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 * The default key generation method for the regular 2-prime RSA keys was
56 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
57 Conditions Based on Auxiliary Probable Primes). This method is slower
58 than the original method.
62 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
63 They are replaced with the BN_check_prime() function that avoids possible
64 misuse and always uses at least 64 rounds of the Miller-Rabin
65 primality test. At least 64 rounds of the Miller-Rabin test are now also
66 used for all prime generation, including RSA key generation.
67 This increases key generation time, especially for larger keys.
71 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
72 as they are not useful with non-deprecated functions.
76 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
77 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
78 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
79 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
80 were used to collect all necessary data to form a HTTP request, and to
81 perform the HTTP transfer with that request. With OpenSSL 3.0, the
82 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
83 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
84 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
85 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
86 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
87 OSSL_HTTP_REQ_CTX_set_max_response_length().
89 *Rich Salz and Richard Levitte*
91 * Validation of SM2 keys has been separated from the validation of regular EC
92 keys, allowing to improve the SM2 validation process to reject loaded private
93 keys that are not conforming to the SM2 ISO standard.
94 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
99 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
100 switches: a validation failure triggers an early exit, returning a failure
101 exit status to the parent process.
105 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
106 to ignore unknown ciphers.
110 * The `-cipher-commands` and `-digest-commands` options
111 of the command line utility `list` have been deprecated.
112 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
116 * All of the low level EC_KEY functions have been deprecated including:
118 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
119 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
120 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
121 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
122 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
123 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
124 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
125 EC_KEY_METHOD_get_verify,
126 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
127 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
128 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
129 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
130 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
131 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
132 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
133 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
134 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
135 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
136 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
137 Applications that need to implement an EC_KEY_METHOD need to consider
138 implementation of the functionality in a special provider.
139 For replacement of the functions manipulating the EC_KEY objects
140 see the EVP_PKEY-EC(7) manual page.
142 Additionally functions that read and write EC_KEY objects such as
143 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
144 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
145 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
146 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
147 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
148 have also been deprecated. Applications should instead use the
149 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
151 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
152 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
153 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
154 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
155 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
157 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
159 * Deprecated all the libcrypto and libssl error string loading
160 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
161 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
162 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
163 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
164 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
165 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
166 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
167 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
168 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
169 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
171 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
172 now loads error strings automatically.
176 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
177 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
178 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
179 are to be used by servers requiring ephemeral DH keys. Instead applications
180 should consider using the built-in DH parameters that are available by
181 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
182 necessary then applications can use the alternative functions
183 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
184 replacement for the "callback" functions. The callback was originally useful
185 in order to have different parameters for export and non-export ciphersuites.
186 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
187 functions should be replaced by one of the other methods described above.
191 * The `-crypt` option to the `passwd` command line tool has been removed.
195 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
200 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
202 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
203 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
204 The inverse ciphers use AES decryption for wrapping, and
205 AES encryption for unwrapping.
209 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
210 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
211 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
212 OpenSSL 3.0 these are replaced by the more generic functions
213 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
214 The old versions have been converted to deprecated macros that just call the
219 * The security callback, which can be customised by application code, supports
220 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
221 in the "other" parameter. In most places this is what is passed. All these
222 places occur server side. However there was one client side call of this
223 security operation and it passed a DH object instead. This is incorrect
224 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
225 of the other locations. Therefore this client side call has been changed to
226 pass an EVP_PKEY instead.
230 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
231 interface. Their functionality remains unchanged.
235 * Added new option for 'openssl list', '-providers', which will display the
236 list of loaded providers, their names, version and status. It optionally
237 displays their gettable parameters.
241 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
242 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
243 type is internally recognised so the workaround is no longer needed.
245 Functionality is still retained as it is, but will only work with
246 EVP_PKEYs with a legacy internal key.
250 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
251 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
255 * Changed all "STACK" functions to be macros instead of inline functions. Macro
256 parameters are still checked for type safety at compile time via helper
261 * Remove the RAND_DRBG API
263 The RAND_DRBG API did not fit well into the new provider concept as
264 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
265 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
266 and some of its API calls are rather low-level. This holds in particular
267 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
269 Adding a compatibility layer to continue supporting the RAND_DRBG API as
270 a legacy API for a regular deprecation period turned out to come at the
271 price of complicating the new provider API unnecessarily. Since the
272 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
275 *Paul Dale and Matthias St. Pierre*
277 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
278 as well as actual hostnames.
282 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
283 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
284 conversely, silently ignore DTLS protocol version bounds when configuring
285 TLS-based contexts. The commands can be repeated to set bounds of both
286 types. The same applies with the corresponding "min_protocol" and
287 "max_protocol" command-line switches, in case some application uses both TLS
290 SSL_CTX instances that are created for a fixed protocol version (e.g.
291 `TLSv1_server_method()`) also silently ignore version bounds. Previously
292 attempts to apply bounds to these protocol versions would result in an
293 error. Now only the "version-flexible" SSL_CTX instances are subject to
294 limits in configuration files in command-line options.
298 * Deprecated the `ENGINE` API. Engines should be replaced with providers
303 * Reworked the recorded ERR codes to make better space for system errors.
304 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
305 given code is a system error (true) or an OpenSSL error (false).
309 * Reworked the test perl framework to better allow parallel testing.
311 *Nicola Tuveri and David von Oheimb*
313 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
314 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
318 * 'Configure' has been changed to figure out the configuration target if
319 none is given on the command line. Consequently, the 'config' script is
320 now only a mere wrapper. All documentation is changed to only mention
323 *Rich Salz and Richard Levitte*
325 * Added a library context `OSSL_LIB_CTX` that applications as well as
326 other libraries can use to form a separate context within which
327 libcrypto operations are performed.
329 There are two ways this can be used:
331 - Directly, by passing a library context to functions that take
332 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
334 - Indirectly, by creating a new library context and then assigning
335 it as the new default, with `OSSL_LIB_CTX_set0_default`.
337 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
338 apart from the functions directly related to `OSSL_LIB_CTX`, accept
339 NULL to indicate that the default library context should be used.
341 Library code that changes the default library context using
342 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
343 second call before returning to the caller.
345 _(Note: the library context was initially called `OPENSSL_CTX` and
346 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
350 * Handshake now fails if Extended Master Secret extension is dropped
355 * Dropped interactive mode from the `openssl` program. From now on,
356 running it without arguments is equivalent to `openssl help`.
360 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
361 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
362 While the old function names have been retained for backward compatibility
363 they should not be used in new developments
364 because their return values are confusing: Unlike other `_cmp()` functions
365 they do not return 0 in case their arguments are equal.
369 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
370 `EC_GROUP_get_field_type()`.
374 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
375 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
376 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
377 Applications should rely on the library automatically assigning a suitable
378 EC_METHOD internally upon EC_GROUP construction.
382 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
383 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
384 assigned internally without application intervention.
385 Users of EC_GROUP_new() should switch to a different suitable constructor.
389 * Add CAdES-BES signature verification support, mostly derived
390 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
392 *Filipe Raimundo da Silva*
394 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
398 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
399 functions are not widely used and now OpenSSL automatically perform this
400 conversion when needed.
404 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
405 EC_KEY_precompute_mult(). These functions are not widely used and
406 applications should instead switch to named curves which OpenSSL has
407 hardcoded lookup tables for.
411 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
412 should instead use the L<EC_POINT_mul(3)> function.
416 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
417 that are not applicable to the new provider model. Applications should
418 instead use EVP_default_properties_is_fips_enabled() and
419 EVP_default_properties_enable_fips().
423 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
424 is set, an unexpected EOF is ignored, it pretends a close notify was received
425 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
429 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
430 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
431 used and applications should instead use the
432 L<EC_POINT_set_affine_coordinates(3)> and
433 L<EC_POINT_get_affine_coordinates(3)> functions.
437 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
438 arrays to be more easily constructed via a series of utility functions.
439 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
440 the various push functions and finally convert to a passable OSSL_PARAM
441 array using OSSL_PARAM_BLD_to_param().
445 * The security strength of SHA1 and MD5 based signatures in TLS has been
446 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
447 working at the default security level of 1 and instead requires security
448 level 0. The security level can be changed either using the cipher string
449 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
453 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
454 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
455 internal keys, if they correspond to one of those built in types.
459 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
460 contain a provider side internal key.
464 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
465 They are old functions that we don't use, and that you could disable with
466 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
470 * Project text documents not yet having a proper file name extension
471 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
472 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
476 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
477 have been converted to Markdown with the goal to produce documents
478 which not only look pretty when viewed online in the browser, but
479 remain well readable inside a plain text editor.
481 To achieve this goal, a 'minimalistic' Markdown style has been applied
482 which avoids formatting elements that interfere too much with the
483 reading flow in the text file. For example, it
485 * avoids [ATX headings][] and uses [setext headings][] instead
486 (which works for `<h1>` and `<h2>` headings only).
487 * avoids [inline links][] and uses [reference links][] instead.
488 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
490 [ATX headings]: https://github.github.com/gfm/#atx-headings
491 [setext headings]: https://github.github.com/gfm/#setext-headings
492 [inline links]: https://github.github.com/gfm/#inline-link
493 [reference links]: https://github.github.com/gfm/#reference-link
494 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
495 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
497 *Matthias St. Pierre*
499 * The test suite is changed to preserve results of each test recipe.
500 A new directory test-runs/ with subdirectories named like the
501 test recipes are created in the build tree for this purpose.
505 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
506 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
507 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
509 *David von Oheimb, Martin Peylo*
511 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
512 The legacy OCSP-focused and only partly documented API is retained for
513 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
517 * Added `util/check-format.pl`, a tool for checking adherence to the
518 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
519 The checks performed are incomplete and yield some false positives.
520 Still the tool should be useful for detecting most typical glitches.
524 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
525 If domain name resolution yields multiple IP addresses all of them are tried
526 after `connect()` failures.
530 * All of the low level RSA functions have been deprecated including:
532 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
533 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
534 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
535 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
536 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
537 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
538 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
539 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
540 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
541 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
542 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
543 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
544 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
545 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
546 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
547 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
548 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
549 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
550 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
551 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
552 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
553 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
554 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
555 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
556 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
557 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
558 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
559 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
561 Use of these low level functions has been informally discouraged for a long
562 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
563 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
564 L<EVP_PKEY_decrypt(3)>.
568 * X509 certificates signed using SHA1 are no longer allowed at security
570 In TLS/SSL the default security level is 1. It can be set either
571 using the cipher string with `@SECLEVEL`, or calling
572 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
573 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
575 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
576 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
577 options of the commands.
581 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
582 modified to use PKEY APIs. These commands are now in maintenance mode
583 and no new features will be added to them.
587 * The command line utility rsautl has been deprecated.
588 Instead use the pkeyutl program.
592 * The command line utilities genrsa and rsa have been modified to use PKEY
593 APIs. They now write PKCS#8 keys by default. These commands are now in
594 maintenance mode and no new features will be added to them.
598 * All of the low level DH functions have been deprecated including:
600 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
601 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
602 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
603 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
604 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
605 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
606 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
607 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
608 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
609 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
610 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
611 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
612 DH_meth_get_generate_params and DH_meth_set_generate_params.
614 Use of these low level functions has been informally discouraged for a long
615 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
616 and L<EVP_PKEY_derive(3)>.
618 Additionally functions that read and write DH objects such as d2i_DHparams,
619 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
620 functions have also been deprecated. Applications should instead use the
621 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
623 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
624 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
625 `EVP_PKEY_set1_DH()` are also deprecated.
626 Applications should instead either read or write an
627 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
628 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
630 *Paul Dale and Matt Caswell*
632 * All of the low level DSA functions have been deprecated including:
634 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
635 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
636 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
637 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
638 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
639 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
640 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
641 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
642 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
643 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
644 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
645 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
646 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
647 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
648 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
649 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
650 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
652 Use of these low level functions has been informally discouraged for a long
653 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
654 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
658 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
659 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
660 This means that applications don't have to look at the curve NID and
661 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
662 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
663 a no-op when the EVP_PKEY is already of the given type.
665 Parameter and key generation is also reworked to make it possible
666 to generate EVP_PKEY_SM2 parameters and keys without having to go
667 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
668 However, code that does the latter will still work as before.
672 * Deprecated low level ECDH and ECDSA functions. These include:
674 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
675 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
678 Use of these low level functions has been informally discouraged for a long
679 time. Instead applications should use the EVP_PKEY_derive(3),
680 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
684 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
685 and EVP_PKEY_decrypt() instead.
686 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
687 and EVP_PKEY_encrypt() instead.
691 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
692 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
693 a new formulation to include all the things it can be used for,
694 as well as words of caution.
698 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
699 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
703 * All of the low level HMAC functions have been deprecated including:
705 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
706 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
709 Use of these low level functions has been informally discouraged for a long
710 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
711 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
712 and L<EVP_MAC_final(3)>.
716 * Over two thousand fixes were made to the documentation, including:
717 - Common options (such as -rand/-writerand, TLS version control, etc)
718 were refactored and point to newly-enhanced descriptions in openssl.pod.
719 - Added style conformance for all options (with help from Richard Levitte),
720 documented all reported missing options, added a CI build to check
721 that all options are documented and that no unimplemented options
723 - Documented some internals, such as all use of environment variables.
724 - Addressed all internal broken L<> references.
728 * All of the low level CMAC functions have been deprecated including:
730 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
731 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
733 Use of these low level functions has been informally discouraged for a long
734 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
735 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
736 and L<EVP_MAC_final(3)>.
740 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
741 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
744 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
745 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
746 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
747 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
748 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
749 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
750 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
751 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
752 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
753 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
755 Use of these low level functions has been informally discouraged
756 for a long time. Applications should use the EVP_DigestInit_ex(3),
757 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
761 * Corrected the documentation of the return values from the `EVP_DigestSign*`
762 set of functions. The documentation mentioned negative values for some
763 errors, but this was never the case, so the mention of negative values
766 Code that followed the documentation and thereby check with something
767 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
771 * All of the low level cipher functions have been deprecated including:
773 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
774 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
775 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
776 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
777 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
778 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
779 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
780 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
781 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
782 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
783 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
784 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
785 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
786 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
787 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
788 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
789 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
790 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
791 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
792 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
793 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
794 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
795 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
796 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
797 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
798 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
799 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
800 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
801 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
803 Use of these low level functions has been informally discouraged for
804 a long time. Applications should use the high level EVP APIs, e.g.
805 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
806 equivalently named decrypt functions instead.
808 *Matt Caswell and Paul Dale*
810 * Removed include/openssl/opensslconf.h.in and replaced it with
811 include/openssl/configuration.h.in, which differs in not including
812 <openssl/macros.h>. A short header include/openssl/opensslconf.h
813 was added to include both.
815 This allows internal hacks where one might need to modify the set
816 of configured macros, for example this if deprecated symbols are
817 still supposed to be available internally:
819 #include <openssl/configuration.h>
821 #undef OPENSSL_NO_DEPRECATED
822 #define OPENSSL_SUPPRESS_DEPRECATED
824 #include <openssl/macros.h>
826 This should not be used by applications that use the exported
827 symbols, as that will lead to linking errors.
831 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
832 used in exponentiation with 512-bit moduli. No EC algorithms are
833 affected. Analysis suggests that attacks against 2-prime RSA1024,
834 3-prime RSA1536, and DSA1024 as a result of this defect would be very
835 difficult to perform and are not believed likely. Attacks against DH512
836 are considered just feasible. However, for an attack the target would
837 have to re-use the DH512 private key, which is not recommended anyway.
838 Also applications directly using the low level API BN_mod_exp may be
839 affected if they use BN_FLG_CONSTTIME.
844 * Most memory-debug features have been deprecated, and the functionality
845 replaced with no-ops.
849 * Added documentation for the STACK API.
853 * Introduced a new method type and API, OSSL_ENCODER, to
854 represent generic encoders. An implementation is expected to
855 be able to encode an object associated with a given name (such
856 as an algorithm name for an asymmetric key) into forms given by
857 implementation properties.
859 Encoders are primarily used from inside libcrypto, through
860 calls to functions like EVP_PKEY_print_private(),
861 PEM_write_bio_PrivateKey() and similar.
863 Encoders are specified in such a way that they can be made to
864 directly handle the provider side portion of an object, if this
865 provider side part comes from the same provider as the encoder
866 itself, but can also be made to handle objects in parametrized
867 form (as an OSSL_PARAM array of data). This allows a provider to
868 offer generic encoders as a service for any other provider.
872 * Added a .pragma directive to the syntax of configuration files, to
873 allow varying behavior in a supported and predictable manner.
874 Currently added pragma:
878 This allows dollar signs to be a keyword character unless it's
879 followed by a opening brace or parenthesis. This is useful for
880 platforms where dollar signs are commonly used in names, such as
881 volume names and system directory names on VMS.
885 * Added functionality to create an EVP_PKEY from user data. This
886 is effectively the same as creating a RSA, DH or DSA object and
887 then assigning them to an EVP_PKEY, but directly using algorithm
888 agnostic EVP functions. A benefit is that this should be future
889 proof for public key algorithms to come.
893 * Change the interpretation of the '--api' configuration option to
894 mean that this is a desired API compatibility level with no
895 further meaning. The previous interpretation, that this would
896 also mean to remove all deprecated symbols up to and including
897 the given version, no requires that 'no-deprecated' is also used
898 in the configuration.
900 When building applications, the desired API compatibility level
901 can be set with the OPENSSL_API_COMPAT macro like before. For
902 API compatibility version below 3.0, the old style numerical
903 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
904 For version 3.0 and on, the value is expected to be the decimal
905 value calculated from the major and minor version like this:
907 MAJOR * 10000 + MINOR * 100
911 -DOPENSSL_API_COMPAT=30000 For 3.0
912 -DOPENSSL_API_COMPAT=30200 For 3.2
914 To hide declarations that are deprecated up to and including the
915 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
916 given when building the application as well.
920 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
921 access to certificate and CRL stores via URIs and OSSL_STORE
924 This adds the following functions:
926 - X509_LOOKUP_store()
927 - X509_STORE_load_file()
928 - X509_STORE_load_path()
929 - X509_STORE_load_store()
930 - SSL_add_store_cert_subjects_to_stack()
931 - SSL_CTX_set_default_verify_store()
932 - SSL_CTX_load_verify_file()
933 - SSL_CTX_load_verify_dir()
934 - SSL_CTX_load_verify_store()
938 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
939 The presence of this system service is determined at run-time.
943 * Added functionality to create an EVP_PKEY context based on data
944 for methods from providers. This takes an algorithm name and a
945 property query string and simply stores them, with the intent
946 that any operation that uses this context will use those strings
947 to fetch the needed methods implicitly, thereby making the port
948 of application written for pre-3.0 OpenSSL easier.
952 * The undocumented function NCONF_WIN32() has been deprecated; for
953 conversion details see the HISTORY section of doc/man5/config.pod
957 * Introduced the new functions EVP_DigestSignInit_ex() and
958 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
959 EVP_DigestVerifyUpdate() have been converted to functions. See the man
960 pages for further details.
964 * Over two thousand fixes were made to the documentation, including:
965 adding missing command flags, better style conformance, documentation
968 *Rich Salz, Richard Levitte*
970 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
971 X25519, X448, Ed25519 and Ed448.
975 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
980 * Deprecated the public definition of `ERR_STATE` as well as the function
981 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
986 * Added ERR functionality to give callers access to the stored function
987 names that have replaced the older function code based functions.
989 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
990 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
991 ERR_peek_error_all() and ERR_peek_last_error_all().
993 These functions have become deprecated: ERR_get_error_line(),
994 ERR_get_error_line_data(), ERR_peek_error_line_data(),
995 ERR_peek_last_error_line_data() and ERR_func_error_string().
997 Users are recommended to use ERR_get_error_all(), or to pick information
998 with ERR_peek functions and finish off with getting the error code by using
1003 * Extended testing to be verbose for failing tests only. The make variables
1004 VERBOSE_FAILURE or VF can be used to enable this:
1006 $ make VF=1 test # Unix
1007 $ mms /macro=(VF=1) test ! OpenVMS
1008 $ nmake VF=1 test # Windows
1012 * Added the `-copy_extensions` option to the `x509` command for use with
1013 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1014 all extensions in the request are copied to the certificate or vice versa.
1016 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1018 * Added the `-copy_extensions` option to the `req` command for use with
1019 `-x509`. When given with the `copy` or `copyall` argument,
1020 all extensions in the certification request are copied to the certificate.
1024 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1025 they generate are by default RFC 5280 compliant in the following sense:
1026 There is a subjectKeyIdentifier extension with a hash value of the public key
1027 and for not self-signed certs there is an authorityKeyIdentifier extension
1028 with a keyIdentifier field or issuer information identifying the signing key.
1029 This is done unless some configuration overrides the new default behavior,
1030 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1034 * Added several checks to `X509_verify_cert()` according to requirements in
1035 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1036 (which may be done by using the CLI option `-x509_strict`):
1037 * The basicConstraints of CA certificates must be marked critical.
1038 * CA certificates must explicitly include the keyUsage extension.
1039 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1040 * The issuer name of any certificate must not be empty.
