and a fix for the INSTALL
17824 *Ulf Moeller <ulf@fitug.de>*
17826 * Remove some legacy and erroneous uses of malloc, free instead of
17829 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17831 * Make rsa_oaep_test return non-zero on error.
17833 *Ulf Moeller <ulf@fitug.de>*
17835 * Add support for native Solaris shared libraries. Configure
17836 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17837 if someone would make that last step automatic.
17839 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17841 * ctx_size was not built with the right compiler during "make links". Fixed.
17845 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17846 except NULL ciphers". This means the default cipher list will no longer
17847 enable NULL ciphers. They need to be specifically enabled e.g. with
17848 the string "DEFAULT:eNULL".
17852 * Fix to RSA private encryption routines: if p < q then it would
17853 occasionally produce an invalid result. This will only happen with
17854 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17858 * Be less restrictive and allow also `perl util/perlpath.pl
17859 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17860 because this way one can also use an interpreter named `perl5` (which is
17861 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17862 installed as `perl`).
17864 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17866 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17868 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17870 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17871 advapi32.lib to Win32 build and change the pem test comparison
17872 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17873 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17874 and crypto/des/ede_cbcm_enc.c.
17878 * DES quad checksum was broken on big-endian architectures. Fixed.
17882 * Comment out two functions in bio.h that aren't implemented. Fix up the
17883 Win32 test batch file so it (might) work again. The Win32 test batch file
17884 is horrible: I feel ill....
17888 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17889 in e_os.h. Audit of header files to check ANSI and non ANSI
17890 sections: 10 functions were absent from non ANSI section and not exported
17891 from Windows DLLs. Fixed up libeay.num for new functions.
17895 * Make `openssl version` output lines consistent.
17897 *Ralf S. Engelschall*
17899 * Fix Win32 symbol export lists for BIO functions: Added
17900 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17901 to ms/libeay{16,32}.def.
17903 *Ralf S. Engelschall*
17905 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17906 fine under Unix and passes some trivial tests I've now added. But the
17907 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17908 added to make sure no one expects that this stuff really works in the
17909 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17910 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17913 *Ralf S. Engelschall*
17915 * Fix the generation of two part addresses in perl.
17917 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17919 * Add config entry for Linux on MIPS.
17921 *John Tobey <jtobey@channel1.com>*
17923 * Make links whenever Configure is run, unless we are on Windoze.
17927 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17928 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17933 * Add a useful kludge to allow package maintainers to specify compiler and
17934 other platforms details on the command line without having to patch the
17935 Configure script every time: One now can use
17936 `perl Configure <id>:<details>`,
17937 i.e. platform ids are allowed to have details appended
17938 to them (separated by colons). This is treated as there would be a static
17939 pre-configured entry in Configure's %table under key `<id>` with value
17940 `<details>` and `perl Configure <id>` is called. So, when you want to
17941 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17942 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17943 now, which overrides the FreeBSD-elf entry on-the-fly.
17945 *Ralf S. Engelschall*
17947 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17951 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17952 on the `perl Configure ...` command line. This way one can compile
17953 OpenSSL libraries with Position Independent Code (PIC) which is needed
17954 for linking it into DSOs.
17956 *Ralf S. Engelschall*
17958 * Remarkably, export ciphers were totally broken and no-one had noticed!
17963 * Cleaned up the LICENSE document: The official contact for any license
17964 questions now is the OpenSSL core team under openssl-core@openssl.org.
17965 And add a paragraph about the dual-license situation to make sure people
17966 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17967 to the OpenSSL toolkit.
17969 *Ralf S. Engelschall*
17971 * General source tree makefile cleanups: Made `making xxx in yyy...`
17972 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17973 Additionally cleaned up the `make links` target: Remove unnecessary
17974 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17975 to speed processing and no longer clutter the display with confusing
17976 stuff. Instead only the actually done links are displayed.