1041 * The subject name of CA certs, certs with keyUsage crlSign,
1042 and certs without subjectAlternativeName must not be empty.
1043 * If a subjectAlternativeName extension is given it must not be empty.
1044 * The signatureAlgorithm field and the cert signature must be consistent.
1045 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1046 must not be marked critical.
1047 * The authorityKeyIdentifier must be given for X.509v3 certs
1048 unless they are self-signed.
1049 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1053 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1054 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1058 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1059 used even when parsing explicit parameters, when loading a encoded key
1060 or calling `EC_GROUP_new_from_ecpkparameters()`/
1061 `EC_GROUP_new_from_ecparameters()`.
1062 This prevents bypass of security hardening and performance gains,
1063 especially for curves with specialized EC_METHODs.
1064 By default, if a key encoded with explicit parameters is loaded and later
1065 encoded, the output is still encoded with explicit parameters, even if
1066 internally a "named" EC_GROUP is used for computation.
1070 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1071 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1072 NULL. After this change, only the cofactor parameter can be NULL. It also
1073 does some minimal sanity checks on the passed order.
1078 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1079 An attack is simple, if the first CMS_recipientInfo is valid but the
1080 second CMS_recipientInfo is chosen ciphertext. If the second
1081 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1082 encryption key will be replaced by garbage, and the message cannot be
1083 decoded, but if the RSA decryption fails, the correct encryption key is
1084 used and the recipient will not notice the attack.
1085 As a work around for this potential attack the length of the decrypted
1086 key must be equal to the cipher default key length, in case the
1087 certifiate is not given and all recipientInfo are tried out.
1088 The old behaviour can be re-enabled in the CMS code by setting the
1089 CMS_DEBUG_DECRYPT flag.
1093 * Early start up entropy quality from the DEVRANDOM seed source has been
1094 improved for older Linux systems. The RAND subsystem will wait for
1095 /dev/random to be producing output before seeding from /dev/urandom.
1096 The seeded state is stored for future library initialisations using
1097 a system global shared memory segment. The shared memory identifier
1098 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1099 the desired value. The default identifier is 114.
1103 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1104 when primes for RSA keys are computed.
1105 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1106 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1107 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1108 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1109 This avoids possible fingerprinting of newly generated RSA modules.
1113 * Correct the extended master secret constant on EBCDIC systems. Without this
1114 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1115 negotiate EMS will fail. Unfortunately this also means that TLS connections
1116 between EBCDIC systems with this fix, and EBCDIC systems without this
1117 fix will fail if they negotiate EMS.
1121 * Changed the library initialisation so that the config file is now loaded
1122 by default. This was already the case for libssl. It now occurs for both
1123 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1124 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1128 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1129 where the former acts as a replacement for `ERR_put_error()`, and the
1130 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1131 `ERR_raise_data()` adds more flexibility by taking a format string and
1132 an arbitrary number of arguments following it, to be processed with
1137 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1138 to check if a named provider is loaded and available. When called, it
1139 will also activate all fallback providers if such are still present.
1143 * Enforce a minimum DH modulus size of 512 bits.
1147 * Changed DH parameters to generate the order q subgroup instead of 2q.
1148 Previously generated DH parameters are still accepted by DH_check
1149 but DH_generate_key works around that by clearing bit 0 of the
1150 private key for those. This avoids leaking bit 0 of the private key.
1154 * Significantly reduce secure memory usage by the randomness pools.
1158 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1163 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1164 algorithms. An implementation of a key exchange algorithm can be obtained
1165 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1166 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1167 the older EVP_PKEY_derive_init() function. See the man pages for the new
1168 functions for further details.
1172 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1176 * Removed the function names from error messages and deprecated the
1179 * Removed NextStep support and the macro OPENSSL_UNISTD
1183 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1184 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1185 Also removed "export var as function" capability; we do not export
1186 variables, only functions.
1190 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1191 an error and 1 indicating success. In previous versions of OpenSSL this
1192 was a void type. If a key was set longer than the maximum possible this
1197 * Support SM2 signing and verification schemes with X509 certificate.
1201 * Use SHA256 as the default digest for TS query in the `ts` app.
1205 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1206 This checks that the salt length is at least 128 bits, the derived key
1207 length is at least 112 bits, and that the iteration count is at least 1000.
1208 For backwards compatibility these checks are disabled by default in the
1209 default provider, but are enabled by default in the fips provider.
1210 To enable or disable these checks use the control
1211 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1215 * Default cipher lists/suites are now available via a function, the
1216 #defines are deprecated.
1220 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1221 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1222 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1226 * Join the directories crypto/x509 and crypto/x509v3
1230 * Added command 'openssl kdf' that uses the EVP_KDF API.
1234 * Added command 'openssl mac' that uses the EVP_MAC API.
1238 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1239 as default directories. Also added the command 'openssl info'
1240 for scripting purposes.
1244 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1245 deprecated. These undocumented functions were never integrated into the EVP
1246 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1247 Bi-directional IGE mode. These modes were never formally standardised and
1248 usage of these functions is believed to be very small. In particular
1249 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1250 is ever used. The security implications are believed to be minimal, but
1251 this issue was never fixed for backwards compatibility reasons. New code
1252 should not use these modes.
1256 * Add prediction resistance to the DRBG reseeding process.
1260 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1261 mandated by IEEE Std 1619-2018.
1265 * Added newline escaping functionality to a filename when using openssl dgst.
1266 This output format is to replicate the output format found in the `*sum`
1267 checksum programs. This aims to preserve backward compatibility.
1269 *Matt Eaton, Richard Levitte, and Paul Dale*
1271 * Removed the heartbeat message in DTLS feature, as it has very
1272 little usage and doesn't seem to fulfill a valuable purpose.
1273 The configuration option is now deprecated.
1277 * Changed the output of 'openssl {digestname} < file' to display the
1278 digest name in its output.
1282 * Added a new generic trace API which provides support for enabling
1283 instrumentation through trace output. This feature is mainly intended
1284 as an aid for developers and is disabled by default. To utilize it,
1285 OpenSSL needs to be configured with the `enable-trace` option.
1287 If the tracing API is enabled, the application can activate trace output
1288 by registering BIOs as trace channels for a number of tracing and debugging
1291 The `openssl` program has been expanded to enable any of the types
1292 available via environment variables defined by the user, and serves as
1293 one possible example on how to use this functionality.
1295 *Richard Levitte & Matthias St. Pierre*
1297 * Added build tests for C++. These are generated files that only do one
1298 thing, to include one public OpenSSL head file each. This tests that
1299 the public header files can be usefully included in a C++ application.
1301 This test isn't enabled by default. It can be enabled with the option
1302 'enable-buildtest-c++'.
1306 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1310 * Add KMAC to EVP_MAC.
1314 * Added property based algorithm implementation selection framework to
1319 * Added SCA hardening for modular field inversion in EC_GROUP through
1320 a new dedicated field_inv() pointer in EC_METHOD.
1321 This also addresses a leakage affecting conversions from projective
1322 to affine coordinates.
1324 *Billy Bob Brumley, Nicola Tuveri*
1326 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1327 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1328 those algorithms that were already supported through the EVP_PKEY API
1329 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1330 and scrypt are now wrappers that call EVP_KDF.
1334 * Build devcrypto engine as a dynamic engine.
1336 *Eneas U de Queiroz*
1338 * Add keyed BLAKE2 to EVP_MAC.
1342 * Fix a bug in the computation of the endpoint-pair shared secret used
1343 by DTLS over SCTP. This breaks interoperability with older versions
1344 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1345 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1346 interoperability with such broken implementations. However, enabling
1347 this switch breaks interoperability with correct implementations.
1349 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1350 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1354 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1358 * Change the license to the Apache License v2.0.
1362 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1364 - Major releases (indicated by incrementing the MAJOR release number)
1365 may introduce incompatible API/ABI changes.
1366 - Minor releases (indicated by incrementing the MINOR release number)
1367 may introduce new features but retain API/ABI compatibility.
1368 - Patch releases (indicated by incrementing the PATCH number)
1369 are intended for bug fixes and other improvements of existing
1370 features only (like improving performance or adding documentation)
1371 and retain API/ABI compatibility.
1375 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1379 * Remove the 'dist' target and add a tarball building script. The
1380 'dist' target has fallen out of use, and it shouldn't be
1381 necessary to configure just to create a source distribution.
1385 * Recreate the OS390-Unix config target. It no longer relies on a
1386 special script like it did for OpenSSL pre-1.1.0.
1390 * Instead of having the source directories listed in Configure, add
1391 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1396 * Add GMAC to EVP_MAC.
1400 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1404 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1405 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1406 to facilitate the continued use of MACs through raw private keys in
1407 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1411 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1412 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1416 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1417 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1418 are retained for backwards compatibility.
1422 * AES-XTS mode now enforces that its two keys are different to mitigate
1423 the attacked described in "Efficient Instantiations of Tweakable
1424 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1425 Details of this attack can be obtained from:
1426 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1430 * Rename the object files, i.e. give them other names than in previous
1431 versions. Their names now include the name of the final product, as
1432 well as its type mnemonic (bin, lib, shlib).
1436 * Added new option for 'openssl list', '-objects', which will display the
1437 list of built in objects, i.e. OIDs with names.
1441 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1442 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1447 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1448 improves application performance by removing data copies and providing
1449 applications with zero-copy system calls such as sendfile and splice.
1453 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1454 option is set, openssl cleanses (zeroize) plaintext bytes from
1455 internal buffers after delivering them to the application. Note,
1456 the application is still responsible for cleansing other copies
1457 (e.g.: data received by SSL_read(3)).
1461 * `PKCS12_parse` now maintains the order of the parsed certificates
1462 when outputting them via `*ca` (rather than reversing it).
1466 * Deprecated pthread fork support methods. These were unused so no
1467 replacement is required.
1469 - OPENSSL_fork_prepare()
1470 - OPENSSL_fork_parent()
1471 - OPENSSL_fork_child()
1478 ### Changes between 1.1.1i and 1.1.1j [xx XXX xxxx]
1480 * Fixed SRP_Calc_client_key so that it uses constant time. The previous
1481 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1482 could be exploited in a side channel attack to recover the password. Since
1483 the attack is local host only this is outside of the current OpenSSL
1484 threat model and therefore no CVE is assigned.
1486 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1491 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1493 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1494 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1495 If an attacker can control both items being compared then this could lead
1496 to a possible denial of service attack. OpenSSL itself uses the
1497 GENERAL_NAME_cmp function for two purposes:
1498 1) Comparing CRL distribution point names between an available CRL and a
1499 CRL distribution point embedded in an X509 certificate
1500 2) When verifying that a timestamp response token signer matches the
1501 timestamp authority name (exposed via the API functions
1502 TS_RESP_verify_response and TS_RESP_verify_token)
1507 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1509 * Certificates with explicit curve parameters are now disallowed in
1510 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1514 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1515 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1516 conversely, silently ignore DTLS protocol version bounds when configuring
1517 TLS-based contexts. The commands can be repeated to set bounds of both
1518 types. The same applies with the corresponding "min_protocol" and
1519 "max_protocol" command-line switches, in case some application uses both TLS
1522 SSL_CTX instances that are created for a fixed protocol version (e.g.
1523 TLSv1_server_method()) also silently ignore version bounds. Previously
1524 attempts to apply bounds to these protocol versions would result in an
1525 error. Now only the "version-flexible" SSL_CTX instances are subject to
1526 limits in configuration files in command-line options.
1530 * Handshake now fails if Extended Master Secret extension is dropped
1535 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1537 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1539 * Fixed segmentation fault in SSL_check_chain()
1540 Server or client applications that call the SSL_check_chain() function
1541 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1542 dereference as a result of incorrect handling of the
1543 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1544 or unrecognised signature algorithm is received from the peer. This could
1545 be exploited by a malicious peer in a Denial of Service attack.
1550 * Added AES consttime code for no-asm configurations
1551 an optional constant time support for AES was added
1552 when building openssl for no-asm.
1553 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1554 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1555 At this time this feature is by default disabled.
1556 It will be enabled by default in 3.0.
1560 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1562 * Revert the change of EOF detection while reading in libssl to avoid
1563 regressions in applications depending on the current way of reporting
1564 the EOF. As the existing method is not fully accurate the change to
1565 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1566 branch and will be present in the 3.0 release.
1570 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1571 when primes for RSA keys are computed.
1572 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1573 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1574 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1575 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1576 This avoids possible fingerprinting of newly generated RSA modules.
1580 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1582 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1583 while reading in libssl then we would report an error back to the
1584 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1585 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1586 therefore give a hint as to what went wrong.
1590 * Check that ed25519 and ed448 are allowed by the security level. Previously
1591 signature algorithms not using an MD were not being checked that they were
1592 allowed by the security level.
1596 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1597 was not quite right. The behaviour was not consistent between resumption
1598 and normal handshakes, and also not quite consistent with historical
1599 behaviour. The behaviour in various scenarios has been clarified and
1600 it has been updated to make it match historical behaviour as closely as
1605 * *[VMS only]* The header files that the VMS compilers include automatically,
1606 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1607 that the C++ compiler doesn't understand. This is a shortcoming in the
1608 compiler, but can be worked around with `__cplusplus` guards.
1610 C++ applications that use OpenSSL libraries must be compiled using the
1611 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1612 functions. Otherwise, only functions with symbols of less than 31
1613 characters can be used, as the linker will not be able to successfully
1614 resolve symbols with longer names.
1618 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1619 The presence of this system service is determined at run-time.
1623 * Added newline escaping functionality to a filename when using openssl dgst.
1624 This output format is to replicate the output format found in the `*sum`
1625 checksum programs. This aims to preserve backward compatibility.
1627 *Matt Eaton, Richard Levitte, and Paul Dale*
1629 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1634 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1636 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1637 number generator (RNG). This was intended to include protection in the
1638 event of a fork() system call in order to ensure that the parent and child
1639 processes did not share the same RNG state. However this protection was not
1640 being used in the default case.
1642 A partial mitigation for this issue is that the output from a high
1643 precision timer is mixed into the RNG state so the likelihood of a parent
1644 and child process sharing state is significantly reduced.
1646 If an application already calls OPENSSL_init_crypto() explicitly using
1647 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1650 *Matthias St. Pierre*
1652 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1653 used even when parsing explicit parameters, when loading a encoded key
1654 or calling `EC_GROUP_new_from_ecpkparameters()`/
1655 `EC_GROUP_new_from_ecparameters()`.
1656 This prevents bypass of security hardening and performance gains,
1657 especially for curves with specialized EC_METHODs.
1658 By default, if a key encoded with explicit parameters is loaded and later
1659 encoded, the output is still encoded with explicit parameters, even if
1660 internally a "named" EC_GROUP is used for computation.
1664 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1665 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1666 NULL. After this change, only the cofactor parameter can be NULL. It also
1667 does some minimal sanity checks on the passed order.
1672 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1673 An attack is simple, if the first CMS_recipientInfo is valid but the
1674 second CMS_recipientInfo is chosen ciphertext. If the second
1675 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1676 encryption key will be replaced by garbage, and the message cannot be
1677 decoded, but if the RSA decryption fails, the correct encryption key is
1678 used and the recipient will not notice the attack.
1679 As a work around for this potential attack the length of the decrypted
1680 key must be equal to the cipher default key length, in case the
1681 certifiate is not given and all recipientInfo are tried out.
1682 The old behaviour can be re-enabled in the CMS code by setting the
1683 CMS_DEBUG_DECRYPT flag.
1688 * Early start up entropy quality from the DEVRANDOM seed source has been
1689 improved for older Linux systems. The RAND subsystem will wait for
1690 /dev/random to be producing output before seeding from /dev/urandom.
1691 The seeded state is stored for future library initialisations using
1692 a system global shared memory segment. The shared memory identifier
1693 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1694 the desired value. The default identifier is 114.
1698 * Correct the extended master secret constant on EBCDIC systems. Without this
1699 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1700 negotiate EMS will fail. Unfortunately this also means that TLS connections
1701 between EBCDIC systems with this fix, and EBCDIC systems without this
1702 fix will fail if they negotiate EMS.
1706 * Use Windows installation paths in the mingw builds
1708 Mingw isn't a POSIX environment per se, which means that Windows
1709 paths should be used for installation.
1714 * Changed DH_check to accept parameters with order q and 2q subgroups.
1715 With order 2q subgroups the bit 0 of the private key is not secret
1716 but DH_generate_key works around that by clearing bit 0 of the
1717 private key for those. This avoids leaking bit 0 of the private key.
1721 * Significantly reduce secure memory usage by the randomness pools.
1725 * Revert the DEVRANDOM_WAIT feature for Linux systems
1727 The DEVRANDOM_WAIT feature added a select() call to wait for the
1728 /dev/random device to become readable before reading from the
1729 /dev/urandom device.
1731 It turned out that this change had negative side effects on
1732 performance which were not acceptable. After some discussion it
1733 was decided to revert this feature and leave it up to the OS
1734 resp. the platform maintainer to ensure a proper initialization
1735 during early boot time.
1737 *Matthias St. Pierre*
1739 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1741 * Add build tests for C++. These are generated files that only do one
1742 thing, to include one public OpenSSL head file each. This tests that
1743 the public header files can be usefully included in a C++ application.
1745 This test isn't enabled by default. It can be enabled with the option
1746 'enable-buildtest-c++'.
1750 * Enable SHA3 pre-hashing for ECDSA and DSA.
1754 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1755 This changes the size when using the `genpkey` command when no size is given.
1756 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1757 generation commands to use 2048 bits by default.
1761 * Reorganize the manual pages to consistently have RETURN VALUES,
1762 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1763 util/fix-doc-nits accordingly.
1765 *Paul Yang, Joshua Lock*
1767 * Add the missing accessor EVP_PKEY_get0_engine()
1771 * Have commands like `s_client` and `s_server` output the signature scheme
1772 along with other cipher suite parameters when debugging.
1776 * Make OPENSSL_config() error agnostic again.
1780 * Do the error handling in RSA decryption constant time.
1784 * Prevent over long nonces in ChaCha20-Poly1305.
1786 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1787 for every encryption operation. RFC 7539 specifies that the nonce value
1788 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1789 and front pads the nonce with 0 bytes if it is less than 12
1790 bytes. However it also incorrectly allows a nonce to be set of up to 16
1791 bytes. In this case only the last 12 bytes are significant and any
1792 additional leading bytes are ignored.
1794 It is a requirement of using this cipher that nonce values are
1795 unique. Messages encrypted using a reused nonce value are susceptible to
1796 serious confidentiality and integrity attacks. If an application changes
1797 the default nonce length to be longer than 12 bytes and then makes a
1798 change to the leading bytes of the nonce expecting the new value to be a
1799 new unique nonce then such an application could inadvertently encrypt
1800 messages with a reused nonce.
1802 Additionally the ignored bytes in a long nonce are not covered by the
1803 integrity guarantee of this cipher. Any application that relies on the
1804 integrity of these ignored leading bytes of a long nonce may be further
1805 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1806 is safe because no such use sets such a long nonce value. However user
1807 applications that use this cipher directly and set a non-default nonce
1808 length to be longer than 12 bytes may be vulnerable.
1810 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1816 * Add DEVRANDOM_WAIT feature for Linux systems
1818 On older Linux systems where the getrandom() system call is not available,
1819 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1820 Contrary to getrandom(), the /dev/urandom device will not block during
1821 early boot when the kernel CSPRNG has not been seeded yet.
1823 To mitigate this known weakness, use select() to wait for /dev/random to
1824 become readable before reading from /dev/urandom.
1826 * Ensure that SM2 only uses SM3 as digest algorithm
1830 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1832 * Change the info callback signals for the start and end of a post-handshake
1833 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1834 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1835 confused by this and assume that a TLSv1.2 renegotiation has started. This
1836 can break KeyUpdate handling. Instead we no longer signal the start and end
1837 of a post handshake message exchange (although the messages themselves are
1838 still signalled). This could break some applications that were expecting
1839 the old signals. However without this KeyUpdate is not usable for many
1844 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1846 * Timing vulnerability in DSA signature generation
1848 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1849 timing side channel attack. An attacker could use variations in the signing
1850 algorithm to recover the private key.
1852 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1857 * Timing vulnerability in ECDSA signature generation
1859 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1860 timing side channel attack. An attacker could use variations in the signing
1861 algorithm to recover the private key.
1863 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1868 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1869 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1870 of two gigabytes and the error handling improved.
1872 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1873 categorized as a normal bug, not a security issue, because the DRBG reseeds
1874 automatically and is fully functional even without additional randomness
1875 provided by the application.
1877 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1879 * Add a new ClientHello callback. Provides a callback interface that gives
1880 the application the ability to adjust the nascent SSL object at the
1881 earliest stage of ClientHello processing, immediately after extensions have
1882 been collected but before they have been processed. In particular, this
1883 callback can adjust the supported TLS versions in response to the contents
1888 * Add SM2 base algorithm support.
1892 * s390x assembly pack: add (improved) hardware-support for the following
1893 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1894 aes-cfb/cfb8, aes-ecb.
1898 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1899 parameter is no longer accepted, as it leads to a corrupt table. NULL
1900 pem_str is reserved for alias entries only.