17978 *Ralf S. Engelschall*
17980 * Permit null encryption ciphersuites, used for authentication only. It used
17981 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17982 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17987 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17988 signed attributes when verifying signatures (this would break them),
17989 the detached data encoding was wrong and public keys obtained using
17990 X509_get_pubkey() weren't freed.
17994 * Add text documentation for the BUFFER functions. Also added a work around
17995 to a Win95 console bug. This was triggered by the password read stuff: the
17996 last character typed gets carried over to the next fread(). If you were
17997 generating a new cert request using 'req' for example then the last
17998 character of the passphrase would be CR which would then enter the first
18003 * Added the new 'Includes OpenSSL Cryptography Software' button as
18004 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18005 button and can be used by applications based on OpenSSL to show the
18006 relationship to the OpenSSL project.
18008 *Ralf S. Engelschall*
18010 * Remove confusing variables in function signatures in files
18011 ssl/ssl_lib.c and ssl/ssl.h.
18013 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18015 * Don't install bss_file.c under PREFIX/include/
18017 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18019 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18020 functions that return function pointers and has support for NT specific
18021 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18022 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18023 unsigned to signed types: this was killing the Win32 compile.
18027 * Add new certificate file to stack functions,
18028 SSL_add_dir_cert_subjects_to_stack() and
18029 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18030 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18031 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18032 This means that Apache-SSL and similar packages don't have to mess around
18033 to add as many CAs as they want to the preferred list.
18037 * Experiment with doxygen documentation. Currently only partially applied to
18039 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18040 openssl.doxy as the configuration file.
18044 * Get rid of remaining C++-style comments which strict C compilers hate.
18046 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18048 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18049 compiled in by default: it has problems with large keys.
18053 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18054 DH private keys and/or callback functions which directly correspond to
18055 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18056 is needed for applications which have to configure certificates on a
18057 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18059 For the RSA certificate situation is makes no difference, but
18060 for the DSA certificate situation this fixes the "no shared cipher"
18061 problem where the OpenSSL cipher selection procedure failed because the
18062 temporary keys were not overtaken from the context and the API provided
18063 no way to reconfigure them.
18064 The new functions now let applications reconfigure the stuff and they
18065 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18066 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18067 non-public-API function ssl_cert_instantiate() is used as a helper
18068 function and also to reduce code redundancy inside ssl_rsa.c.
18070 *Ralf S. Engelschall*
18072 * Move s_server -dcert and -dkey options out of the undocumented feature
18073 area because they are useful for the DSA situation and should be
18074 recognized by the users.
18076 *Ralf S. Engelschall*
18078 * Fix the cipher decision scheme for export ciphers: the export bits are
18079 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18080 SSL_EXP_MASK. So, the original variable has to be used instead of the
18081 already masked variable.
18083 *Richard Levitte <levitte@stacken.kth.se>*
18085 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18087 *Richard Levitte <levitte@stacken.kth.se>*
18089 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18090 from `int` to `unsigned int` because it is a length and initialized by
18091 EVP_DigestFinal() which expects an `unsigned int *`.
18093 *Richard Levitte <levitte@stacken.kth.se>*
18095 * Don't hard-code path to Perl interpreter on shebang line of Configure
18096 script. Instead use the usual Shell->Perl transition trick.
18098 *Ralf S. Engelschall*
18100 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18101 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18102 -noout -modulus` as it's already the case for `openssl rsa -noout
18103 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18104 currently the public key is printed (a decision which was already done by
18105 `openssl dsa -modulus` in the past) which serves a similar purpose.
18106 Additionally the NO_RSA no longer completely removes the whole -modulus
18107 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18110 *Ralf S. Engelschall*
18112 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18113 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18115 *Arne Ansper <arne@ats.cyber.ee>*
18117 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18118 to be added. Now both 'req' and 'ca' can use new objects defined in the
18123 * Add cool BIO that does syslog (or event log on NT).
18125 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18127 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18128 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18129 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18130 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18134 * Add preliminary config info for new extension code.