1904 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1905 step for prime curves. The new implementation is based on formulae from
1906 differential addition-and-doubling in homogeneous projective coordinates
1907 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1908 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1909 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1910 to work in projective coordinates.
1912 *Billy Bob Brumley, Nicola Tuveri*
1914 * Change generating and checking of primes so that the error rate of not
1915 being prime depends on the intended use based on the size of the input.
1916 For larger primes this will result in more rounds of Miller-Rabin.
1917 The maximal error rate for primes with more than 1080 bits is lowered
1920 *Kurt Roeckx, Annie Yousar*
1922 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1926 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1927 moving between systems, and to avoid confusion when a Windows build is
1928 done with mingw vs with MSVC. For POSIX installs, there's still a
1929 symlink or copy named 'tsget' to avoid that confusion as well.
1933 * Revert blinding in ECDSA sign and instead make problematic addition
1934 length-invariant. Switch even to fixed-length Montgomery multiplication.
1938 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1939 step for binary curves. The new implementation is based on formulae from
1940 differential addition-and-doubling in mixed Lopez-Dahab projective
1941 coordinates, modified to independently blind the operands.
1943 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1945 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1946 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1947 EC_METHODs to implement their own specialized "ladder step", to take
1948 advantage of more favorable coordinate systems or more efficient
1949 differential addition-and-doubling algorithms.
1951 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1953 * Modified the random device based seed sources to keep the relevant
1954 file descriptors open rather than reopening them on each access.
1955 This allows such sources to operate in a chroot() jail without
1956 the associated device nodes being available. This behaviour can be
1957 controlled using RAND_keep_random_devices_open().
1961 * Numerous side-channel attack mitigations have been applied. This may have
1962 performance impacts for some algorithms for the benefit of improved
1963 security. Specific changes are noted in this change log by their respective
1968 * AIX shared library support overhaul. Switch to AIX "natural" way of
1969 handling shared libraries, which means collecting shared objects of
1970 different versions and bitnesses in one common archive. This allows to
1971 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1972 doesn't affect the way 3rd party applications are linked, only how
1973 multi-version installation is managed.
1977 * Make ec_group_do_inverse_ord() more robust and available to other
1978 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1979 mitigations are applied to the fallback BN_mod_inverse().
1980 When using this function rather than BN_mod_inverse() directly, new
1981 EC cryptosystem implementations are then safer-by-default.
1985 * Add coordinate blinding for EC_POINT and implement projective
1986 coordinate blinding for generic prime curves as a countermeasure to
1987 chosen point SCA attacks.
1989 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1991 * Add blinding to ECDSA and DSA signatures to protect against side channel
1992 attacks discovered by Keegan Ryan (NCC Group).
1996 * Enforce checking in the `pkeyutl` command to ensure that the input
1997 length does not exceed the maximum supported digest length when performing
1998 a sign, verify or verifyrecover operation.
2002 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2003 I/O in combination with something like select() or poll() will hang. This
2004 can be turned off again using SSL_CTX_clear_mode().
2005 Many applications do not properly handle non-application data records, and
2006 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2007 around the problems in those applications, but can also break some.
2008 It's recommended to read the manpages about SSL_read(), SSL_write(),
2009 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2010 SSL_CTX_set_read_ahead() again.
2014 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2015 now allow empty (zero character) pass phrases.
2019 * Apply blinding to binary field modular inversion and remove patent
2020 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2024 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2025 binary and prime elliptic curves.
2029 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2030 constant time fixed point multiplication.
2034 * Revise elliptic curve scalar multiplication with timing attack
2035 defenses: ec_wNAF_mul redirects to a constant time implementation
2036 when computing fixed point and variable point multiplication (which
2037 in OpenSSL are mostly used with secret scalars in keygen, sign,
2038 ECDH derive operations).
2039 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2042 * Updated CONTRIBUTING
2046 * Updated DRBG / RAND to request nonce and additional low entropy
2047 randomness from the system.
2049 *Matthias St. Pierre*
2051 * Updated 'openssl rehash' to use OpenSSL consistent default.
2055 * Moved the load of the ssl_conf module to libcrypto, which helps
2056 loading engines that libssl uses before libssl is initialised.
2060 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2064 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2066 *Ingo Schwarze, Rich Salz*
2068 * Added output of accepting IP address and port for 'openssl s_server'
2072 * Added a new API for TLSv1.3 ciphersuites:
2073 SSL_CTX_set_ciphersuites()
2074 SSL_set_ciphersuites()
2078 * Memory allocation failures consistently add an error to the error
2083 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2084 in libcrypto when run as setuid/setgid.
2088 * Load any config file by default when libssl is used.
2092 * Added new public header file <openssl/rand_drbg.h> and documentation
2093 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2095 *Matthias St. Pierre*
2097 * QNX support removed (cannot find contributors to get their approval
2098 for the license change).
2102 * TLSv1.3 replay protection for early data has been implemented. See the
2103 SSL_read_early_data() man page for further details.
2107 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2108 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2109 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2110 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2111 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2112 configuration has been separated out. See the ciphers man page or the
2113 SSL_CTX_set_ciphersuites() man page for more information.
2117 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2118 in responder mode now supports the new "-multi" option, which
2119 spawns the specified number of child processes to handle OCSP
2120 requests. The "-timeout" option now also limits the OCSP
2121 responder's patience to wait to receive the full client request
2122 on a newly accepted connection. Child processes are respawned
2123 as needed, and the CA index file is automatically reloaded
2124 when changed. This makes it possible to run the "ocsp" responder
2125 as a long-running service, making the OpenSSL CA somewhat more
2126 feature-complete. In this mode, most diagnostic messages logged
2127 after entering the event loop are logged via syslog(3) rather than
2132 * Added support for X448 and Ed448. Heavily based on original work by
2137 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2138 objects loaded. This adds the functions OSSL_STORE_expect() and
2139 OSSL_STORE_find() as well as needed tools to construct searches and
2140 get the search data out of them.
2144 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2145 version of OpenSSL should review their configuration settings to ensure
2146 that they are still appropriate for TLSv1.3. For further information see:
2147 <https://wiki.openssl.org/index.php/TLS1.3>
2151 * Grand redesign of the OpenSSL random generator
2153 The default RAND method now utilizes an AES-CTR DRBG according to
2154 NIST standard SP 800-90Ar1. The new random generator is essentially
2155 a port of the default random generator from the OpenSSL FIPS 2.0
2156 object module. It is a hybrid deterministic random bit generator
2157 using an AES-CTR bit stream and which seeds and reseeds itself
2158 automatically using trusted system entropy sources.
2160 Some of its new features are:
2161 - Support for multiple DRBG instances with seed chaining.
2162 - The default RAND method makes use of a DRBG.
2163 - There is a public and private DRBG instance.
2164 - The DRBG instances are fork-safe.
2165 - Keep all global DRBG instances on the secure heap if it is enabled.
2166 - The public and private DRBG instance are per thread for lock free
2169 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2171 * Changed Configure so it only says what it does and doesn't dump
2172 so much data. Instead, ./configdata.pm should be used as a script
2173 to display all sorts of configuration data.
2177 * Added processing of "make variables" to Configure.
2181 * Added SHA512/224 and SHA512/256 algorithm support.
2185 * The last traces of Netware support, first removed in 1.1.0, have
2190 * Get rid of Makefile.shared, and in the process, make the processing
2191 of certain files (rc.obj, or the .def/.map/.opt files produced from
2192 the ordinal files) more visible and hopefully easier to trace and
2193 debug (or make silent).
2197 * Make it possible to have environment variable assignments as
2198 arguments to config / Configure.
2202 * Add multi-prime RSA (RFC 8017) support.
2206 * Add SM3 implemented according to GB/T 32905-2016
2207 *Jack Lloyd <jack.lloyd@ribose.com>,*
2208 *Ronald Tse <ronald.tse@ribose.com>,*
2209 *Erick Borsboom <erick.borsboom@ribose.com>*
2211 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2212 as documented in RFC6066.
2213 Based on a patch from Tomasz Moń
2215 *Filipe Raimundo da Silva*
2217 * Add SM4 implemented according to GB/T 32907-2016.
2218 *Jack Lloyd <jack.lloyd@ribose.com>,*
2219 *Ronald Tse <ronald.tse@ribose.com>,*
2220 *Erick Borsboom <erick.borsboom@ribose.com>*
2222 * Reimplement -newreq-nodes and ERR_error_string_n; the
2223 original author does not agree with the license change.
2227 * Add ARIA AEAD TLS support.
2231 * Some macro definitions to support VS6 have been removed. Visual
2232 Studio 6 has not worked since 1.1.0
2236 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2237 without clearing the errors.
2241 * Add "atfork" functions. If building on a system that without
2242 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2243 requirements. The RAND facility now uses/requires this.
2251 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2252 not possible to disable entirely. However, it's still possible to
2253 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2256 To disable, configure with 'no-ui-console'. 'no-ui' is still
2257 possible to use as an alias. Check at compile time with the
2258 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2259 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2263 * Add a STORE module, which implements a uniform and URI based reader of
2264 stores that can contain keys, certificates, CRLs and numerous other
2265 objects. The main API is loosely based on a few stdio functions,
2266 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2267 OSSL_STORE_error and OSSL_STORE_close.
2268 The implementation uses backends called "loaders" to implement arbitrary
2269 URI schemes. There is one built in "loader" for the 'file' scheme.
2273 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2274 then adjusted to work on FreeBSD 8.4 as well.
2275 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2276 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2280 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2281 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2282 error code calls like this:
2284 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2286 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2287 that can be encoded in C. For the foreseeable future, this will only
2290 *Richard Levitte and Tim Hudson*
2292 * Removed BSD cryptodev engine.
2296 * Add a build target 'build_all_generated', to build all generated files
2297 and only that. This can be used to prepare everything that requires
2298 things like perl for a system that lacks perl and then move everything
2299 to that system and do the rest of the build there.
2303 * In the UI interface, make it possible to duplicate the user data. This
2304 can be used by engines that need to retain the data for a longer time
2305 than just the call where this user data is passed.
2309 * Ignore the '-named_curve auto' value for compatibility of applications
2312 *Tomáš Mráz <tmraz@fedoraproject.org>*
2314 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2315 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2316 alerts across multiple records (some of which could be empty). In practice
2317 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2318 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2319 support this at all. Supporting it adds significant complexity to the
2320 record layer, and its removal is unlikely to cause interoperability
2325 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2326 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2327 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2332 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2333 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2335 *Richard Levitte, Andy Polyakov*
2337 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2342 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2343 platform rather than 'mingw'.
2347 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2348 success if they are asked to add an object which already exists
2349 in the store. This change cascades to other functions which load
2350 certificates and CRLs.
2354 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2355 facilitate stack unwinding even from assembly subroutines.
2359 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2360 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2364 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2365 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2366 which is the minimum version we support.
2370 * Certificate time validation (X509_cmp_time) enforces stricter
2371 compliance with RFC 5280. Fractional seconds and timezone offsets
2372 are no longer allowed.
2376 * Add support for ARIA
2380 * s_client will now send the Server Name Indication (SNI) extension by
2381 default unless the new "-noservername" option is used. The server name is
2382 based on the host provided to the "-connect" option unless overridden by
2383 using "-servername".
2387 * Add support for SipHash
2391 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2392 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2393 prevent issues where no progress is being made and the peer continually
2394 sends unrecognised record types, using up resources processing them.
2398 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2399 using the algorithm defined in
2400 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2404 * Heartbeat support has been removed; the ABI is changed for now.
2406 *Richard Levitte, Rich Salz*
2408 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2412 * The RSA "null" method, which was partially supported to avoid patent
2413 issues, has been replaced to always returns NULL.
2420 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2422 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2423 used even when parsing explicit parameters, when loading a encoded key
2424 or calling `EC_GROUP_new_from_ecpkparameters()`/
2425 `EC_GROUP_new_from_ecparameters()`.
2426 This prevents bypass of security hardening and performance gains,
2427 especially for curves with specialized EC_METHODs.
2428 By default, if a key encoded with explicit parameters is loaded and later
2429 encoded, the output is still encoded with explicit parameters, even if
2430 internally a "named" EC_GROUP is used for computation.
2434 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2435 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2436 NULL. After this change, only the cofactor parameter can be NULL. It also
2437 does some minimal sanity checks on the passed order.
2442 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2443 An attack is simple, if the first CMS_recipientInfo is valid but the
2444 second CMS_recipientInfo is chosen ciphertext. If the second
2445 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2446 encryption key will be replaced by garbage, and the message cannot be
2447 decoded, but if the RSA decryption fails, the correct encryption key is
2448 used and the recipient will not notice the attack.
2449 As a work around for this potential attack the length of the decrypted
2450 key must be equal to the cipher default key length, in case the
2451 certifiate is not given and all recipientInfo are tried out.
2452 The old behaviour can be re-enabled in the CMS code by setting the
2453 CMS_DEBUG_DECRYPT flag.
2458 * Use Windows installation paths in the mingw builds
2460 Mingw isn't a POSIX environment per se, which means that Windows
2461 paths should be used for installation.
2466 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2468 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2469 This changes the size when using the `genpkey` command when no size is given.
2470 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2471 generation commands to use 2048 bits by default.
2475 * Prevent over long nonces in ChaCha20-Poly1305.
2477 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2478 for every encryption operation. RFC 7539 specifies that the nonce value
2479 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2480 and front pads the nonce with 0 bytes if it is less than 12
2481 bytes. However it also incorrectly allows a nonce to be set of up to 16
2482 bytes. In this case only the last 12 bytes are significant and any
2483 additional leading bytes are ignored.
2485 It is a requirement of using this cipher that nonce values are
2486 unique. Messages encrypted using a reused nonce value are susceptible to
2487 serious confidentiality and integrity attacks. If an application changes
2488 the default nonce length to be longer than 12 bytes and then makes a
2489 change to the leading bytes of the nonce expecting the new value to be a
2490 new unique nonce then such an application could inadvertently encrypt
2491 messages with a reused nonce.
2493 Additionally the ignored bytes in a long nonce are not covered by the
2494 integrity guarantee of this cipher. Any application that relies on the
2495 integrity of these ignored leading bytes of a long nonce may be further
2496 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2497 is safe because no such use sets such a long nonce value. However user
2498 applications that use this cipher directly and set a non-default nonce
2499 length to be longer than 12 bytes may be vulnerable.
2501 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2507 * Added SCA hardening for modular field inversion in EC_GROUP through
2508 a new dedicated field_inv() pointer in EC_METHOD.
2509 This also addresses a leakage affecting conversions from projective
2510 to affine coordinates.
2512 *Billy Bob Brumley, Nicola Tuveri*
2514 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2515 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2519 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2523 * Remove the 'dist' target and add a tarball building script. The
2524 'dist' target has fallen out of use, and it shouldn't be
2525 necessary to configure just to create a source distribution.
2529 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2531 * Timing vulnerability in DSA signature generation
2533 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2534 timing side channel attack. An attacker could use variations in the signing
2535 algorithm to recover the private key.
2537 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2542 * Timing vulnerability in ECDSA signature generation
2544 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2545 timing side channel attack. An attacker could use variations in the signing
2546 algorithm to recover the private key.
2548 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2553 * Add coordinate blinding for EC_POINT and implement projective
2554 coordinate blinding for generic prime curves as a countermeasure to
2555 chosen point SCA attacks.
2557 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2559 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2561 * Client DoS due to large DH parameter
2563 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2564 malicious server can send a very large prime value to the client. This will
2565 cause the client to spend an unreasonably long period of time generating a
2566 key for this prime resulting in a hang until the client has finished. This
2567 could be exploited in a Denial Of Service attack.
2569 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2574 * Cache timing vulnerability in RSA Key Generation
2576 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2577 a cache timing side channel attack. An attacker with sufficient access to
2578 mount cache timing attacks during the RSA key generation process could
2579 recover the private key.
2581 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2582 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2587 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2588 parameter is no longer accepted, as it leads to a corrupt table. NULL
2589 pem_str is reserved for alias entries only.
2593 * Revert blinding in ECDSA sign and instead make problematic addition
2594 length-invariant. Switch even to fixed-length Montgomery multiplication.
2598 * Change generating and checking of primes so that the error rate of not
2599 being prime depends on the intended use based on the size of the input.
2600 For larger primes this will result in more rounds of Miller-Rabin.
2601 The maximal error rate for primes with more than 1080 bits is lowered
2604 *Kurt Roeckx, Annie Yousar*
2606 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2610 * Add blinding to ECDSA and DSA signatures to protect against side channel
2611 attacks discovered by Keegan Ryan (NCC Group).
2615 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2616 now allow empty (zero character) pass phrases.
2620 * Certificate time validation (X509_cmp_time) enforces stricter
2621 compliance with RFC 5280. Fractional seconds and timezone offsets
2622 are no longer allowed.
2626 * Fixed a text canonicalisation bug in CMS
2628 Where a CMS detached signature is used with text content the text goes
2629 through a canonicalisation process first prior to signing or verifying a
2630 signature. This process strips trailing space at the end of lines, converts
2631 line terminators to CRLF and removes additional trailing line terminators
2632 at the end of a file. A bug in the canonicalisation process meant that
2633 some characters, such as form-feed, were incorrectly treated as whitespace
2634 and removed. This is contrary to the specification (RFC5485). This fix
2635 could mean that detached text data signed with an earlier version of
2636 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2637 signed with a fixed OpenSSL may fail to verify with an earlier version of
2638 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2639 and use the "-binary" flag (for the "cms" command line application) or set
2640 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2644 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2646 * Constructed ASN.1 types with a recursive definition could exceed the stack
2648 Constructed ASN.1 types with a recursive definition (such as can be found
2649 in PKCS7) could eventually exceed the stack given malicious input with
2650 excessive recursion. This could result in a Denial Of Service attack. There
2651 are no such structures used within SSL/TLS that come from untrusted sources
2652 so this is considered safe.
2654 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2660 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2662 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2663 effectively reduced to only comparing the least significant bit of each
2664 byte. This allows an attacker to forge messages that would be considered as
2665 authenticated in an amount of tries lower than that guaranteed by the
2666 security claims of the scheme. The module can only be compiled by the
2667 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2669 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2675 * Add a build target 'build_all_generated', to build all generated files
2676 and only that. This can be used to prepare everything that requires
2677 things like perl for a system that lacks perl and then move everything
2678 to that system and do the rest of the build there.
2682 * Backport SSL_OP_NO_RENGOTIATION
2684 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2685 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2686 changes this is no longer possible in 1.1.0. Therefore the new
2687 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2688 1.1.0 to provide equivalent functionality.
2690 Note that if an application built against 1.1.0h headers (or above) is run
2691 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2692 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2696 * Removed the OS390-Unix config target. It relied on a script that doesn't
2701 * rsaz_1024_mul_avx2 overflow bug on x86_64
2703 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2704 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2705 Analysis suggests that attacks against RSA and DSA as a result of this
2706 defect would be very difficult to perform and are not believed likely.
2707 Attacks against DH1024 are considered just feasible, because most of the
2708 work necessary to deduce information about a private key may be performed
2709 offline. The amount of resources required for such an attack would be
2710 significant. However, for an attack on TLS to be meaningful, the server
2711 would have to share the DH1024 private key among multiple clients, which is
2712 no longer an option since CVE-2016-0701.
2714 This only affects processors that support the AVX2 but not ADX extensions
2715 like Intel Haswell (4th generation).
2717 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2718 was originally found via the OSS-Fuzz project.
2723 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2725 * bn_sqrx8x_internal carry bug on x86_64
2727 There is a carry propagating bug in the x86_64 Montgomery squaring
2728 procedure. No EC algorithms are affected. Analysis suggests that attacks
2729 against RSA and DSA as a result of this defect would be very difficult to
2730 perform and are not believed likely. Attacks against DH are considered just
2731 feasible (although very difficult) because most of the work necessary to
2732 deduce information about a private key may be performed offline. The amount
2733 of resources required for such an attack would be very significant and
2734 likely only accessible to a limited number of attackers. An attacker would
2735 additionally need online access to an unpatched system using the target
2736 private key in a scenario with persistent DH parameters and a private
2737 key that is shared between multiple clients.
2739 This only affects processors that support the BMI1, BMI2 and ADX extensions
2740 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2742 This issue was reported to OpenSSL by the OSS-Fuzz project.
2747 * Malformed X.509 IPAddressFamily could cause OOB read
2749 If an X.509 certificate has a malformed IPAddressFamily extension,
2750 OpenSSL could do a one-byte buffer overread. The most likely result
2751 would be an erroneous display of the certificate in text format.
2753 This issue was reported to OpenSSL by the OSS-Fuzz project.
2758 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2760 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2761 platform rather than 'mingw'.
2765 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2766 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2767 which is the minimum version we support.
2771 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2773 * Encrypt-Then-Mac renegotiation crash
2775 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2776 negotiated where it was not in the original handshake (or vice-versa) then
2777 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2778 and servers are affected.
2780 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2785 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2787 * Truncated packet could crash via OOB read
2789 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2790 cipher is being used, then a truncated packet can cause that host to
2791 perform an out-of-bounds read, usually resulting in a crash.
2793 This issue was reported to OpenSSL by Robert Święcki of Google.
2798 * Bad (EC)DHE parameters cause a client crash
2800 If a malicious server supplies bad parameters for a DHE or ECDHE key
2801 exchange then this can result in the client attempting to dereference a
2802 NULL pointer leading to a client crash. This could be exploited in a Denial
2805 This issue was reported to OpenSSL by Guido Vranken.
2810 * BN_mod_exp may produce incorrect results on x86_64
2812 There is a carry propagating bug in the x86_64 Montgomery squaring
2813 procedure. No EC algorithms are affected. Analysis suggests that attacks
2814 against RSA and DSA as a result of this defect would be very difficult to
2815 perform and are not believed likely. Attacks against DH are considered just
2816 feasible (although very difficult) because most of the work necessary to
2817 deduce information about a private key may be performed offline. The amount
2818 of resources required for such an attack would be very significant and
2819 likely only accessible to a limited number of attackers. An attacker would
2820 additionally need online access to an unpatched system using the target
2821 private key in a scenario with persistent DH parameters and a private
2822 key that is shared between multiple clients. For example this can occur by
2823 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2824 similar to CVE-2015-3193 but must be treated as a separate problem.
2826 This issue was reported to OpenSSL by the OSS-Fuzz project.
2831 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2833 * ChaCha20/Poly1305 heap-buffer-overflow
2835 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2836 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2837 crash. This issue is not considered to be exploitable beyond a DoS.
2839 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2844 * CMS Null dereference
2846 Applications parsing invalid CMS structures can crash with a NULL pointer
2847 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2848 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2849 structure callback if an attempt is made to free certain invalid encodings.
2850 Only CHOICE structures using a callback which do not handle NULL value are
2853 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2858 * Montgomery multiplication may produce incorrect results
2860 There is a carry propagating bug in the Broadwell-specific Montgomery
2861 multiplication procedure that handles input lengths divisible by, but
2862 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2863 and DH private keys are impossible. This is because the subroutine in
2864 question is not used in operations with the private key itself and an input
2865 of the attacker's direct choice. Otherwise the bug can manifest itself as
2866 transient authentication and key negotiation failures or reproducible
2867 erroneous outcome of public-key operations with specially crafted input.
2868 Among EC algorithms only Brainpool P-512 curves are affected and one
2869 presumably can attack ECDH key negotiation. Impact was not analyzed in
2870 detail, because pre-requisites for attack are considered unlikely. Namely
2871 multiple clients have to choose the curve in question and the server has to
2872 share the private key among them, neither of which is default behaviour.
2873 Even then only clients that chose the curve will be affected.
2875 This issue was publicly reported as transient failures and was not
2876 initially recognized as a security issue. Thanks to Richard Morgan for
2877 providing reproducible case.
2882 * Removed automatic addition of RPATH in shared libraries and executables,
2883 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2887 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2889 * Fix Use After Free for large message sizes
2891 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2892 message larger than approx 16k is received then the underlying buffer to
2893 store the incoming message is reallocated and moved. Unfortunately a
2894 dangling pointer to the old location is left which results in an attempt to
2895 write to the previously freed location. This is likely to result in a
2896 crash, however it could potentially lead to execution of arbitrary code.
2898 This issue only affects OpenSSL 1.1.0a.
2900 This issue was reported to OpenSSL by Robert Święcki.
2905 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2907 * OCSP Status Request extension unbounded memory growth
2909 A malicious client can send an excessively large OCSP Status Request
2910 extension. If that client continually requests renegotiation, sending a
2911 large OCSP Status Request extension each time, then there will be unbounded
2912 memory growth on the server. This will eventually lead to a Denial Of
2913 Service attack through memory exhaustion. Servers with a default
2914 configuration are vulnerable even if they do not support OCSP. Builds using
2915 the "no-ocsp" build time option are not affected.
2917 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2922 * SSL_peek() hang on empty record
2924 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2925 sends an empty record. This could be exploited by a malicious peer in a
2926 Denial Of Service attack.
2928 This issue was reported to OpenSSL by Alex Gaynor.
2933 * Excessive allocation of memory in tls_get_message_header() and
2934 dtls1_preprocess_fragment()
2936 A (D)TLS message includes 3 bytes for its length in the header for the
2937 message. This would allow for messages up to 16Mb in length. Messages of
2938 this length are excessive and OpenSSL includes a check to ensure that a
2939 peer is sending reasonably sized messages in order to avoid too much memory
2940 being consumed to service a connection. A flaw in the logic of version
2941 1.1.0 means that memory for the message is allocated too early, prior to
2942 the excessive message length check. Due to way memory is allocated in
2943 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2944 to service a connection. This could lead to a Denial of Service through
2945 memory exhaustion. However, the excessive message length check still takes
2946 place, and this would cause the connection to immediately fail. Assuming
2947 that the application calls SSL_free() on the failed connection in a timely
2948 manner then the 21Mb of allocated memory will then be immediately freed
2949 again. Therefore the excessive memory allocation will be transitory in
2950 nature. This then means that there is only a security impact if:
2952 1) The application does not call SSL_free() in a timely manner in the event
2953 that the connection fails
2955 2) The application is working in a constrained environment where there is
2956 very little free memory
2958 3) The attacker initiates multiple connection attempts such that there are
2959 multiple connections in a state where memory has been allocated for the
2960 connection; SSL_free() has not yet been called; and there is insufficient
2961 memory to service the multiple requests.
2963 Except in the instance of (1) above any Denial Of Service is likely to be
2964 transitory because as soon as the connection fails the memory is
2965 subsequently freed again in the SSL_free() call. However there is an
2966 increased risk during this period of application crashes due to the lack of
2967 memory - which would then mean a more serious Denial of Service.
2969 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2970 (CVE-2016-6307 and CVE-2016-6308)
2974 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2975 had to be removed. Primary reason is that vendor assembler can't
2976 assemble our modules with -KPIC flag. As result it, assembly
2977 support, was not even available as option. But its lack means
2978 lack of side-channel resistant code, which is incompatible with
2979 security by todays standards. Fortunately gcc is readily available
2980 prepackaged option, which we firmly point at...
2984 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2986 * Windows command-line tool supports UTF-8 opt-in option for arguments
2987 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2988 (to any value) allows Windows user to access PKCS#12 file generated
2989 with Windows CryptoAPI and protected with non-ASCII password, as well
2990 as files generated under UTF-8 locale on Linux also protected with
2995 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2996 have been disabled by default and removed from DEFAULT, just like RC4.
2997 See the RC4 item below to re-enable both.
3001 * The method for finding the storage location for the Windows RAND seed file
3002 has changed. First we check %RANDFILE%. If that is not set then we check
3003 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3004 all else fails we fall back to C:\.
3008 * The EVP_EncryptUpdate() function has had its return type changed from void
3009 to int. A return of 0 indicates and error while a return of 1 indicates
3014 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3015 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3016 off the constant time implementation for RSA, DSA and DH have been made
3017 no-ops and deprecated.
3021 * Windows RAND implementation was simplified to only get entropy by
3022 calling CryptGenRandom(). Various other RAND-related tickets
3025 *Joseph Wylie Yandle, Rich Salz*
3027 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3028 and `OPENSSL_LH_`, respectively. The old names are available
3029 with API compatibility. They new names are now completely documented.
3033 * Unify TYPE_up_ref(obj) methods signature.
3034 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3035 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3036 int (instead of void) like all others TYPE_up_ref() methods.
3037 So now these methods also check the return value of CRYPTO_atomic_add(),
3038 and the validity of object reference counter.
3040 *fdasilvayy@gmail.com*
3042 * With Windows Visual Studio builds, the .pdb files are installed
3043 alongside the installed libraries and executables. For a static
3044 library installation, ossl_static.pdb is the associate compiler
3045 generated .pdb file to be used when linking programs.
3049 * Remove openssl.spec. Packaging files belong with the packagers.
3053 * Automatic Darwin/OSX configuration has had a refresh, it will now
3054 recognise x86_64 architectures automatically. You can still decide
3055 to build for a different bitness with the environment variable
3056 KERNEL_BITS (can be 32 or 64), for example:
3058 KERNEL_BITS=32 ./config
3062 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3063 256 bit AES and HMAC with SHA256.
3067 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3071 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3075 * To enable users to have their own config files and build file templates,
3076 Configure looks in the directory indicated by the environment variable
3077 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3078 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3079 name and is used as is.
3083 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3084 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3085 X509_CERT_FILE_CTX was removed.
3089 * "shared" builds are now the default. To create only static libraries use
3090 the "no-shared" Configure option.
3094 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3095 All of these option have not worked for some while and are fundamental
3100 * Make various cleanup routines no-ops and mark them as deprecated. Most
3101 global cleanup functions are no longer required because they are handled
3102 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3103 Explicitly de-initing can cause problems (e.g. where a library that uses
3104 OpenSSL de-inits, but an application is still using it). The affected
3105 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3106 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3107 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3108 COMP_zlib_cleanup().
3112 * --strict-warnings no longer enables runtime debugging options
3113 such as REF_DEBUG. Instead, debug options are automatically
3114 enabled with '--debug' builds.
3116 *Andy Polyakov, Emilia Käsper*
3118 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3119 have been moved out of the public header files. New functions for managing
3120 these have been added.
3124 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3125 objects have been moved out of the public header files. New
3126 functions for managing these have been added.
3130 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3131 have been moved out of the public header files. New functions for managing
3132 these have been added.
3136 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3137 moved out of the public header files. New functions for managing these
3142 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3146 * Removed the mk1mf build scripts.
3150 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3151 it is always safe to #include a header now.
3155 * Removed the aged BC-32 config and all its supporting scripts
3159 * Removed support for Ultrix, Netware, and OS/2.
3163 * Add support for HKDF.
3165 *Alessandro Ghedini*
3167 * Add support for blake2b and blake2s
3171 * Added support for "pipelining". Ciphers that have the
3172 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3173 encryptions/decryptions simultaneously. There are currently no built-in
3174 ciphers with this property but the expectation is that engines will be able
3175 to offer it to significantly improve throughput. Support has been extended
3176 into libssl so that multiple records for a single connection can be
3177 processed in one go (for >=TLS 1.1).
3181 * Added the AFALG engine. This is an async capable engine which is able to
3182 offload work to the Linux kernel. In this initial version it only supports
3183 AES128-CBC. The kernel must be version 4.1.0 or greater.
3187 * OpenSSL now uses a new threading API. It is no longer necessary to
3188 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3189 are two supported threading models: pthreads and windows threads. It is
3190 also possible to configure OpenSSL at compile time for "no-threads". The
3191 old threading API should no longer be used. The functions have been
3192 replaced with "no-op" compatibility macros.
3194 *Alessandro Ghedini, Matt Caswell*
3196 * Modify behavior of ALPN to invoke callback after SNI/servername
3197 callback, such that updates to the SSL_CTX affect ALPN.
3201 * Add SSL_CIPHER queries for authentication and key-exchange.
3205 * Changes to the DEFAULT cipherlist:
3206 - Prefer (EC)DHE handshakes over plain RSA.
3207 - Prefer AEAD ciphers over legacy ciphers.
3208 - Prefer ECDSA over RSA when both certificates are available.
3209 - Prefer TLSv1.2 ciphers/PRF.
3210 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3215 * Change the ECC default curve list to be this, in order: x25519,
3216 secp256r1, secp521r1, secp384r1.
3220 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3221 disabled by default. They can be re-enabled using the
3222 enable-weak-ssl-ciphers option to Configure.
3226 * If the server has ALPN configured, but supports no protocols that the
3227 client advertises, send a fatal "no_application_protocol" alert.
3228 This behaviour is SHALL in RFC 7301, though it isn't universally
3229 implemented by other servers.
3233 * Add X25519 support.
3234 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3235 for public and private key encoding using the format documented in
3236 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3237 key generation and key derivation.
3239 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3244 * Deprecate SRP_VBASE_get_by_user.
3245 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3246 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3247 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3248 seed, even if the seed is configured.
3250 Users should use SRP_VBASE_get1_by_user instead. Note that in
3251 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3252 also that even though configuring the SRP seed attempts to hide
3253 invalid usernames by continuing the handshake with fake
3254 credentials, this behaviour is not constant time and no strong
3255 guarantees are made that the handshake is indistinguishable from
3256 that of a valid user.
3260 * Configuration change; it's now possible to build dynamic engines
3261 without having to build shared libraries and vice versa. This
3262 only applies to the engines in `engines/`, those in `crypto/engine/`
3263 will always be built into libcrypto (i.e. "static").
3265 Building dynamic engines is enabled by default; to disable, use
3266 the configuration option "disable-dynamic-engine".
3268 The only requirements for building dynamic engines are the
3269 presence of the DSO module and building with position independent
3270 code, so they will also automatically be disabled if configuring
3271 with "disable-dso" or "disable-pic".
3273 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3274 are also taken away from openssl/opensslconf.h, as they are
3279 * Configuration change; if there is a known flag to compile
3280 position independent code, it will always be applied on the
3281 libcrypto and libssl object files, and never on the application
3282 object files. This means other libraries that use routines from
3283 libcrypto / libssl can be made into shared libraries regardless
3284 of how OpenSSL was configured.
3286 If this isn't desirable, the configuration options "disable-pic"
3287 or "no-pic" can be used to disable the use of PIC. This will
3288 also disable building shared libraries and dynamic engines.
3292 * Removed JPAKE code. It was experimental and has no wide use.
3296 * The INSTALL_PREFIX Makefile variable has been renamed to
3297 DESTDIR. That makes for less confusion on what this variable
3298 is for. Also, the configuration option --install_prefix is
3303 * Heartbeat for TLS has been removed and is disabled by default
3304 for DTLS; configure with enable-heartbeats. Code that uses the
3305 old #define's might need to be updated.
3307 *Emilia Käsper, Rich Salz*
3309 * Rename REF_CHECK to REF_DEBUG.
3313 * New "unified" build system
3315 The "unified" build system is aimed to be a common system for all
3316 platforms we support. With it comes new support for VMS.
3318 This system builds supports building in a different directory tree
3319 than the source tree. It produces one Makefile (for unix family
3320 or lookalikes), or one descrip.mms (for VMS).
3322 The source of information to make the Makefile / descrip.mms is
3323 small files called 'build.info', holding the necessary
3324 information for each directory with source to compile, and a
3325 template in Configurations, like unix-Makefile.tmpl or
3328 With this change, the library names were also renamed on Windows
3329 and on VMS. They now have names that are closer to the standard
3330 on Unix, and include the major version number, and in certain
3331 cases, the architecture they are built for. See "Notes on shared
3332 libraries" in INSTALL.
3334 We rely heavily on the perl module Text::Template.
3338 * Added support for auto-initialisation and de-initialisation of the library.
3339 OpenSSL no longer requires explicit init or deinit routines to be called,
3340 except in certain circumstances. See the OPENSSL_init_crypto() and
3341 OPENSSL_init_ssl() man pages for further information.
3345 * The arguments to the DTLSv1_listen function have changed. Specifically the
3346 "peer" argument is now expected to be a BIO_ADDR object.
3348 * Rewrite of BIO networking library. The BIO library lacked consistent
3349 support of IPv6, and adding it required some more extensive
3350 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3351 which hold all types of addresses and chains of address information.
3352 It also introduces a new API, with functions like BIO_socket,
3353 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3354 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3355 have been adapted accordingly.
3359 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3364 * CRIME protection: disable compression by default, even if OpenSSL is
3365 compiled with zlib enabled. Applications can still enable compression
3366 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3367 using the SSL_CONF library to configure compression.
3371 * The signature of the session callback configured with
3372 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3373 was explicitly marked as `const unsigned char*` instead of
3378 * Always DPURIFY. Remove the use of uninitialized memory in the
3379 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3383 * Removed many obsolete configuration items, including
3384 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3385 MD2_CHAR, MD2_INT, MD2_LONG
3387 IDEA_SHORT, IDEA_LONG
3388 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3390 *Rich Salz, with advice from Andy Polyakov*
3392 * Many BN internals have been moved to an internal header file.
3394 *Rich Salz with help from Andy Polyakov*
3396 * Configuration and writing out the results from it has changed.
3397 Files such as Makefile include/openssl/opensslconf.h and are now
3398 produced through general templates, such as Makefile.in and
3399 crypto/opensslconf.h.in and some help from the perl module
3402 Also, the center of configuration information is no longer
3403 Makefile. Instead, Configure produces a perl module in
3404 configdata.pm which holds most of the config data (in the hash
3405 table %config), the target data that comes from the target
3406 configuration in one of the `Configurations/*.conf` files (in
3411 * To clarify their intended purposes, the Configure options
3412 --prefix and --openssldir change their semantics, and become more
3413 straightforward and less interdependent.
3415 --prefix shall be used exclusively to give the location INSTALLTOP
3416 where programs, scripts, libraries, include files and manuals are
3417 going to be installed. The default is now /usr/local.
3419 --openssldir shall be used exclusively to give the default
3420 location OPENSSLDIR where certificates, private keys, CRLs are
3421 managed. This is also where the default openssl.cnf gets
3423 If the directory given with this option is a relative path, the
3424 values of both the --prefix value and the --openssldir value will
3425 be combined to become OPENSSLDIR.
3426 The default for --openssldir is INSTALLTOP/ssl.
3428 Anyone who uses --openssldir to specify where OpenSSL is to be
3429 installed MUST change to use --prefix instead.
3433 * The GOST engine was out of date and therefore it has been removed. An up
3434 to date GOST engine is now being maintained in an external repository.
3435 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3436 support for GOST ciphersuites (these are only activated if a GOST engine
3441 * EGD is no longer supported by default; use enable-egd when
3444 *Ben Kaduk and Rich Salz*
3446 * The distribution now has Makefile.in files, which are used to
3447 create Makefile's when Configure is run. *Configure must be run
3448 before trying to build now.*
3452 * The return value for SSL_CIPHER_description() for error conditions
3457 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3459 Obtaining and performing DNSSEC validation of TLSA records is
3460 the application's responsibility. The application provides
3461 the TLSA records of its choice to OpenSSL, and these are then
3462 used to authenticate the peer.
3464 The TLSA records need not even come from DNS. They can, for
3465 example, be used to implement local end-entity certificate or
3466 trust-anchor "pinning", where the "pin" data takes the form
3467 of TLSA records, which can augment or replace verification
3468 based on the usual WebPKI public certification authorities.
3472 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3473 continues to support deprecated interfaces in default builds.
3474 However, applications are strongly advised to compile their
3475 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3476 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3477 or the 1.1.0 releases.
3479 In environments in which all applications have been ported to
3480 not use any deprecated interfaces OpenSSL's Configure script
3481 should be used with the --api=1.1.0 option to entirely remove
3482 support for the deprecated features from the library and
3483 unconditionally disable them in the installed headers.
3484 Essentially the same effect can be achieved with the "no-deprecated"
3485 argument to Configure, except that this will always restrict
3486 the build to just the latest API, rather than a fixed API
3489 As applications are ported to future revisions of the API,
3490 they should update their compile-time OPENSSL_API_COMPAT define
3491 accordingly, but in most cases should be able to continue to
3492 compile with later releases.
3494 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3495 0x10000000L and 0x00908000L, respectively. However those
3496 versions did not support the OPENSSL_API_COMPAT feature, and
3497 so applications are not typically tested for explicit support
3498 of just the undeprecated features of either release.
3502 * Add support for setting the minimum and maximum supported protocol.
3503 It can bet set via the SSL_set_min_proto_version() and
3504 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3505 MaxProtocol. It's recommended to use the new APIs to disable
3506 protocols instead of disabling individual protocols using
3507 SSL_set_options() or SSL_CONF's Protocol. This change also
3508 removes support for disabling TLS 1.2 in the OpenSSL TLS
3509 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3513 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3517 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3518 and integrates ECDSA and ECDH functionality into EC. Implementations can
3519 now redirect key generation and no longer need to convert to or from
3522 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3523 include the ec.h header file instead.
3527 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3528 ciphers who are no longer supported and drops support the ephemeral RSA key
3529 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3533 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3534 opaque. For HMAC_CTX, the following constructors and destructors
3537 HMAC_CTX *HMAC_CTX_new(void);
3538 void HMAC_CTX_free(HMAC_CTX *ctx);
3540 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3541 destroy such methods has been added. See EVP_MD_meth_new(3) and
3542 EVP_CIPHER_meth_new(3) for documentation.
3545 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3546 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3547 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3548 an already created structure.
3549 2) For consistency with the majority of our object creators and
3550 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3551 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3552 for deprecated builds.
3556 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3557 cryptographic operations to be performed asynchronously as long as an
3558 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3559 further details. Libssl has also had this capability integrated with the
3560 introduction of the new mode SSL_MODE_ASYNC and associated error
3561 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3562 pages. This work was developed in partnership with Intel Corp.
3566 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3567 always enabled now. If you want to disable the support you should
3568 exclude it using the list of supported ciphers. This also means that the
3569 "-no_ecdhe" option has been removed from s_server.
3573 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3574 SSL_{CTX_}set1_curves() which can set a list.
3578 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3579 curve you want to support using SSL_{CTX_}set1_curves().