18138 * Make RSA_NO_PADDING really use no padding.
18140 *Ulf Moeller <ulf@fitug.de>*
18142 * Generate errors when private/public key check is done.
18146 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18147 for some CRL extensions and new objects added.
18151 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18152 key usage extension and fuller support for authority key id.
18156 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18157 padding method for RSA, which is recommended for new applications in PKCS
18158 #1 v2.0 (RFC 2437, October 1998).
18159 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18160 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18161 against Bleichbacher's attack on RSA.
18162 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18165 * Updates to the new SSL compression code
18167 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18169 * Fix so that the version number in the master secret, when passed
18170 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18171 (because the server will not accept higher), that the version number
18172 is 0x03,0x01, not 0x03,0x00
18174 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18176 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18177 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18178 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18182 * Support for RAW extensions where an arbitrary extension can be
18183 created by including its DER encoding. See `apps/openssl.cnf` for
18188 * Make sure latest Perl versions don't interpret some generated C array
18189 code as Perl array code in the crypto/err/err_genc.pl script.
18191 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18193 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18194 not many people have the assembler. Various Win32 compilation fixes and
18195 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18196 build instructions.
18200 * Modify configure script 'Configure' to automatically create crypto/date.h
18201 file under Win32 and also build pem.h from pem.org. New script
18202 util/mkfiles.pl to create the MINFO file on environments that can't do a
18203 'make files': perl util/mkfiles.pl >MINFO should work.
18207 * Major rework of DES function declarations, in the pursuit of correctness
18208 and purity. As a result, many evil casts evaporated, and some weirdness,
18209 too. You may find this causes warnings in your code. Zapping your evil
18210 casts will probably fix them. Mostly.
18214 * Fix for a typo in asn1.h. Bug fix to object creation script
18215 obj_dat.pl. It considered a zero in an object definition to mean
18216 "end of object": none of the objects in objects.h have any zeros
18217 so it wasn't spotted.
18219 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18221 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18222 Masking (CBCM). In the absence of test vectors, the best I have been able
18223 to do is check that the decrypt undoes the encrypt, so far. Send me test
18224 vectors if you have them.
18228 * Correct calculation of key length for export ciphers (too much space was
18229 allocated for null ciphers). This has not been tested!
18233 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18234 message is now correct (it understands "crypto" and "ssl" on its
18235 command line). There is also now an "update" option. This will update
18236 the util/ssleay.num and util/libeay.num files with any new functions.
18238 perl util/mkdef.pl crypto ssl update
18239 it will update them.
18243 * Overhauled the Perl interface:
18244 - ported BN stuff to OpenSSL's different BN library
18245 - made the perl/ source tree CVS-aware
18246 - renamed the package from SSLeay to OpenSSL (the files still contain
18247 their history because I've copied them in the repository)
18248 - removed obsolete files (the test scripts will be replaced
18249 by better Test::Harness variants in the future)
18251 *Ralf S. Engelschall*
18253 * First cut for a very conservative source tree cleanup:
18254 1. merge various obsolete readme texts into doc/ssleay.txt
18255 where we collect the old documents and readme texts.
18256 2. remove the first part of files where I'm already sure that we no
18257 longer need them because of three reasons: either they are just temporary
18258 files which were left by Eric or they are preserved original files where
18259 I've verified that the diff is also available in the CVS via "cvs diff
18260 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18261 the crypto/md/ stuff).
18263 *Ralf S. Engelschall*
18265 * More extension code. Incomplete support for subject and issuer alt
18266 name, issuer and authority key id. Change the i2v function parameters
18267 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18268 what that's for :-) Fix to ASN1 macro which messed up
18269 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18273 * Preliminary support for ENUMERATED type. This is largely copied from the
18278 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18280 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18282 * Make sure `make rehash` target really finds the `openssl` program.
18284 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18286 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18287 like to hear about it if this slows down other processors.
18291 * Add CygWin32 platform information to Configure script.
18293 *Alan Batie <batie@aahz.jf.intel.com>*
18295 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18297 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18299 * New program nseq to manipulate netscape certificate sequences
18303 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18308 * Fixes to BN code. Previously the default was to define BN_RECURSION
18309 but the BN code had some problems that would cause failures when
18310 doing certificate verification and some other functions.
18312 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18314 * Add ASN1 and PEM code to support netscape certificate sequences.
18318 * Add ASN1 and PEM code to support netscape certificate sequences.
18322 * Add several PKIX and private extended key usage OIDs.
18326 * Modify the 'ca' program to handle the new extension code. Modify
18327 openssl.cnf for new extension format, add comments.
18331 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18332 and add a sample to openssl.cnf so req -x509 now adds appropriate
18337 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18338 error code, add initial support to X509_print() and x509 application.
18342 * Takes a deep breath and start adding X509 V3 extension support code. Add
18343 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18344 stuff is currently isolated and isn't even compiled yet.
18348 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18349 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18350 Removed the versions check from X509 routines when loading extensions:
18351 this allows certain broken certificates that don't set the version
18352 properly to be processed.
18356 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18357 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18358 can still be regenerated with "make depend".
18362 * Spelling mistake in C version of CAST-128.
18364 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18366 * Changes to the error generation code. The perl script err-code.pl
18367 now reads in the old error codes and retains the old numbers, only
18368 adding new ones if necessary. It also only changes the .err files if new
18369 codes are added. The makefiles have been modified to only insert errors
18370 when needed (to avoid needlessly modifying header files). This is done
18371 by only inserting errors if the .err file is newer than the auto generated
18372 C file. To rebuild all the error codes from scratch (the old behaviour)
18373 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18374 or delete all the .err files.
18378 * CAST-128 was incorrectly implemented for short keys. The C version has
18379 been fixed, but is untested. The assembler versions are also fixed, but
18380 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18381 to regenerate it if needed.
18382 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18383 Hagino <itojun@kame.net>*
18385 * File was opened incorrectly in randfile.c.
18387 *Ulf Möller <ulf@fitug.de>*
18389 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18390 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18391 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18392 al: it's just almost always a UTCTime. Note this patch adds new error
18393 codes so do a "make errors" if there are problems.
18397 * Correct Linux 1 recognition in config.
18399 *Ulf Möller <ulf@fitug.de>*
18401 * Remove pointless MD5 hash when using DSA keys in ca.
18403 *Anonymous <nobody@replay.com>*
18405 * Generate an error if given an empty string as a cert directory. Also
18406 generate an error if handed NULL (previously returned 0 to indicate an
18407 error, but didn't set one).
18409 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18411 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18415 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18416 parameters. This was causing a warning which killed off the Win32 compile.
18420 * Remove C++ style comments from crypto/bn/bn_local.h.
18422 *Neil Costigan <neil.costigan@celocom.com>*
18424 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18425 based on a text string, looking up short and long names and finally
18426 "dot" format. The "dot" format stuff didn't work. Added new function
18427 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18428 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18429 OID is not part of the table.
18433 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18434 X509_LOOKUP_by_alias().
18438 * Sort openssl functions by name.
18442 * Get the `gendsa` command working and add it to the `list` command. Remove
18443 encryption from sample DSA keys (in case anyone is interested the password
18448 * Make *all* `*_free` functions accept a NULL pointer.
18450 *Frans Heymans <fheymans@isaserver.be>*
18452 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18455 *Anonymous <nobody@replay.com>*
18457 * s_server should send the CAfile as acceptable CAs, not its own cert.
18459 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18461 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18463 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18465 * Temp key "for export" tests were wrong in s3_srvr.c.
18467 *Anonymous <nobody@replay.com>*
18469 * Add prototype for temp key callback functions
18470 SSL_CTX_set_tmp_{rsa,dh}_callback().
18474 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18475 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18479 * X509_name_add_entry() freed the wrong thing after an error.