3583 * State machine rewrite. The state machine code has been significantly
3584 refactored in order to remove much duplication of code and solve issues
3585 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3586 further details). This change does have some associated API changes.
3587 Notably the SSL_state() function has been removed and replaced by
3588 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3589 SSL_set_state() has been removed altogether. The previous handshake states
3590 defined in ssl.h and ssl3.h have also been removed.
3594 * All instances of the string "ssleay" in the public API were replaced
3595 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3596 Some error codes related to internal RSA_eay API's were renamed.
3600 * The demo files in crypto/threads were moved to demo/threads.
3604 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3607 *Matt Caswell, Rich Salz*
3609 * New ASN.1 embed macro.
3611 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3612 structure is not allocated: it is part of the parent. That is instead of
3620 This reduces memory fragmentation and make it impossible to accidentally
3621 set a mandatory field to NULL.
3623 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3624 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3625 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3630 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3634 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3635 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3636 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3637 DES and RC4 ciphersuites.
3641 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3642 This changes the decoding behaviour for some invalid messages,
3643 though the change is mostly in the more lenient direction, and
3644 legacy behaviour is preserved as much as possible.
3648 * Fix no-stdio build.
3649 *David Woodhouse <David.Woodhouse@intel.com> and also*
3650 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3652 * New testing framework
3653 The testing framework has been largely rewritten and is now using
3654 perl and the perl modules Test::Harness and an extended variant of
3655 Test::More called OpenSSL::Test to do its work. All test scripts in
3656 test/ have been rewritten into test recipes, and all direct calls to
3657 executables in test/Makefile have become individual recipes using the
3658 simplified testing OpenSSL::Test::Simple.
3660 For documentation on our testing modules, do:
3662 perldoc test/testlib/OpenSSL/Test/Simple.pm
3663 perldoc test/testlib/OpenSSL/Test.pm
3667 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3668 are used; the latter aborts on memory leaks (usually checked on exit).
3669 Some undocumented "set malloc, etc., hooks" functions were removed
3670 and others were changed. All are now documented.
3674 * In DSA_generate_parameters_ex, if the provided seed is too short,
3677 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3679 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3680 from RFC4279, RFC4785, RFC5487, RFC5489.
3682 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3683 original RSA_PSK patch.
3687 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3688 era flag was never set throughout the codebase (only read). Also removed
3689 SSL3_FLAGS_POP_BUFFER which was only used if
3690 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3694 * Changed the default name options in the "ca", "crl", "req" and "x509"
3695 to be "oneline" instead of "compat".
3699 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3700 not aware of clients that still exhibit this bug, and the workaround
3701 hasn't been working properly for a while.
3705 * The return type of BIO_number_read() and BIO_number_written() as well as
3706 the corresponding num_read and num_write members in the BIO structure has
3707 changed from unsigned long to uint64_t. On platforms where an unsigned
3708 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3713 * Given the pervasive nature of TLS extensions it is inadvisable to run
3714 OpenSSL without support for them. It also means that maintaining
3715 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3716 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3720 * Removed support for the two export grade static DH ciphersuites
3721 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3722 were newly added (along with a number of other static DH ciphersuites) to
3723 1.0.2. However the two export ones have *never* worked since they were
3724 introduced. It seems strange in any case to be adding new export
3725 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3729 * Version negotiation has been rewritten. In particular SSLv23_method(),
3730 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3731 and turned into macros which simply call the new preferred function names
3732 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3733 should use the new names instead. Also as part of this change the ssl23.h
3734 header file has been removed.
3738 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3739 code and the associated standard is no longer considered fit-for-purpose.
3743 * RT2547 was closed. When generating a private key, try to make the
3744 output file readable only by the owner. This behavior change might
3745 be noticeable when interacting with other software.
3747 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3752 * Added HTTP GET support to the ocsp command.
3756 * Changed default digest for the dgst and enc commands from MD5 to
3761 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3765 * Added support for TLS extended master secret from
3766 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3767 initial patch which was a great help during development.
3771 * All libssl internal structures have been removed from the public header
3772 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3773 now redundant). Users should not attempt to access internal structures
3774 directly. Instead they should use the provided API functions.
3778 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3779 Access to deprecated functions can be re-enabled by running config with
3780 "enable-deprecated". In addition applications wishing to use deprecated
3781 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3782 will, by default, disable some transitive includes that previously existed
3783 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3787 * Added support for OCB mode. OpenSSL has been granted a patent license
3788 compatible with the OpenSSL license for use of OCB. Details are available
3789 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3790 for OCB can be removed by calling config with no-ocb.
3794 * SSLv2 support has been removed. It still supports receiving a SSLv2
3795 compatible client hello.
3799 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3800 done while fixing the error code for the key-too-small case.
3802 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3804 * CA.sh has been removed; use CA.pl instead.
3808 * Removed old DES API.
3812 * Remove various unsupported platforms:
3818 Sinix/ReliantUNIX RM400
3823 16-bit platforms such as WIN16
3827 * Clean up OPENSSL_NO_xxx #define's
3828 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3829 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3830 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3831 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3832 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3833 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3834 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3835 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3836 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3837 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3841 * Cleaned up dead code
3842 Remove all but one '#ifdef undef' which is to be looked at.
3846 * Clean up calling of xxx_free routines.
3847 Just like free(), fix most of the xxx_free routines to accept
3848 NULL. Remove the non-null checks from callers. Save much code.
3852 * Add secure heap for storage of private keys (when possible).
3853 Add BIO_s_secmem(), CBIGNUM, etc.
3854 Contributed by Akamai Technologies under our Corporate CLA.
3858 * Experimental support for a new, fast, unbiased prime candidate generator,
3859 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3861 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3863 * New output format NSS in the sess_id command line tool. This allows
3864 exporting the session id and the master key in NSS keylog format.
3866 *Martin Kaiser <martin@kaiser.cx>*
3868 * Harmonize version and its documentation. -f flag is used to display
3871 *mancha <mancha1@zoho.com>*
3873 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3874 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3876 *mancha <mancha1@zoho.com>*
3878 * Fix some double frees. These are not thought to be exploitable.
3880 *mancha <mancha1@zoho.com>*
3882 * A missing bounds check in the handling of the TLS heartbeat extension
3883 can be used to reveal up to 64k of memory to a connected client or
3886 Thanks for Neel Mehta of Google Security for discovering this bug and to
3887 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3888 preparing the fix ([CVE-2014-0160])
3890 *Adam Langley, Bodo Moeller*
3892 * Fix for the attack described in the paper "Recovering OpenSSL
3893 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3894 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3895 <http://eprint.iacr.org/2014/140>
3897 Thanks to Yuval Yarom and Naomi Benger for discovering this
3898 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3900 *Yuval Yarom and Naomi Benger*
3902 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3903 this fixes a limitation in previous versions of OpenSSL.
3907 * Experimental encrypt-then-mac support.
3909 Experimental support for encrypt then mac from
3910 draft-gutmann-tls-encrypt-then-mac-02.txt
3912 To enable it set the appropriate extension number (0x42 for the test
3913 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3915 For non-compliant peers (i.e. just about everything) this should have no
3918 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3922 * Add EVP support for key wrapping algorithms, to avoid problems with
3923 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3924 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3925 algorithms and include tests cases.
3929 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3934 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3935 MGF1 digest and OAEP label.
3939 * Make openssl verify return errors.
3941 *Chris Palmer <palmer@google.com> and Ben Laurie*
3943 * New function ASN1_TIME_diff to calculate the difference between two
3944 ASN1_TIME structures or one structure and the current time.
3948 * Update fips_test_suite to support multiple command line options. New
3949 test to induce all self test errors in sequence and check expected
3954 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3955 sign or verify all in one operation.
3959 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3960 test programs and fips_test_suite. Includes functionality to parse
3961 the minimal script output of fipsalgest.pl directly.
3965 * Add authorisation parameter to FIPS_module_mode_set().
3969 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3973 * Use separate DRBG fields for internal and external flags. New function
3974 FIPS_drbg_health_check() to perform on demand health checking. Add
3975 generation tests to fips_test_suite with reduced health check interval to
3976 demonstrate periodic health checking. Add "nodh" option to
3977 fips_test_suite to skip very slow DH test.
3981 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3986 * More extensive health check for DRBG checking many more failure modes.
3987 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3988 combination: call this in fips_test_suite.
3992 * Add support for canonical generation of DSA parameter 'g'. See
3995 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3996 POST to handle HMAC cases.
4000 * Add functions FIPS_module_version() and FIPS_module_version_text()
4001 to return numerical and string versions of the FIPS module number.
4005 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4006 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4007 outside the validated module in the FIPS capable OpenSSL.
4011 * Minor change to DRBG entropy callback semantics. In some cases
4012 there is no multiple of the block length between min_len and
4013 max_len. Allow the callback to return more than max_len bytes
4014 of entropy but discard any extra: it is the callback's responsibility
4015 to ensure that the extra data discarded does not impact the
4016 requested amount of entropy.
4020 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4021 information in FIPS186-3, SP800-57 and SP800-131A.
4025 * CCM support via EVP. Interface is very similar to GCM case except we
4026 must supply all data in one chunk (i.e. no update, final) and the
4027 message length must be supplied if AAD is used. Add algorithm test
4032 * Initial version of POST overhaul. Add POST callback to allow the status
4033 of POST to be monitored and/or failures induced. Modify fips_test_suite
4034 to use callback. Always run all selftests even if one fails.
4038 * XTS support including algorithm test driver in the fips_gcmtest program.
4039 Note: this does increase the maximum key length from 32 to 64 bytes but
4040 there should be no binary compatibility issues as existing applications
4041 will never use XTS mode.
4045 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4046 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4047 performs algorithm blocking for unapproved PRNG types. Also do not
4048 set PRNG type in FIPS_mode_set(): leave this to the application.
4049 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4050 the standard OpenSSL PRNG: set additional data to a date time vector.
4054 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4055 This shouldn't present any incompatibility problems because applications
4056 shouldn't be using these directly and any that are will need to rethink
4057 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4061 * Extensive self tests and health checking required by SP800-90 DRBG.
4062 Remove strength parameter from FIPS_drbg_instantiate and always
4063 instantiate at maximum supported strength.
4067 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4071 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4075 * New function DH_compute_key_padded() to compute a DH key and pad with
4076 leading zeroes if needed: this complies with SP800-56A et al.
4080 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4081 anything, incomplete, subject to change and largely untested at present.
4085 * Modify fipscanisteronly build option to only build the necessary object
4086 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4090 * Add experimental option FIPSSYMS to give all symbols in
4091 fipscanister.o and FIPS or fips prefix. This will avoid
4092 conflicts with future versions of OpenSSL. Add perl script
4093 util/fipsas.pl to preprocess assembly language source files
4094 and rename any affected symbols.
4098 * Add selftest checks and algorithm block of non-fips algorithms in
4099 FIPS mode. Remove DES2 from selftests.
4103 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4104 return internal method without any ENGINE dependencies. Add new
4105 tiny fips sign and verify functions.
4109 * New build option no-ec2m to disable characteristic 2 code.
4113 * New build option "fipscanisteronly". This only builds fipscanister.o
4114 and (currently) associated fips utilities. Uses the file Makefile.fips
4115 instead of Makefile.org as the prototype.
4119 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4120 Update fips_gcmtest to use IV generator.
4124 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4125 setting output buffer to NULL. The `*Final` function must be
4126 called although it will not retrieve any additional data. The tag
4127 can be set or retrieved with a ctrl. The IV length is by default 12
4128 bytes (96 bits) but can be set to an alternative value. If the IV
4129 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4134 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4135 underlying do_cipher function handles all cipher semantics itself
4136 including padding and finalisation. This is useful if (for example)
4137 an ENGINE cipher handles block padding itself. The behaviour of
4138 do_cipher is subtly changed if this flag is set: the return value
4139 is the number of characters written to the output buffer (zero is
4140 no longer an error code) or a negative error code. Also if the
4141 input buffer is NULL and length 0 finalisation should be performed.
4145 * If a candidate issuer certificate is already part of the constructed
4146 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4150 * Improve forward-security support: add functions
4152 void SSL_CTX_set_not_resumable_session_callback(
4153 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4154 void SSL_set_not_resumable_session_callback(
4155 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4157 for use by SSL/TLS servers; the callback function will be called whenever a
4158 new session is created, and gets to decide whether the session may be
4159 cached to make it resumable (return 0) or not (return 1). (As by the
4160 SSL/TLS protocol specifications, the session_id sent by the server will be
4161 empty to indicate that the session is not resumable; also, the server will
4162 not generate RFC 4507 (RFC 5077) session tickets.)
4164 A simple reasonable callback implementation is to return is_forward_secure.
4165 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4166 by the SSL/TLS server library, indicating whether it can provide forward
4169 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4171 * New -verify_name option in command line utilities to set verification
4176 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4177 Add CMAC pkey methods.
4181 * Experimental renegotiation in s_server -www mode. If the client
4182 browses /reneg connection is renegotiated. If /renegcert it is
4183 renegotiated requesting a certificate.
4187 * Add an "external" session cache for debugging purposes to s_server. This
4188 should help trace issues which normally are only apparent in deployed
4189 multi-process servers.
4193 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4194 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4195 BIO_set_cipher() and some obscure PEM functions were changed so they
4196 can now return an error. The RAND changes required a change to the
4197 RAND_METHOD structure.
4201 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4202 a gcc attribute to warn if the result of a function is ignored. This
4203 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4204 whose return value is often ignored.
4208 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4209 These allow SCTs (signed certificate timestamps) to be requested and
4210 validated when establishing a connection.
4212 *Rob Percival <robpercival@google.com>*
4217 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4219 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4220 used even when parsing explicit parameters, when loading a encoded key
4221 or calling `EC_GROUP_new_from_ecpkparameters()`/
4222 `EC_GROUP_new_from_ecparameters()`.
4223 This prevents bypass of security hardening and performance gains,
4224 especially for curves with specialized EC_METHODs.
4225 By default, if a key encoded with explicit parameters is loaded and later
4226 encoded, the output is still encoded with explicit parameters, even if
4227 internally a "named" EC_GROUP is used for computation.
4231 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4232 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4233 NULL. After this change, only the cofactor parameter can be NULL. It also
4234 does some minimal sanity checks on the passed order.
4239 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4240 An attack is simple, if the first CMS_recipientInfo is valid but the
4241 second CMS_recipientInfo is chosen ciphertext. If the second
4242 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4243 encryption key will be replaced by garbage, and the message cannot be
4244 decoded, but if the RSA decryption fails, the correct encryption key is
4245 used and the recipient will not notice the attack.
4246 As a work around for this potential attack the length of the decrypted
4247 key must be equal to the cipher default key length, in case the
4248 certifiate is not given and all recipientInfo are tried out.
4249 The old behaviour can be re-enabled in the CMS code by setting the
4250 CMS_DEBUG_DECRYPT flag.
4255 * Document issue with installation paths in diverse Windows builds
4257 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4258 binaries and run-time config file.
4263 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4265 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4266 This changes the size when using the `genpkey` command when no size is given.
4267 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4268 generation commands to use 2048 bits by default.
4272 * Add FIPS support for Android Arm 64-bit
4274 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4275 Module in Version 2.0.10. For some reason, the corresponding target
4276 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4277 built with FIPS support on Android Arm 64-bit. This omission has been
4280 *Matthias St. Pierre*
4282 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4284 * 0-byte record padding oracle
4286 If an application encounters a fatal protocol error and then calls
4287 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4288 then OpenSSL can respond differently to the calling application if a 0 byte
4289 record is received with invalid padding compared to if a 0 byte record is
4290 received with an invalid MAC. If the application then behaves differently
4291 based on that in a way that is detectable to the remote peer, then this
4292 amounts to a padding oracle that could be used to decrypt data.
4294 In order for this to be exploitable "non-stitched" ciphersuites must be in
4295 use. Stitched ciphersuites are optimised implementations of certain
4296 commonly used ciphersuites. Also the application must call SSL_shutdown()
4297 twice even if a protocol error has occurred (applications should not do
4298 this but some do anyway).
4300 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4301 Aviram, with additional investigation by Steven Collison and Andrew
4302 Hourselt. It was reported to OpenSSL on 10th December 2018.
4307 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4311 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4313 * Microarchitecture timing vulnerability in ECC scalar multiplication
4315 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4316 shown to be vulnerable to a microarchitecture timing side channel attack.
4317 An attacker with sufficient access to mount local timing attacks during
4318 ECDSA signature generation could recover the private key.
4320 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4321 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4327 * Timing vulnerability in DSA signature generation
4329 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4330 timing side channel attack. An attacker could use variations in the signing
4331 algorithm to recover the private key.
4333 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4338 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4339 Module, accidentally introduced while backporting security fixes from the
4340 development branch and hindering the use of ECC in FIPS mode.
4344 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4346 * Client DoS due to large DH parameter
4348 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4349 malicious server can send a very large prime value to the client. This will
4350 cause the client to spend an unreasonably long period of time generating a
4351 key for this prime resulting in a hang until the client has finished. This
4352 could be exploited in a Denial Of Service attack.
4354 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4359 * Cache timing vulnerability in RSA Key Generation
4361 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4362 a cache timing side channel attack. An attacker with sufficient access to
4363 mount cache timing attacks during the RSA key generation process could
4364 recover the private key.
4366 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4367 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4372 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4373 parameter is no longer accepted, as it leads to a corrupt table. NULL
4374 pem_str is reserved for alias entries only.
4378 * Revert blinding in ECDSA sign and instead make problematic addition
4379 length-invariant. Switch even to fixed-length Montgomery multiplication.
4383 * Change generating and checking of primes so that the error rate of not
4384 being prime depends on the intended use based on the size of the input.
4385 For larger primes this will result in more rounds of Miller-Rabin.
4386 The maximal error rate for primes with more than 1080 bits is lowered
4389 *Kurt Roeckx, Annie Yousar*
4391 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4395 * Add blinding to ECDSA and DSA signatures to protect against side channel
4396 attacks discovered by Keegan Ryan (NCC Group).
4400 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4401 now allow empty (zero character) pass phrases.
4405 * Certificate time validation (X509_cmp_time) enforces stricter
4406 compliance with RFC 5280. Fractional seconds and timezone offsets
4407 are no longer allowed.
4411 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4413 * Constructed ASN.1 types with a recursive definition could exceed the stack
4415 Constructed ASN.1 types with a recursive definition (such as can be found
4416 in PKCS7) could eventually exceed the stack given malicious input with
4417 excessive recursion. This could result in a Denial Of Service attack. There
4418 are no such structures used within SSL/TLS that come from untrusted sources
4419 so this is considered safe.
4421 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4427 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4429 * Read/write after SSL object in error state
4431 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4432 mechanism. The intent was that if a fatal error occurred during a handshake
4433 then OpenSSL would move into the error state and would immediately fail if
4434 you attempted to continue the handshake. This works as designed for the
4435 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4436 SSL_connect()), however due to a bug it does not work correctly if
4437 SSL_read() or SSL_write() is called directly. In that scenario, if the
4438 handshake fails then a fatal error will be returned in the initial function
4439 call. If SSL_read()/SSL_write() is subsequently called by the application
4440 for the same SSL object then it will succeed and the data is passed without
4441 being decrypted/encrypted directly from the SSL/TLS record layer.
4443 In order to exploit this issue an application bug would have to be present
4444 that resulted in a call to SSL_read()/SSL_write() being issued after having
4445 already received a fatal error.
4447 This issue was reported to OpenSSL by David Benjamin (Google).
4452 * rsaz_1024_mul_avx2 overflow bug on x86_64
4454 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4455 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4456 Analysis suggests that attacks against RSA and DSA as a result of this
4457 defect would be very difficult to perform and are not believed likely.
4458 Attacks against DH1024 are considered just feasible, because most of the
4459 work necessary to deduce information about a private key may be performed
4460 offline. The amount of resources required for such an attack would be
4461 significant. However, for an attack on TLS to be meaningful, the server
4462 would have to share the DH1024 private key among multiple clients, which is
4463 no longer an option since CVE-2016-0701.
4465 This only affects processors that support the AVX2 but not ADX extensions
4466 like Intel Haswell (4th generation).
4468 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4469 was originally found via the OSS-Fuzz project.
4474 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4476 * bn_sqrx8x_internal carry bug on x86_64
4478 There is a carry propagating bug in the x86_64 Montgomery squaring
4479 procedure. No EC algorithms are affected. Analysis suggests that attacks
4480 against RSA and DSA as a result of this defect would be very difficult to
4481 perform and are not believed likely. Attacks against DH are considered just
4482 feasible (although very difficult) because most of the work necessary to
4483 deduce information about a private key may be performed offline. The amount
4484 of resources required for such an attack would be very significant and
4485 likely only accessible to a limited number of attackers. An attacker would
4486 additionally need online access to an unpatched system using the target
4487 private key in a scenario with persistent DH parameters and a private
4488 key that is shared between multiple clients.
4490 This only affects processors that support the BMI1, BMI2 and ADX extensions
4491 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4493 This issue was reported to OpenSSL by the OSS-Fuzz project.
4498 * Malformed X.509 IPAddressFamily could cause OOB read
4500 If an X.509 certificate has a malformed IPAddressFamily extension,
4501 OpenSSL could do a one-byte buffer overread. The most likely result
4502 would be an erroneous display of the certificate in text format.