18481 *Arne Ansper <arne@ats.cyber.ee>*
18483 * rsa_eay.c would attempt to free a NULL context.
18485 *Arne Ansper <arne@ats.cyber.ee>*
18487 * BIO_s_socket() had a broken should_retry() on Windoze.
18489 *Arne Ansper <arne@ats.cyber.ee>*
18491 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18493 *Arne Ansper <arne@ats.cyber.ee>*
18495 * Make sure the already existing X509_STORE->depth variable is initialized
18496 in X509_STORE_new(), but document the fact that this variable is still
18497 unused in the certificate verification process.
18499 *Ralf S. Engelschall*
18501 * Fix the various library and `apps/` files to free up pkeys obtained from
18502 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18506 * Fix reference counting in X509_PUBKEY_get(). This makes
18507 demos/maurice/example2.c work, amongst others, probably.
18509 *Steve Henson and Ben Laurie*
18511 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18512 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18513 are no longer created. This way we have a single and consistent command
18514 line interface `openssl <command>`, similar to `cvs <command>`.
18516 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18518 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18519 BIT STRING wrapper always have zero unused bits.
18523 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18527 * Make the top-level INSTALL documentation easier to understand.
18531 * Makefiles updated to exit if an error occurs in a sub-directory
18532 make (including if user presses ^C) [Paul Sutton]
18534 * Make Montgomery context stuff explicit in RSA data structure.
18538 * Fix build order of pem and err to allow for generated pem.h.
18542 * Fix renumbering bug in X509_NAME_delete_entry().
18546 * Enhanced the err-ins.pl script so it makes the error library number
18547 global and can add a library name. This is needed for external ASN1 and
18548 other error libraries.
18552 * Fixed sk_insert which never worked properly.
18556 * Fix ASN1 macros so they can handle indefinite length constructed
18557 EXPLICIT tags. Some non standard certificates use these: they can now
18562 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18563 into a single doc/ssleay.txt bundle. This way the information is still
18564 preserved but no longer messes up this directory. Now it's new room for
18565 the new set of documentation files.
18567 *Ralf S. Engelschall*
18569 * SETs were incorrectly DER encoded. This was a major pain, because they
18570 shared code with SEQUENCEs, which aren't coded the same. This means that
18571 almost everything to do with SETs or SEQUENCEs has either changed name or
18572 number of arguments.
18574 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18576 * Fix test data to work with the above.
18580 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18581 was already fixed by Eric for 0.9.1 it seems.
18583 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18585 * Autodetect FreeBSD3.
18589 * Fix various bugs in Configure. This affects the following platforms:
18593 unixware-2.0-pentium
18598 * Eliminate generated files from CVS. Reorder tests to regenerate files
18599 before they are needed.
18603 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18607 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18609 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18610 changed SSLeay to OpenSSL in version strings.
18612 *Ralf S. Engelschall*
18614 * Some fixups to the top-level documents.
18618 * Fixed the nasty bug where rsaref.h was not found under compile-time
18619 because the symlink to include/ was missing.
18621 *Ralf S. Engelschall*
18623 * Incorporated the popular no-RSA/DSA-only patches
18624 which allow to compile a RSA-free SSLeay.
18626 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18628 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18629 when "ssleay" is still not found.
18631 *Ralf S. Engelschall*
18633 * Added more platforms to Configure: Cray T3E, HPUX 11,
18635 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18637 * Updated the README file.
18639 *Ralf S. Engelschall*
18641 * Added various .cvsignore files in the CVS repository subdirs
18642 to make a "cvs update" really silent.
18644 *Ralf S. Engelschall*
18646 * Recompiled the error-definition header files and added
18647 missing symbols to the Win32 linker tables.
18649 *Ralf S. Engelschall*
18651 * Cleaned up the top-level documents;
18652 o new files: CHANGES and LICENSE
18653 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18654 o merged COPYRIGHT into LICENSE
18655 o removed obsolete TODO file
18656 o renamed MICROSOFT to INSTALL.W32
18658 *Ralf S. Engelschall*
18660 * Removed dummy files from the 0.9.1b source tree:
18662 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18663 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18664 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18665 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18666 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18669 *Ralf S. Engelschall*
18671 * Added various platform portability fixes.
18675 * The Genesis of the OpenSSL rpject:
18676 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18677 Young and Tim J. Hudson created while they were working for C2Net until
18680 *The OpenSSL Project*
18682 ### Changes between 0.9.0b and 0.9.1b [not released]
18684 * Updated a few CA certificates under certs/
18688 * Changed some BIGNUM api stuff.
18692 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18693 DGUX x86, Linux Alpha, etc.
18697 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18698 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18703 * Add -strparse option to asn1pars program which parses nested
18706 *Dr Stephen Henson <shenson@bigfoot.com>*
18708 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18712 * DSA fix for "ca" program.
18716 * Added "-genkey" option to "dsaparam" program.
18720 * Added RIPE MD160 (rmd160) message digest.
18724 * Added -a (all) option to "ssleay version" command.
18728 * Added PLATFORM define which is the id given to Configure.
18732 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18736 * Extended the ASN.1 parser routines.
18740 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18744 * Added a BN_CTX to the BN library.
18748 * Fixed the weak key values in DES library
18752 * Changed API in EVP library for cipher aliases.
18756 * Added support for RC2/64bit cipher.
18760 * Converted the lhash library to the crypto/mem.c functions.
18764 * Added more recognized ASN.1 object ids.
18768 * Added more RSA padding checks for SSL/TLS.
18772 * Added BIO proxy/filter functionality.
18776 * Added extra_certs to SSL_CTX which can be used
18777 send extra CA certificates to the client in the CA cert chain sending
18778 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18782 * Now Fortezza is denied in the authentication phase because
18783 this is key exchange mechanism is not supported by SSLeay at all.
18787 * Additional PKCS1 checks.
18791 * Support the string "TLSv1" for all TLS v1 ciphers.
18795 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18796 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18800 * Fixed a few memory leaks.
18804 * Fixed various code and comment typos.
18808 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18809 bytes sent in the client random.
18811 *Edward Bishop <ebishop@spyglass.com>*
18815 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18816 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18817 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18818 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18819 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18820 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18821 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18822 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18823 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18824 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18825 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18826 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18827 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18828 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18829 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18830 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18831 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18832 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18833 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18834 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18835 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18836 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18837 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18838 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18839 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18840 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18841 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18842 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18843 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18844 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18845 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18846 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18847 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18848 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18849 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18850 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18851 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18852 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18853 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18854 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18855 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18856 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18857 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18858 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18859 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18860 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18861 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18862 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18863 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18864 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18865 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18866 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18867 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18868 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18869 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18870 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18871 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18872 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18873 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18874 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18875 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18876 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18877 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18878 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18879 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18880 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18881 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18882 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18883 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18884 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18885 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18886 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18887 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18888 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18889 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18890 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18891 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18892 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18893 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18894 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18895 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18896 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18897 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18898 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18899 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18900 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18901 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18902 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18903 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18904 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18905 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18906 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18907 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18908 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18909 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18910 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18911 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18912 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18913 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18914 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18915 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18916 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18917 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18918 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18919 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18920 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18921 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18922 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18923 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18924 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18925 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18926 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18927 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18928 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18929 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18930 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18931 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18932 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18933 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18934 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18935 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18936 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18937 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18938 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18939 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18940 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18941 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18942 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18943 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18944 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18945 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18946 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18947 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18948 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18949 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18950 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18951 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18952 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18953 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18954 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18955 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18956 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18957 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18958 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18959 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18960 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18961 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18962 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18963 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18964 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18965 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18966 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18967 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18968 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18969 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18970 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18971 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18972 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18973 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18974 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18975 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18976 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655