4504 This issue was reported to OpenSSL by the OSS-Fuzz project.
4509 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4511 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4512 platform rather than 'mingw'.
4516 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4518 * Truncated packet could crash via OOB read
4520 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4521 cipher is being used, then a truncated packet can cause that host to
4522 perform an out-of-bounds read, usually resulting in a crash.
4524 This issue was reported to OpenSSL by Robert Święcki of Google.
4529 * BN_mod_exp may produce incorrect results on x86_64
4531 There is a carry propagating bug in the x86_64 Montgomery squaring
4532 procedure. No EC algorithms are affected. Analysis suggests that attacks
4533 against RSA and DSA as a result of this defect would be very difficult to
4534 perform and are not believed likely. Attacks against DH are considered just
4535 feasible (although very difficult) because most of the work necessary to
4536 deduce information about a private key may be performed offline. The amount
4537 of resources required for such an attack would be very significant and
4538 likely only accessible to a limited number of attackers. An attacker would
4539 additionally need online access to an unpatched system using the target
4540 private key in a scenario with persistent DH parameters and a private
4541 key that is shared between multiple clients. For example this can occur by
4542 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4543 similar to CVE-2015-3193 but must be treated as a separate problem.
4545 This issue was reported to OpenSSL by the OSS-Fuzz project.
4550 * Montgomery multiplication may produce incorrect results
4552 There is a carry propagating bug in the Broadwell-specific Montgomery
4553 multiplication procedure that handles input lengths divisible by, but
4554 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4555 and DH private keys are impossible. This is because the subroutine in
4556 question is not used in operations with the private key itself and an input
4557 of the attacker's direct choice. Otherwise the bug can manifest itself as
4558 transient authentication and key negotiation failures or reproducible
4559 erroneous outcome of public-key operations with specially crafted input.
4560 Among EC algorithms only Brainpool P-512 curves are affected and one
4561 presumably can attack ECDH key negotiation. Impact was not analyzed in
4562 detail, because pre-requisites for attack are considered unlikely. Namely
4563 multiple clients have to choose the curve in question and the server has to
4564 share the private key among them, neither of which is default behaviour.
4565 Even then only clients that chose the curve will be affected.
4567 This issue was publicly reported as transient failures and was not
4568 initially recognized as a security issue. Thanks to Richard Morgan for
4569 providing reproducible case.
4574 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4575 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4576 prevent issues where no progress is being made and the peer continually
4577 sends unrecognised record types, using up resources processing them.
4581 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4583 * Missing CRL sanity check
4585 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4586 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4587 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4589 This issue only affects the OpenSSL 1.0.2i
4594 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4596 * OCSP Status Request extension unbounded memory growth
4598 A malicious client can send an excessively large OCSP Status Request
4599 extension. If that client continually requests renegotiation, sending a
4600 large OCSP Status Request extension each time, then there will be unbounded
4601 memory growth on the server. This will eventually lead to a Denial Of
4602 Service attack through memory exhaustion. Servers with a default
4603 configuration are vulnerable even if they do not support OCSP. Builds using
4604 the "no-ocsp" build time option are not affected.
4606 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4611 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4614 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4620 * OOB write in MDC2_Update()
4622 An overflow can occur in MDC2_Update() either if called directly or
4623 through the EVP_DigestUpdate() function using MDC2. If an attacker
4624 is able to supply very large amounts of input data after a previous
4625 call to EVP_EncryptUpdate() with a partial block then a length check
4626 can overflow resulting in a heap corruption.
4628 The amount of data needed is comparable to SIZE_MAX which is impractical
4631 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4636 * Malformed SHA512 ticket DoS
4638 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4639 DoS attack where a malformed ticket will result in an OOB read which will
4642 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4643 a custom server callback and ticket lookup mechanism.
4645 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4650 * OOB write in BN_bn2dec()
4652 The function BN_bn2dec() does not check the return value of BN_div_word().
4653 This can cause an OOB write if an application uses this function with an
4654 overly large BIGNUM. This could be a problem if an overly large certificate
4655 or CRL is printed out from an untrusted source. TLS is not affected because
4656 record limits will reject an oversized certificate before it is parsed.
4658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4663 * OOB read in TS_OBJ_print_bio()
4665 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4666 the total length the OID text representation would use and not the amount
4667 of data written. This will result in OOB reads when large OIDs are
4670 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4675 * Pointer arithmetic undefined behaviour
4677 Avoid some undefined pointer arithmetic
4679 A common idiom in the codebase is to check limits in the following manner:
4682 Where "p" points to some malloc'd data of SIZE bytes and
4685 "len" here could be from some externally supplied data (e.g. from a TLS
4688 The rules of C pointer arithmetic are such that "p + len" is only well
4689 defined where len <= SIZE. Therefore the above idiom is actually
4690 undefined behaviour.
4692 For example this could cause problems if some malloc implementation
4693 provides an address for "p" such that "p + len" actually overflows for
4694 values of len that are too big and therefore p + len < limit.
4696 This issue was reported to OpenSSL by Guido Vranken
4701 * Constant time flag not preserved in DSA signing
4703 Operations in the DSA signing algorithm should run in constant time in
4704 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4705 implementation means that a non-constant time codepath is followed for
4706 certain operations. This has been demonstrated through a cache-timing
4707 attack to be sufficient for an attacker to recover the private DSA key.
4709 This issue was reported by César Pereida (Aalto University), Billy Brumley
4710 (Tampere University of Technology), and Yuval Yarom (The University of
4711 Adelaide and NICTA).
4716 * DTLS buffered message DoS
4718 In a DTLS connection where handshake messages are delivered out-of-order
4719 those messages that OpenSSL is not yet ready to process will be buffered
4720 for later use. Under certain circumstances, a flaw in the logic means that
4721 those messages do not get removed from the buffer even though the handshake
4722 has been completed. An attacker could force up to approx. 15 messages to
4723 remain in the buffer when they are no longer required. These messages will
4724 be cleared when the DTLS connection is closed. The default maximum size for
4725 a message is 100k. Therefore the attacker could force an additional 1500k
4726 to be consumed per connection. By opening many simulataneous connections an
4727 attacker could cause a DoS attack through memory exhaustion.
4729 This issue was reported to OpenSSL by Quan Luo.
4734 * DTLS replay protection DoS
4736 A flaw in the DTLS replay attack protection mechanism means that records
4737 that arrive for future epochs update the replay protection "window" before
4738 the MAC for the record has been validated. This could be exploited by an
4739 attacker by sending a record for the next epoch (which does not have to
4740 decrypt or have a valid MAC), with a very large sequence number. This means
4741 that all subsequent legitimate packets are dropped causing a denial of
4742 service for a specific DTLS connection.
4744 This issue was reported to OpenSSL by the OCAP audit team.
4749 * Certificate message OOB reads
4751 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4752 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4753 theoretical DoS risk but this has not been observed in practice on common
4756 The messages affected are client certificate, client certificate request
4757 and server certificate. As a result the attack can only be performed
4758 against a client or a server which enables client authentication.
4760 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4765 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4767 * Prevent padding oracle in AES-NI CBC MAC check
4769 A MITM attacker can use a padding oracle attack to decrypt traffic
4770 when the connection uses an AES CBC cipher and the server support
4773 This issue was introduced as part of the fix for Lucky 13 padding
4774 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4775 constant time by making sure that always the same bytes are read and
4776 compared against either the MAC or padding bytes. But it no longer
4777 checked that there was enough data to have both the MAC and padding
4780 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4785 * Fix EVP_EncodeUpdate overflow
4787 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4788 Base64 encoding of binary data. If an attacker is able to supply very large
4789 amounts of input data then a length check can overflow resulting in a heap
4792 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4793 the `PEM_write_bio*` family of functions. These are mainly used within the
4794 OpenSSL command line applications, so any application which processes data
4795 from an untrusted source and outputs it as a PEM file should be considered
4796 vulnerable to this issue. User applications that call these APIs directly
4797 with large amounts of untrusted data may also be vulnerable.
4799 This issue was reported by Guido Vranken.
4804 * Fix EVP_EncryptUpdate overflow
4806 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4807 is able to supply very large amounts of input data after a previous call to
4808 EVP_EncryptUpdate() with a partial block then a length check can overflow
4809 resulting in a heap corruption. Following an analysis of all OpenSSL
4810 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4811 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4812 the first called function after an EVP_EncryptInit(), and therefore that
4813 specific call must be safe. The second form is where the length passed to
4814 EVP_EncryptUpdate() can be seen from the code to be some small value and
4815 therefore there is no possibility of an overflow. Since all instances are
4816 one of these two forms, it is believed that there can be no overflows in
4817 internal code due to this problem. It should be noted that
4818 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4819 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4820 of these calls have also been analysed too and it is believed there are no
4821 instances in internal usage where an overflow could occur.
4823 This issue was reported by Guido Vranken.
4828 * Prevent ASN.1 BIO excessive memory allocation
4830 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4831 a short invalid encoding can cause allocation of large amounts of memory
4832 potentially consuming excessive resources or exhausting memory.
4834 Any application parsing untrusted data through d2i BIO functions is
4835 affected. The memory based functions such as d2i_X509() are *not* affected.
4836 Since the memory based functions are used by the TLS library, TLS
4837 applications are not affected.
4839 This issue was reported by Brian Carpenter.
4846 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4847 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4848 in arbitrary stack data being returned in the buffer.
4850 This issue was reported by Guido Vranken.
4855 * Modify behavior of ALPN to invoke callback after SNI/servername
4856 callback, such that updates to the SSL_CTX affect ALPN.
4860 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4865 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4866 methods are enabled and ssl2 is disabled the methods return NULL.
4870 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4872 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4873 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4874 provide any "EXPORT" or "LOW" strength ciphers.
4878 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4879 is by default disabled at build-time. Builds that are not configured with
4880 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4881 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4882 will need to explicitly call either of:
4884 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4886 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4888 as appropriate. Even if either of those is used, or the application
4889 explicitly uses the version-specific SSLv2_method() or its client and
4890 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4891 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4892 ciphers, and SSLv2 56-bit DES are no longer available.
4897 * Fix a double-free in DSA code
4899 A double free bug was discovered when OpenSSL parses malformed DSA private
4900 keys and could lead to a DoS attack or memory corruption for applications
4901 that receive DSA private keys from untrusted sources. This scenario is
4904 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4910 * Disable SRP fake user seed to address a server memory leak.
4912 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4914 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4915 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4916 was changed to ignore the "fake user" SRP seed, even if the seed
4919 Users should use SRP_VBASE_get1_by_user instead. Note that in
4920 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4921 also that even though configuring the SRP seed attempts to hide
4922 invalid usernames by continuing the handshake with fake
4923 credentials, this behaviour is not constant time and no strong
4924 guarantees are made that the handshake is indistinguishable from
4925 that of a valid user.
4930 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4932 In the BN_hex2bn function the number of hex digits is calculated using an
4933 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4934 large values of `i` this can result in `bn_expand` not allocating any
4935 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4936 field as NULL leading to a subsequent NULL ptr deref. For very large values
4937 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4938 In this case memory is allocated to the internal BIGNUM data field, but it
4939 is insufficiently sized leading to heap corruption. A similar issue exists
4940 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4941 is ever called by user applications with very large untrusted hex/dec data.
4942 This is anticipated to be a rare occurrence.
4944 All OpenSSL internal usage of these functions use data that is not expected
4945 to be untrusted, e.g. config file data or application command line
4946 arguments. If user developed applications generate config file data based
4947 on untrusted data then it is possible that this could also lead to security
4948 consequences. This is also anticipated to be rare.
4950 This issue was reported to OpenSSL by Guido Vranken.
4955 * Fix memory issues in `BIO_*printf` functions
4957 The internal `fmtstr` function used in processing a "%s" format string in
4958 the `BIO_*printf` functions could overflow while calculating the length of a
4959 string and cause an OOB read when printing very long strings.
4961 Additionally the internal `doapr_outch` function can attempt to write to an
4962 OOB memory location (at an offset from the NULL pointer) in the event of a
4963 memory allocation failure. In 1.0.2 and below this could be caused where
4964 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4965 could be in processing a very long "%s" format string. Memory leaks can
4968 The first issue may mask the second issue dependent on compiler behaviour.
4969 These problems could enable attacks where large amounts of untrusted data
4970 is passed to the `BIO_*printf` functions. If applications use these functions
4971 in this way then they could be vulnerable. OpenSSL itself uses these
4972 functions when printing out human-readable dumps of ASN.1 data. Therefore
4973 applications that print this data could be vulnerable if the data is from
4974 untrusted sources. OpenSSL command line applications could also be
4975 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4976 as command line arguments.
4978 Libssl is not considered directly vulnerable. Additionally certificates etc
4979 received via remote connections via libssl are also unlikely to be able to
4980 trigger these issues because of message size limits enforced within libssl.
4982 This issue was reported to OpenSSL Guido Vranken.
4987 * Side channel attack on modular exponentiation
4989 A side-channel attack was found which makes use of cache-bank conflicts on
4990 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4991 of RSA keys. The ability to exploit this issue is limited as it relies on
4992 an attacker who has control of code in a thread running on the same
4993 hyper-threaded core as the victim thread which is performing decryptions.
4995 This issue was reported to OpenSSL by Yuval Yarom, The University of
4996 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4997 Nadia Heninger, University of Pennsylvania with more information at
4998 <http://cachebleed.info>.
5003 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5004 if no keysize is specified with default_bits. This fixes an
5005 omission in an earlier change that changed all RSA/DSA key generation
5006 commands to use 2048 bits by default.
5010 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5012 * DH small subgroups
5014 Historically OpenSSL only ever generated DH parameters based on "safe"
5015 primes. More recently (in version 1.0.2) support was provided for
5016 generating X9.42 style parameter files such as those required for RFC 5114
5017 support. The primes used in such files may not be "safe". Where an
5018 application is using DH configured with parameters based on primes that are
5019 not "safe" then an attacker could use this fact to find a peer's private
5020 DH exponent. This attack requires that the attacker complete multiple
5021 handshakes in which the peer uses the same private DH exponent. For example
5022 this could be used to discover a TLS server's private DH exponent if it's
5023 reusing the private DH exponent or it's using a static DH ciphersuite.
5025 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5026 TLS. It is not on by default. If the option is not set then the server
5027 reuses the same private DH exponent for the life of the server process and
5028 would be vulnerable to this attack. It is believed that many popular
5029 applications do set this option and would therefore not be at risk.
5031 The fix for this issue adds an additional check where a "q" parameter is
5032 available (as is the case in X9.42 based parameters). This detects the
5033 only known attack, and is the only possible defense for static DH
5034 ciphersuites. This could have some performance impact.
5036 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5037 default and cannot be disabled. This could have some performance impact.
5039 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5044 * SSLv2 doesn't block disabled ciphers
5046 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5047 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5048 been disabled, provided that the SSLv2 protocol was not also disabled via
5051 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5052 and Sebastian Schinzel.
5057 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5059 * BN_mod_exp may produce incorrect results on x86_64
5061 There is a carry propagating bug in the x86_64 Montgomery squaring
5062 procedure. No EC algorithms are affected. Analysis suggests that attacks
5063 against RSA and DSA as a result of this defect would be very difficult to
5064 perform and are not believed likely. Attacks against DH are considered just
5065 feasible (although very difficult) because most of the work necessary to
5066 deduce information about a private key may be performed offline. The amount
5067 of resources required for such an attack would be very significant and
5068 likely only accessible to a limited number of attackers. An attacker would
5069 additionally need online access to an unpatched system using the target
5070 private key in a scenario with persistent DH parameters and a private
5071 key that is shared between multiple clients. For example this can occur by
5072 default in OpenSSL DHE based SSL/TLS ciphersuites.
5074 This issue was reported to OpenSSL by Hanno Böck.
5079 * Certificate verify crash with missing PSS parameter
5081 The signature verification routines will crash with a NULL pointer
5082 dereference if presented with an ASN.1 signature using the RSA PSS
5083 algorithm and absent mask generation function parameter. Since these
5084 routines are used to verify certificate signature algorithms this can be
5085 used to crash any certificate verification operation and exploited in a
5086 DoS attack. Any application which performs certificate verification is
5087 vulnerable including OpenSSL clients and servers which enable client
5090 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5095 * X509_ATTRIBUTE memory leak
5097 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5098 memory. This structure is used by the PKCS#7 and CMS routines so any
5099 application which reads PKCS#7 or CMS data from untrusted sources is
5100 affected. SSL/TLS is not affected.
5102 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5108 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5109 This changes the decoding behaviour for some invalid messages,
5110 though the change is mostly in the more lenient direction, and
5111 legacy behaviour is preserved as much as possible.
5115 * In DSA_generate_parameters_ex, if the provided seed is too short,
5118 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5120 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5122 * Alternate chains certificate forgery
5124 During certificate verification, OpenSSL will attempt to find an
5125 alternative certificate chain if the first attempt to build such a chain
5126 fails. An error in the implementation of this logic can mean that an
5127 attacker could cause certain checks on untrusted certificates to be
5128 bypassed, such as the CA flag, enabling them to use a valid leaf
5129 certificate to act as a CA and "issue" an invalid certificate.
5131 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5136 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5138 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5139 incompatibility in the handling of HMAC. The previous ABI has now been
5144 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5146 * Malformed ECParameters causes infinite loop
5148 When processing an ECParameters structure OpenSSL enters an infinite loop
5149 if the curve specified is over a specially malformed binary polynomial
5152 This can be used to perform denial of service against any
5153 system which processes public keys, certificate requests or
5154 certificates. This includes TLS clients and TLS servers with
5155 client authentication enabled.
5157 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5162 * Exploitable out-of-bounds read in X509_cmp_time
5164 X509_cmp_time does not properly check the length of the ASN1_TIME
5165 string and can read a few bytes out of bounds. In addition,
5166 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5169 An attacker can use this to craft malformed certificates and CRLs of
5170 various sizes and potentially cause a segmentation fault, resulting in
5171 a DoS on applications that verify certificates or CRLs. TLS clients
5172 that verify CRLs are affected. TLS clients and servers with client
5173 authentication enabled may be affected if they use custom verification
5176 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5177 independently by Hanno Böck.
5182 * PKCS7 crash with missing EnvelopedContent
5184 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5185 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5186 with missing content and trigger a NULL pointer dereference on parsing.
5188 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5189 structures from untrusted sources are affected. OpenSSL clients and
5190 servers are not affected.
5192 This issue was reported to OpenSSL by Michal Zalewski (Google).
5197 * CMS verify infinite loop with unknown hash function
5199 When verifying a signedData message the CMS code can enter an infinite loop
5200 if presented with an unknown hash function OID. This can be used to perform
5201 denial of service against any system which verifies signedData messages using
5203 This issue was reported to OpenSSL by Johannes Bauer.
5208 * Race condition handling NewSessionTicket
5210 If a NewSessionTicket is received by a multi-threaded client when attempting to
5211 reuse a previous ticket then a race condition can occur potentially leading to
5212 a double free of the ticket data.
5217 * Only support 256-bit or stronger elliptic curves with the
5218 'ecdh_auto' setting (server) or by default (client). Of supported
5219 curves, prefer P-256 (both).
5223 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5225 * ClientHello sigalgs DoS fix
5227 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5228 invalid signature algorithms extension a NULL pointer dereference will
5229 occur. This can be exploited in a DoS attack against the server.
5231 This issue was was reported to OpenSSL by David Ramos of Stanford
5235 *Stephen Henson and Matt Caswell*
5237 * Multiblock corrupted pointer fix
5239 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5240 feature only applies on 64 bit x86 architecture platforms that support AES
5241 NI instructions. A defect in the implementation of "multiblock" can cause
5242 OpenSSL's internal write buffer to become incorrectly set to NULL when
5243 using non-blocking IO. Typically, when the user application is using a
5244 socket BIO for writing, this will only result in a failed connection.
5245 However if some other BIO is used then it is likely that a segmentation
5246 fault will be triggered, thus enabling a potential DoS attack.
5248 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5253 * Segmentation fault in DTLSv1_listen fix
5255 The DTLSv1_listen function is intended to be stateless and processes the
5256 initial ClientHello from many peers. It is common for user code to loop
5257 over the call to DTLSv1_listen until a valid ClientHello is received with
5258 an associated cookie. A defect in the implementation of DTLSv1_listen means
5259 that state is preserved in the SSL object from one invocation to the next
5260 that can lead to a segmentation fault. Errors processing the initial
5261 ClientHello can trigger this scenario. An example of such an error could be
5262 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5265 This issue was reported to OpenSSL by Per Allansson.
5270 * Segmentation fault in ASN1_TYPE_cmp fix
5272 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5273 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5274 certificate signature algorithm consistency this can be used to crash any
5275 certificate verification operation and exploited in a DoS attack. Any
5276 application which performs certificate verification is vulnerable including
5277 OpenSSL clients and servers which enable client authentication.
5282 * Segmentation fault for invalid PSS parameters fix
5284 The signature verification routines will crash with a NULL pointer
5285 dereference if presented with an ASN.1 signature using the RSA PSS
5286 algorithm and invalid parameters. Since these routines are used to verify
5287 certificate signature algorithms this can be used to crash any
5288 certificate verification operation and exploited in a DoS attack. Any
5289 application which performs certificate verification is vulnerable including
5290 OpenSSL clients and servers which enable client authentication.
5292 This issue was was reported to OpenSSL by Brian Carpenter.
5297 * ASN.1 structure reuse memory corruption fix
5299 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5300 memory corruption via an invalid write. Such reuse is and has been
5301 strongly discouraged and is believed to be rare.
5303 Applications that parse structures containing CHOICE or ANY DEFINED BY
5304 components may be affected. Certificate parsing (d2i_X509 and related
5305 functions) are however not affected. OpenSSL clients and servers are
5311 * PKCS7 NULL pointer dereferences fix
5313 The PKCS#7 parsing code does not handle missing outer ContentInfo
5314 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5315 missing content and trigger a NULL pointer dereference on parsing.
5317 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5318 otherwise parse PKCS#7 structures from untrusted sources are
5319 affected. OpenSSL clients and servers are not affected.
5321 This issue was reported to OpenSSL by Michal Zalewski (Google).
5326 * DoS via reachable assert in SSLv2 servers fix
5328 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5329 servers that both support SSLv2 and enable export cipher suites by sending
5330 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5332 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5333 (OpenSSL development team).
5338 * Empty CKE with client auth and DHE fix
5340 If client auth is used then a server can seg fault in the event of a DHE
5341 ciphersuite being selected and a zero length ClientKeyExchange message
5342 being sent by the client. This could be exploited in a DoS attack.
5347 * Handshake with unseeded PRNG fix
5349 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5350 with an unseeded PRNG. The conditions are:
5351 - The client is on a platform where the PRNG has not been seeded
5352 automatically, and the user has not seeded manually
5353 - A protocol specific client method version has been used (i.e. not
5354 SSL_client_methodv23)
5355 - A ciphersuite is used that does not require additional random data from
5356 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5358 If the handshake succeeds then the client random that has been used will
5359 have been generated from a PRNG with insufficient entropy and therefore the
5360 output may be predictable.
5362 For example using the following command with an unseeded openssl will
5363 succeed on an unpatched platform:
5365 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5370 * Use After Free following d2i_ECPrivatekey error fix
5372 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5373 could cause a use after free condition. This, in turn, could cause a double
5374 free in several private key parsing functions (such as d2i_PrivateKey
5375 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5376 for applications that receive EC private keys from untrusted
5377 sources. This scenario is considered rare.
5379 This issue was discovered by the BoringSSL project and fixed in their
5385 * X509_to_X509_REQ NULL pointer deref fix
5387 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5388 the certificate key is invalid. This function is rarely used in practice.
5390 This issue was discovered by Brian Carpenter.
5395 * Removed the export ciphers from the DEFAULT ciphers
5399 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5401 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5402 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5403 So far those who have to target multiple platforms would compromise
5404 and argue that binary targeting say ARMv5 would still execute on
5405 ARMv8. "Universal" build resolves this compromise by providing
5406 near-optimal performance even on newer platforms.
5410 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5411 (other platforms pending).
5413 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5415 * Add support for the SignedCertificateTimestampList certificate and
5416 OCSP response extensions from RFC6962.
5420 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5421 for corner cases. (Certain input points at infinity could lead to
5422 bogus results, with non-infinity inputs mapped to infinity too.)
5426 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5427 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5428 common cases are optimized and there still is room for further
5429 improvements. Vector Permutation AES for Altivec is also added.
5433 * Add support for little-endian ppc64 Linux target.
5435 *Marcelo Cerri (IBM)*
5437 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5438 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5439 are optimized and there still is room for further improvements.
5440 Both 32- and 64-bit modes are supported.
5442 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5444 * Improved ARMv7 NEON support.
5448 * Support for SPARC Architecture 2011 crypto extensions, first
5449 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5450 SHA256/512, MD5, GHASH and modular exponentiation.
5452 *Andy Polyakov, David Miller*
5454 * Accelerated modular exponentiation for Intel processors, a.k.a.
5457 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5459 * Support for new and upcoming Intel processors, including AVX2,
5460 BMI and SHA ISA extensions. This includes additional "stitched"
5461 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5464 This work was sponsored by Intel Corp.
5468 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5469 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5470 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5474 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5475 this fixes a limitation in previous versions of OpenSSL.
5479 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5480 MGF1 digest and OAEP label.
5484 * Add EVP support for key wrapping algorithms, to avoid problems with
5485 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5486 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5487 algorithms and include tests cases.
5491 * Add functions to allocate and set the fields of an ECDSA_METHOD
5494 *Douglas E. Engert, Steve Henson*
5496 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5497 difference in days and seconds between two tm or ASN1_TIME structures.
5501 * Add -rev test option to s_server to just reverse order of characters
5502 received by client and send back to server. Also prints an abbreviated
5503 summary of the connection parameters.
5507 * New option -brief for s_client and s_server to print out a brief summary
5508 of connection parameters.
5512 * Add callbacks for arbitrary TLS extensions.
5514 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5516 * New option -crl_download in several openssl utilities to download CRLs
5517 from CRLDP extension in certificates.
5521 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5525 * New function X509_CRL_diff to generate a delta CRL from the difference
5526 of two full CRLs. Add support to "crl" utility.
5530 * New functions to set lookup_crls function and to retrieve
5531 X509_STORE from X509_STORE_CTX.
5535 * Print out deprecated issuer and subject unique ID fields in
5540 * Extend OCSP I/O functions so they can be used for simple general purpose
5541 HTTP as well as OCSP. New wrapper function which can be used to download
5542 CRLs using the OCSP API.
5546 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5550 * `SSL_CONF*` functions. These provide a common framework for application
5551 configuration using configuration files or command lines.
5555 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5556 message callback and prints the results. Needs compile time option
5557 "enable-ssl-trace". New options to s_client and s_server to enable
5562 * New ctrl and macro to retrieve supported points extensions.
5563 Print out extension in s_server and s_client.
5567 * New functions to retrieve certificate signature and signature
5572 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5577 * New Suite B modes for TLS code. These use and enforce the requirements
5578 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5579 only use Suite B curves. The Suite B modes can be set by using the
5580 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5584 * New chain verification flags for Suite B levels of security. Check
5585 algorithms are acceptable when flags are set in X509_verify_cert.
5589 * Make tls1_check_chain return a set of flags indicating checks passed
5590 by a certificate chain. Add additional tests to handle client
5591 certificates: checks for matching certificate type and issuer name
5596 * If an attempt is made to use a signature algorithm not in the peer
5597 preference list abort the handshake. If client has no suitable
5598 signature algorithms in response to a certificate request do not
5599 use the certificate.
5603 * If server EC tmp key is not in client preference list abort handshake.
5607 * Add support for certificate stores in CERT structure. This makes it
5608 possible to have different stores per SSL structure or one store in
5609 the parent SSL_CTX. Include distinct stores for certificate chain
5610 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5611 to build and store a certificate chain in CERT structure: returning
5612 an error if the chain cannot be built: this will allow applications
5613 to test if a chain is correctly configured.
5615 Note: if the CERT based stores are not set then the parent SSL_CTX
5616 store is used to retain compatibility with existing behaviour.
5620 * New function ssl_set_client_disabled to set a ciphersuite disabled
5621 mask based on the current session, check mask when sending client
5622 hello and checking the requested ciphersuite.
5626 * New ctrls to retrieve and set certificate types in a certificate
5627 request message. Print out received values in s_client. If certificate
5628 types is not set with custom values set sensible values based on
5629 supported signature algorithms.
5633 * Support for distinct client and server supported signature algorithms.
5637 * Add certificate callback. If set this is called whenever a certificate
5638 is required by client or server. An application can decide which
5639 certificate chain to present based on arbitrary criteria: for example
5640 supported signature algorithms. Add very simple example to s_server.
5641 This fixes many of the problems and restrictions of the existing client
5642 certificate callback: for example you can now clear an existing
5643 certificate and specify the whole chain.
5647 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5648 the certificate can be used for (if anything). Set valid_flags field
5649 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5650 to have similar checks in it.
5652 Add new "cert_flags" field to CERT structure and include a "strict mode".
5653 This enforces some TLS certificate requirements (such as only permitting
5654 certificate signature algorithms contained in the supported algorithms
5655 extension) which some implementations ignore: this option should be used
5656 with caution as it could cause interoperability issues.
5660 * Update and tidy signature algorithm extension processing. Work out
5661 shared signature algorithms based on preferences and peer algorithms
5662 and print them out in s_client and s_server. Abort handshake if no
5663 shared signature algorithms.
5667 * Add new functions to allow customised supported signature algorithms
5668 for SSL and SSL_CTX structures. Add options to s_client and s_server
5673 * New function SSL_certs_clear() to delete all references to certificates
5674 from an SSL structure. Before this once a certificate had been added
5675 it couldn't be removed.
5679 * Integrate hostname, email address and IP address checking with certificate
5680 verification. New verify options supporting checking in openssl utility.
5684 * Fixes and wildcard matching support to hostname and email checking
5685 functions. Add manual page.
5687 *Florian Weimer (Red Hat Product Security Team)*
5689 * New functions to check a hostname email or IP address against a
5690 certificate. Add options x509 utility to print results of checks against
5695 * Fix OCSP checking.
5697 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5699 * Initial experimental support for explicitly trusted non-root CAs.
5700 OpenSSL still tries to build a complete chain to a root but if an
5701 intermediate CA has a trust setting included that is used. The first
5702 setting is used: whether to trust (e.g., -addtrust option to the x509
5707 * Add -trusted_first option which attempts to find certificates in the
5708 trusted store even if an untrusted chain is also supplied.
5712 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5713 platform support for Linux and Android.
5717 * Support for linux-x32, ILP32 environment in x86_64 framework.
5721 * Experimental multi-implementation support for FIPS capable OpenSSL.
5722 When in FIPS mode the approved implementations are used as normal,
5723 when not in FIPS mode the internal unapproved versions are used instead.
5724 This means that the FIPS capable OpenSSL isn't forced to use the
5725 (often lower performance) FIPS implementations outside FIPS mode.
5729 * Transparently support X9.42 DH parameters when calling
5730 PEM_read_bio_DHparameters. This means existing applications can handle
5731 the new parameter format automatically.
5735 * Initial experimental support for X9.42 DH parameter format: mainly
5736 to support use of 'q' parameter for RFC5114 parameters.
5740 * Add DH parameters from RFC5114 including test data to dhtest.
5744 * Support for automatic EC temporary key parameter selection. If enabled
5745 the most preferred EC parameters are automatically used instead of
5746 hardcoded fixed parameters. Now a server just has to call:
5747 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5748 support ECDH and use the most appropriate parameters.
5752 * Enhance and tidy EC curve and point format TLS extension code. Use
5753 static structures instead of allocation if default values are used.
5754 New ctrls to set curves we wish to support and to retrieve shared curves.
5755 Print out shared curves in s_server. New options to s_server and s_client
5756 to set list of supported curves.
5760 * New ctrls to retrieve supported signature algorithms and
5761 supported curve values as an array of NIDs. Extend openssl utility
5762 to print out received values.
5766 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5767 between NIDs and the more common NIST names such as "P-256". Enhance
5768 ecparam utility and ECC method to recognise the NIST names for curves.
5772 * Enhance SSL/TLS certificate chain handling to support different
5773 chains for each certificate instead of one chain in the parent SSL_CTX.
5777 * Support for fixed DH ciphersuite client authentication: where both
5778 server and client use DH certificates with common parameters.
5782 * Support for fixed DH ciphersuites: those requiring DH server
5787 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5789 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5790 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5791 X509_CINF_get_signature were reverted post internal team review.
5796 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5798 * OCSP Status Request extension unbounded memory growth
5800 A malicious client can send an excessively large OCSP Status Request
5801 extension. If that client continually requests renegotiation, sending a
5802 large OCSP Status Request extension each time, then there will be unbounded
5803 memory growth on the server. This will eventually lead to a Denial Of
5804 Service attack through memory exhaustion. Servers with a default
5805 configuration are vulnerable even if they do not support OCSP. Builds using
5806 the "no-ocsp" build time option are not affected.
5808 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5813 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5816 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5822 * OOB write in MDC2_Update()
5824 An overflow can occur in MDC2_Update() either if called directly or
5825 through the EVP_DigestUpdate() function using MDC2. If an attacker
5826 is able to supply very large amounts of input data after a previous
5827 call to EVP_EncryptUpdate() with a partial block then a length check
5828 can overflow resulting in a heap corruption.
5830 The amount of data needed is comparable to SIZE_MAX which is impractical
5833 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5838 * Malformed SHA512 ticket DoS
5840 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5841 DoS attack where a malformed ticket will result in an OOB read which will
5844 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5845 a custom server callback and ticket lookup mechanism.
5847 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5852 * OOB write in BN_bn2dec()
5854 The function BN_bn2dec() does not check the return value of BN_div_word().
5855 This can cause an OOB write if an application uses this function with an
5856 overly large BIGNUM. This could be a problem if an overly large certificate
5857 or CRL is printed out from an untrusted source. TLS is not affected because
5858 record limits will reject an oversized certificate before it is parsed.
5860 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5865 * OOB read in TS_OBJ_print_bio()
5867 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5868 the total length the OID text representation would use and not the amount
5869 of data written. This will result in OOB reads when large OIDs are
5872 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5877 * Pointer arithmetic undefined behaviour
5879 Avoid some undefined pointer arithmetic
5881 A common idiom in the codebase is to check limits in the following manner:
5884 Where "p" points to some malloc'd data of SIZE bytes and
5887 "len" here could be from some externally supplied data (e.g. from a TLS
5890 The rules of C pointer arithmetic are such that "p + len" is only well
5891 defined where len <= SIZE. Therefore the above idiom is actually
5892 undefined behaviour.
5894 For example this could cause problems if some malloc implementation
5895 provides an address for "p" such that "p + len" actually overflows for
5896 values of len that are too big and therefore p + len < limit.
5898 This issue was reported to OpenSSL by Guido Vranken
5903 * Constant time flag not preserved in DSA signing
5905 Operations in the DSA signing algorithm should run in constant time in
5906 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5907 implementation means that a non-constant time codepath is followed for
5908 certain operations. This has been demonstrated through a cache-timing
5909 attack to be sufficient for an attacker to recover the private DSA key.
5911 This issue was reported by César Pereida (Aalto University), Billy Brumley
5912 (Tampere University of Technology), and Yuval Yarom (The University of
5913 Adelaide and NICTA).
5918 * DTLS buffered message DoS
5920 In a DTLS connection where handshake messages are delivered out-of-order
5921 those messages that OpenSSL is not yet ready to process will be buffered
5922 for later use. Under certain circumstances, a flaw in the logic means that
5923 those messages do not get removed from the buffer even though the handshake
5924 has been completed. An attacker could force up to approx. 15 messages to
5925 remain in the buffer when they are no longer required. These messages will
5926 be cleared when the DTLS connection is closed. The default maximum size for
5927 a message is 100k. Therefore the attacker could force an additional 1500k
5928 to be consumed per connection. By opening many simulataneous connections an
5929 attacker could cause a DoS attack through memory exhaustion.
5931 This issue was reported to OpenSSL by Quan Luo.
5936 * DTLS replay protection DoS
5938 A flaw in the DTLS replay attack protection mechanism means that records
5939 that arrive for future epochs update the replay protection "window" before
5940 the MAC for the record has been validated. This could be exploited by an
5941 attacker by sending a record for the next epoch (which does not have to
5942 decrypt or have a valid MAC), with a very large sequence number. This means
5943 that all subsequent legitimate packets are dropped causing a denial of
5944 service for a specific DTLS connection.
5946 This issue was reported to OpenSSL by the OCAP audit team.
5951 * Certificate message OOB reads
5953 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5954 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5955 theoretical DoS risk but this has not been observed in practice on common
5958 The messages affected are client certificate, client certificate request
5959 and server certificate. As a result the attack can only be performed
5960 against a client or a server which enables client authentication.
5962 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5967 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5969 * Prevent padding oracle in AES-NI CBC MAC check
5971 A MITM attacker can use a padding oracle attack to decrypt traffic
5972 when the connection uses an AES CBC cipher and the server support
5975 This issue was introduced as part of the fix for Lucky 13 padding
5976 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5977 constant time by making sure that always the same bytes are read and
5978 compared against either the MAC or padding bytes. But it no longer
5979 checked that there was enough data to have both the MAC and padding
5982 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5987 * Fix EVP_EncodeUpdate overflow
5989 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5990 Base64 encoding of binary data. If an attacker is able to supply very large
5991 amounts of input data then a length check can overflow resulting in a heap
5994 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5995 the `PEM_write_bio*` family of functions. These are mainly used within the
5996 OpenSSL command line applications, so any application which processes data
5997 from an untrusted source and outputs it as a PEM file should be considered
5998 vulnerable to this issue. User applications that call these APIs directly
5999 with large amounts of untrusted data may also be vulnerable.
6001 This issue was reported by Guido Vranken.
6006 * Fix EVP_EncryptUpdate overflow
6008 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6009 is able to supply very large amounts of input data after a previous call to
6010 EVP_EncryptUpdate() with a partial block then a length check can overflow
6011 resulting in a heap corruption. Following an analysis of all OpenSSL
6012 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6013 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6014 the first called function after an EVP_EncryptInit(), and therefore that
6015 specific call must be safe. The second form is where the length passed to
6016 EVP_EncryptUpdate() can be seen from the code to be some small value and
6017 therefore there is no possibility of an overflow. Since all instances are
6018 one of these two forms, it is believed that there can be no overflows in
6019 internal code due to this problem. It should be noted that
6020 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6021 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6022 of these calls have also been analysed too and it is believed there are no
6023 instances in internal usage where an overflow could occur.
6025 This issue was reported by Guido Vranken.
6030 * Prevent ASN.1 BIO excessive memory allocation
6032 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6033 a short invalid encoding can casuse allocation of large amounts of memory
6034 potentially consuming excessive resources or exhausting memory.
6036 Any application parsing untrusted data through d2i BIO functions is
6037 affected. The memory based functions such as d2i_X509() are *not* affected.
6038 Since the memory based functions are used by the TLS library, TLS
6039 applications are not affected.
6041 This issue was reported by Brian Carpenter.
6048 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6049 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6050 in arbitrary stack data being returned in the buffer.
6052 This issue was reported by Guido Vranken.
6057 * Modify behavior of ALPN to invoke callback after SNI/servername
6058 callback, such that updates to the SSL_CTX affect ALPN.
6062 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6067 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6068 methods are enabled and ssl2 is disabled the methods return NULL.
6072 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6074 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6075 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6076 provide any "EXPORT" or "LOW" strength ciphers.
6080 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6081 is by default disabled at build-time. Builds that are not configured with
6082 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6083 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6084 will need to explicitly call either of:
6086 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6088 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6090 as appropriate. Even if either of those is used, or the application
6091 explicitly uses the version-specific SSLv2_method() or its client and
6092 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6093 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6094 ciphers, and SSLv2 56-bit DES are no longer available.
6099 * Fix a double-free in DSA code
6101 A double free bug was discovered when OpenSSL parses malformed DSA private
6102 keys and could lead to a DoS attack or memory corruption for applications
6103 that receive DSA private keys from untrusted sources. This scenario is
6106 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6112 * Disable SRP fake user seed to address a server memory leak.
6114 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6116 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6117 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6118 was changed to ignore the "fake user" SRP seed, even if the seed
6121 Users should use SRP_VBASE_get1_by_user instead. Note that in
6122 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6123 also that even though configuring the SRP seed attempts to hide
6124 invalid usernames by continuing the handshake with fake
6125 credentials, this behaviour is not constant time and no strong
6126 guarantees are made that the handshake is indistinguishable from
6127 that of a valid user.
6132 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6134 In the BN_hex2bn function the number of hex digits is calculated using an
6135 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6136 large values of `i` this can result in `bn_expand` not allocating any
6137 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6138 field as NULL leading to a subsequent NULL ptr deref. For very large values
6139 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6140 In this case memory is allocated to the internal BIGNUM data field, but it
6141 is insufficiently sized leading to heap corruption. A similar issue exists
6142 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6143 is ever called by user applications with very large untrusted hex/dec data.
6144 This is anticipated to be a rare occurrence.
6146 All OpenSSL internal usage of these functions use data that is not expected
6147 to be untrusted, e.g. config file data or application command line
6148 arguments. If user developed applications generate config file data based
6149 on untrusted data then it is possible that this could also lead to security
6150 consequences. This is also anticipated to be rare.
6152 This issue was reported to OpenSSL by Guido Vranken.
6157 * Fix memory issues in `BIO_*printf` functions
6159 The internal `fmtstr` function used in processing a "%s" format string in
6160 the `BIO_*printf` functions could overflow while calculating the length of a
6161 string and cause an OOB read when printing very long strings.
6163 Additionally the internal `doapr_outch` function can attempt to write to an
6164 OOB memory location (at an offset from the NULL pointer) in the event of a
6165 memory allocation failure. In 1.0.2 and below this could be caused where
6166 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6167 could be in processing a very long "%s" format string. Memory leaks can
6170 The first issue may mask the second issue dependent on compiler behaviour.
6171 These problems could enable attacks where large amounts of untrusted data
6172 is passed to the `BIO_*printf` functions. If applications use these functions
6173 in this way then they could be vulnerable. OpenSSL itself uses these
6174 functions when printing out human-readable dumps of ASN.1 data. Therefore
6175 applications that print this data could be vulnerable if the data is from
6176 untrusted sources. OpenSSL command line applications could also be
6177 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6178 as command line arguments.
6180 Libssl is not considered directly vulnerable. Additionally certificates etc
6181 received via remote connections via libssl are also unlikely to be able to
6182 trigger these issues because of message size limits enforced within libssl.
6184 This issue was reported to OpenSSL Guido Vranken.
6189 * Side channel attack on modular exponentiation
6191 A side-channel attack was found which makes use of cache-bank conflicts on
6192 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6193 of RSA keys. The ability to exploit this issue is limited as it relies on
6194 an attacker who has control of code in a thread running on the same
6195 hyper-threaded core as the victim thread which is performing decryptions.
6197 This issue was reported to OpenSSL by Yuval Yarom, The University of
6198 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6199 Nadia Heninger, University of Pennsylvania with more information at
6200 <http://cachebleed.info>.
6205 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6206 if no keysize is specified with default_bits. This fixes an
6207 omission in an earlier change that changed all RSA/DSA key generation
6208 commands to use 2048 bits by default.
6212 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6214 * Protection for DH small subgroup attacks
6216 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6217 switched on by default and cannot be disabled. This could have some
6222 * SSLv2 doesn't block disabled ciphers
6224 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6225 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6226 been disabled, provided that the SSLv2 protocol was not also disabled via
6229 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6230 and Sebastian Schinzel.
6235 * Reject DH handshakes with parameters shorter than 1024 bits.
6239 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6241 * Certificate verify crash with missing PSS parameter
6243 The signature verification routines will crash with a NULL pointer
6244 dereference if presented with an ASN.1 signature using the RSA PSS
6245 algorithm and absent mask generation function parameter. Since these
6246 routines are used to verify certificate signature algorithms this can be
6247 used to crash any certificate verification operation and exploited in a
6248 DoS attack. Any application which performs certificate verification is
6249 vulnerable including OpenSSL clients and servers which enable client
6252 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6257 * X509_ATTRIBUTE memory leak
6259 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6260 memory. This structure is used by the PKCS#7 and CMS routines so any
6261 application which reads PKCS#7 or CMS data from untrusted sources is
6262 affected. SSL/TLS is not affected.
6264 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6270 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6271 This changes the decoding behaviour for some invalid messages,
6272 though the change is mostly in the more lenient direction, and
6273 legacy behaviour is preserved as much as possible.
6277 * In DSA_generate_parameters_ex, if the provided seed is too short,
6278 use a random seed, as already documented.
6280 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6282 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6284 * Alternate chains certificate forgery
6286 During certificate verfification, OpenSSL will attempt to find an
6287 alternative certificate chain if the first attempt to build such a chain
6288 fails. An error in the implementation of this logic can mean that an
6289 attacker could cause certain checks on untrusted certificates to be
6290 bypassed, such as the CA flag, enabling them to use a valid leaf
6291 certificate to act as a CA and "issue" an invalid certificate.
6293 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6299 * Race condition handling PSK identify hint
6301 If PSK identity hints are received by a multi-threaded client then
6302 the values are wrongly updated in the parent SSL_CTX structure. This can
6303 result in a race condition potentially leading to a double free of the
6309 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6311 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6312 incompatibility in the handling of HMAC. The previous ABI has now been
6315 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6317 * Malformed ECParameters causes infinite loop
6319 When processing an ECParameters structure OpenSSL enters an infinite loop
6320 if the curve specified is over a specially malformed binary polynomial
6323 This can be used to perform denial of service against any
6324 system which processes public keys, certificate requests or
6325 certificates. This includes TLS clients and TLS servers with
6326 client authentication enabled.
6328 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6333 * Exploitable out-of-bounds read in X509_cmp_time
6335 X509_cmp_time does not properly check the length of the ASN1_TIME
6336 string and can read a few bytes out of bounds. In addition,
6337 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6340 An attacker can use this to craft malformed certificates and CRLs of
6341 various sizes and potentially cause a segmentation fault, resulting in
6342 a DoS on applications that verify certificates or CRLs. TLS clients
6343 that verify CRLs are affected. TLS clients and servers with client
6344 authentication enabled may be affected if they use custom verification
6347 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6348 independently by Hanno Böck.
6353 * PKCS7 crash with missing EnvelopedContent
6355 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6356 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6357 with missing content and trigger a NULL pointer dereference on parsing.
6359 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6360 structures from untrusted sources are affected. OpenSSL clients and
6361 servers are not affected.
6363 This issue was reported to OpenSSL by Michal Zalewski (Google).
6368 * CMS verify infinite loop with unknown hash function
6370 When verifying a signedData message the CMS code can enter an infinite loop
6371 if presented with an unknown hash function OID. This can be used to perform
6372 denial of service against any system which verifies signedData messages using
6374 This issue was reported to OpenSSL by Johannes Bauer.
6379 * Race condition handling NewSessionTicket
6381 If a NewSessionTicket is received by a multi-threaded client when attempting to
6382 reuse a previous ticket then a race condition can occur potentially leading to
6383 a double free of the ticket data.
6388 * Reject DH handshakes with parameters shorter than 768 bits.
6390 *Kurt Roeckx and Emilia Kasper*
6392 * dhparam: generate 2048-bit parameters by default.
6394 *Kurt Roeckx and Emilia Kasper*
6396 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6398 * Segmentation fault in ASN1_TYPE_cmp fix
6400 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6401 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6402 certificate signature algorithm consistency this can be used to crash any
6403 certificate verification operation and exploited in a DoS attack. Any
6404 application which performs certificate verification is vulnerable including
6405 OpenSSL clients and servers which enable client authentication.
6410 * ASN.1 structure reuse memory corruption fix
6412 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6413 memory corruption via an invalid write. Such reuse is and has been
6414 strongly discouraged and is believed to be rare.
6416 Applications that parse structures containing CHOICE or ANY DEFINED BY
6417 components may be affected. Certificate parsing (d2i_X509 and related
6418 functions) are however not affected. OpenSSL clients and servers are
6424 * PKCS7 NULL pointer dereferences fix
6426 The PKCS#7 parsing code does not handle missing outer ContentInfo
6427 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6428 missing content and trigger a NULL pointer dereference on parsing.
6430 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6431 otherwise parse PKCS#7 structures from untrusted sources are
6432 affected. OpenSSL clients and servers are not affected.
6434 This issue was reported to OpenSSL by Michal Zalewski (Google).
6439 * DoS via reachable assert in SSLv2 servers fix
6441 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6442 servers that both support SSLv2 and enable export cipher suites by sending
6443 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6445 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6446 (OpenSSL development team).
6451 * Use After Free following d2i_ECPrivatekey error fix
6453 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6454 could cause a use after free condition. This, in turn, could cause a double
6455 free in several private key parsing functions (such as d2i_PrivateKey
6456 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6457 for applications that receive EC private keys from untrusted
6458 sources. This scenario is considered rare.
6460 This issue was discovered by the BoringSSL project and fixed in their
6466 * X509_to_X509_REQ NULL pointer deref fix
6468 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6469 the certificate key is invalid. This function is rarely used in practice.
6471 This issue was discovered by Brian Carpenter.
6476 * Removed the export ciphers from the DEFAULT ciphers
6480 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6482 * Build fixes for the Windows and OpenVMS platforms
6484 *Matt Caswell and Richard Levitte*
6486 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6488 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6489 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6490 dereference. This could lead to a Denial Of Service attack. Thanks to
6491 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6496 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6497 dtls1_buffer_record function under certain conditions. In particular this
6498 could occur if an attacker sent repeated DTLS records with the same
6499 sequence number but for the next epoch. The memory leak could be exploited
6500 by an attacker in a Denial of Service attack through memory exhaustion.
6501 Thanks to Chris Mueller for reporting this issue.
6506 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6507 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6508 method would be set to NULL which could later result in a NULL pointer
6509 dereference. Thanks to Frank Schmirler for reporting this issue.
6514 * Abort handshake if server key exchange message is omitted for ephemeral
6517 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6518 reporting this issue.
6523 * Remove non-export ephemeral RSA code on client and server. This code
6524 violated the TLS standard by allowing the use of temporary RSA keys in
6525 non-export ciphersuites and could be used by a server to effectively
6526 downgrade the RSA key length used to a value smaller than the server
6527 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6528 INRIA or reporting this issue.
6533 * Fixed issue where DH client certificates are accepted without verification.
6534 An OpenSSL server will accept a DH certificate for client authentication
6535 without the certificate verify message. This effectively allows a client to
6536 authenticate without the use of a private key. This only affects servers
6537 which trust a client certificate authority which issues certificates
6538 containing DH keys: these are extremely rare and hardly ever encountered.
6539 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6545 * Ensure that the session ID context of an SSL is updated when its
6546 SSL_CTX is updated via SSL_set_SSL_CTX.
6548 The session ID context is typically set from the parent SSL_CTX,
6549 and can vary with the CTX.
6553 * Fix various certificate fingerprint issues.
6555 By using non-DER or invalid encodings outside the signed portion of a
6556 certificate the fingerprint can be changed without breaking the signature.
6557 Although no details of the signed portion of the certificate can be changed
6558 this can cause problems with some applications: e.g. those using the
6559 certificate fingerprint for blacklists.
6561 1. Reject signatures with non zero unused bits.
6563 If the BIT STRING containing the signature has non zero unused bits reject
6564 the signature. All current signature algorithms require zero unused bits.
6566 2. Check certificate algorithm consistency.
6568 Check the AlgorithmIdentifier inside TBS matches the one in the
6569 certificate signature. NB: this will result in signature failure
6570 errors for some broken certificates.
6572 Thanks to Konrad Kraszewski from Google for reporting this issue.
6574 3. Check DSA/ECDSA signatures use DER.
6576 Re-encode DSA/ECDSA signatures and compare with the original received
6577 signature. Return an error if there is a mismatch.
6579 This will reject various cases including garbage after signature
6580 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6581 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6582 (negative or with leading zeroes).
6584 Further analysis was conducted and fixes were developed by Stephen Henson
6585 of the OpenSSL core team.
6591 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6592 results on some platforms, including x86_64. This bug occurs at random
6593 with a very low probability, and is not known to be exploitable in any
6594 way, though its exact impact is difficult to determine. Thanks to Pieter
6595 Wuille (Blockstream) who reported this issue and also suggested an initial
6596 fix. Further analysis was conducted by the OpenSSL development team and
6597 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6598 the OpenSSL core team.
6603 * Do not resume sessions on the server if the negotiated protocol
6604 version does not match the session's version. Resuming with a different
6605 version, while not strictly forbidden by the RFC, is of questionable
6606 sanity and breaks all known clients.
6608 *David Benjamin, Emilia Käsper*
6610 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6611 early CCS messages during renegotiation. (Note that because
6612 renegotiation is encrypted, this early CCS was not exploitable.)
6616 * Tighten client-side session ticket handling during renegotiation:
6617 ensure that the client only accepts a session ticket if the server sends
6618 the extension anew in the ServerHello. Previously, a TLS client would
6619 reuse the old extension state and thus accept a session ticket if one was
6620 announced in the initial ServerHello.
6622 Similarly, ensure that the client requires a session ticket if one
6623 was advertised in the ServerHello. Previously, a TLS client would
6624 ignore a missing NewSessionTicket message.
6628 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6632 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6633 sends a carefully crafted handshake message, to cause OpenSSL to fail
6634 to free up to 64k of memory causing a memory leak. This could be
6635 exploited in a Denial Of Service attack. This issue affects OpenSSL
6636 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6637 whether SRTP is used or configured. Implementations of OpenSSL that
6638 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6640 The fix was developed by the OpenSSL team.
6645 * Session Ticket Memory Leak.
6647 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6648 integrity of that ticket is first verified. In the event of a session
6649 ticket integrity check failing, OpenSSL will fail to free memory
6650 causing a memory leak. By sending a large number of invalid session
6651 tickets an attacker could exploit this issue in a Denial Of Service
6657 * Build option no-ssl3 is incomplete.
6659 When OpenSSL is configured with "no-ssl3" as a build option, servers
6660 could accept and complete a SSL 3.0 handshake, and clients could be
6661 configured to send them.
6664 *Akamai and the OpenSSL team*
6666 * Add support for TLS_FALLBACK_SCSV.
6667 Client applications doing fallback retries should call
6668 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6671 *Adam Langley, Bodo Moeller*
6673 * Add additional DigestInfo checks.
6675 Re-encode DigestInto in DER and check against the original when
6676 verifying RSA signature: this will reject any improperly encoded
6677 DigestInfo structures.
6679 Note: this is a precautionary measure and no attacks are currently known.
6683 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6685 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6686 SRP code can be overrun an internal buffer. Add sanity check that
6687 g, A, B < N to SRP code.
6689 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6690 Group for discovering this issue.
6695 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6696 TLS 1.0 instead of higher protocol versions when the ClientHello message
6697 is badly fragmented. This allows a man-in-the-middle attacker to force a
6698 downgrade to TLS 1.0 even if both the server and the client support a
6699 higher protocol version, by modifying the client's TLS records.
6701 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6702 researching this issue.
6707 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6708 to a denial of service attack. A malicious server can crash the client
6709 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6710 ciphersuite and sending carefully crafted handshake messages.
6712 Thanks to Felix Gröbert (Google) for discovering and researching this
6718 * By sending carefully crafted DTLS packets an attacker could cause openssl
6719 to leak memory. This can be exploited through a Denial of Service attack.
6720 Thanks to Adam Langley for discovering and researching this issue.
6725 * An attacker can force openssl to consume large amounts of memory whilst
6726 processing DTLS handshake messages. This can be exploited through a
6727 Denial of Service attack.
6728 Thanks to Adam Langley for discovering and researching this issue.
6733 * An attacker can force an error condition which causes openssl to crash
6734 whilst processing DTLS packets due to memory being freed twice. This
6735 can be exploited through a Denial of Service attack.
6736 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6742 * If a multithreaded client connects to a malicious server using a resumed
6743 session and the server sends an ec point format extension it could write
6744 up to 255 bytes to freed memory.
6746 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6752 * A malicious server can crash an OpenSSL client with a null pointer
6753 dereference (read) by specifying an SRP ciphersuite even though it was not
6754 properly negotiated with the client. This can be exploited through a
6755 Denial of Service attack.
6757 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6758 discovering and researching this issue.
6763 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6764 X509_name_oneline, X509_name_print_ex et al. to leak some information
6765 from the stack. Applications may be affected if they echo pretty printing
6766 output to the attacker.
6768 Thanks to Ivan Fratric (Google) for discovering this issue.
6771 *Emilia Käsper, and Steve Henson*
6773 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6774 for corner cases. (Certain input points at infinity could lead to
6775 bogus results, with non-infinity inputs mapped to infinity too.)
6779 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6781 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6782 handshake can force the use of weak keying material in OpenSSL
6783 SSL/TLS clients and servers.
6785 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6786 researching this issue. ([CVE-2014-0224])
6788 *KIKUCHI Masashi, Steve Henson*
6790 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6791 OpenSSL DTLS client the code can be made to recurse eventually crashing
6794 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6797 *Imre Rad, Steve Henson*
6799 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6800 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6801 client or server. This is potentially exploitable to run arbitrary
6802 code on a vulnerable client or server.
6804 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6806 *Jüri Aedla, Steve Henson*
6808 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6809 are subject to a denial of service attack.
6811 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6812 this issue. ([CVE-2014-3470])
6814 *Felix Gröbert, Ivan Fratric, Steve Henson*
6816 * Harmonize version and its documentation. -f flag is used to display
6819 *mancha <mancha1@zoho.com>*
6821 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6822 in i2d_ECPrivateKey.
6824 *mancha <mancha1@zoho.com>*
6826 * Fix some double frees. These are not thought to be exploitable.
6828 *mancha <mancha1@zoho.com>*
6830 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6832 * A missing bounds check in the handling of the TLS heartbeat extension
6833 can be used to reveal up to 64k of memory to a connected client or
6836 Thanks for Neel Mehta of Google Security for discovering this bug and to
6837 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6838 preparing the fix ([CVE-2014-0160])
6840 *Adam Langley, Bodo Moeller*
6842 * Fix for the attack described in the paper "Recovering OpenSSL
6843 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6844 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6845 <http://eprint.iacr.org/2014/140>
6847 Thanks to Yuval Yarom and Naomi Benger for discovering this
6848 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6850 *Yuval Yarom and Naomi Benger*
6852 * TLS pad extension: draft-agl-tls-padding-03
6854 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6855 TLS client Hello record length value would otherwise be > 255 and
6856 less that 512 pad with a dummy extension containing zeroes so it
6857 is at least 512 bytes long.
6859 *Adam Langley, Steve Henson*
6861 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6863 * Fix for TLS record tampering bug. A carefully crafted invalid
6864 handshake could crash OpenSSL with a NULL pointer exception.
6865 Thanks to Anton Johansson for reporting this issues.
6868 * Keep original DTLS digest and encryption contexts in retransmission
6869 structures so we can use the previous session parameters if they need
6870 to be resent. ([CVE-2013-6450])
6874 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6875 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6876 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6877 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6878 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6879 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6881 *Rob Stradling, Adam Langley*
6883 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6885 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6886 supporting platforms or when small records were transferred.
6888 *Andy Polyakov, Steve Henson*
6890 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6892 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6894 This addresses the flaw in CBC record processing discovered by
6895 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6896 at: <http://www.isg.rhul.ac.uk/tls/>
6898 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6899 Security Group at Royal Holloway, University of London
6900 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6901 Emilia Käsper for the initial patch.
6904 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6906 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6907 ciphersuites which can be exploited in a denial of service attack.
6908 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6909 and detecting this bug and to Wolfgang Ettlinger
6910 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6915 * Return an error when checking OCSP signatures when key is NULL.
6916 This fixes a DoS attack. ([CVE-2013-0166])
6920 * Make openssl verify return errors.
6922 *Chris Palmer <palmer@google.com> and Ben Laurie*
6924 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6925 the right response is stapled. Also change SSL_get_certificate()
6926 so it returns the certificate actually sent.
6927 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6929 *Rob Stradling <rob.stradling@comodo.com>*
6931 * Fix possible deadlock when decoding public keys.
6935 * Don't use TLS 1.0 record version number in initial client hello
6940 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6942 * Sanity check record length before skipping explicit IV in TLS
6943 1.2, 1.1 and DTLS to fix DoS attack.
6945 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6946 fuzzing as a service testing platform.
6951 * Initialise tkeylen properly when encrypting CMS messages.
6952 Thanks to Solar Designer of Openwall for reporting this issue.
6956 * In FIPS mode don't try to use composite ciphers as they are not
6961 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6963 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6964 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6965 mean any application compiled against OpenSSL 1.0.0 headers setting
6966 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6967 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6968 0x10000000L Any application which was previously compiled against
6969 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6970 will need to be recompiled as a result. Letting be results in
6971 inability to disable specifically TLS 1.1 and in client context,
6972 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6976 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6977 disable just protocol X, but all protocols above X *if* there are
6978 protocols *below* X still enabled. In more practical terms it means
6979 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6980 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6981 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6986 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6988 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6989 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6990 in CRYPTO_realloc_clean.
6992 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6993 issue and to Adam Langley <agl@chromium.org> for fixing it.
6996 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6998 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7002 * Workarounds for some broken servers that "hang" if a client hello
7003 record length exceeds 255 bytes.
7005 1. Do not use record version number > TLS 1.0 in initial client
7006 hello: some (but not all) hanging servers will now work.
7007 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7008 the number of ciphers sent in the client hello. This should be
7009 set to an even number, such as 50, for example by passing:
7010 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7011 Most broken servers should now work.
7012 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7013 TLS 1.2 client support entirely.
7017 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7021 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7023 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7024 STRING form instead of a DigestInfo.
7028 * The format used for MDC2 RSA signatures is inconsistent between EVP
7029 and the RSA_sign/RSA_verify functions. This was made more apparent when
7030 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7031 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7032 the correct format in RSA_verify so both forms transparently work.
7036 * Some servers which support TLS 1.0 can choke if we initially indicate
7037 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7038 encrypted premaster secret. As a workaround use the maximum permitted
7039 client version in client hello, this should keep such servers happy
7040 and still work with previous versions of OpenSSL.
7044 * Add support for TLS/DTLS heartbeats.
7046 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7048 * Add support for SCTP.
7050 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7052 * Improved PRNG seeding for VOS.
7054 *Paul Green <Paul.Green@stratus.com>*
7056 * Extensive assembler packs updates, most notably:
7058 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7059 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7060 - x86_64: bit-sliced AES implementation;
7061 - ARM: NEON support, contemporary platforms optimizations;
7062 - s390x: z196 support;
7063 - `*`: GHASH and GF(2^m) multiplication implementations;
7067 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7068 (removal of unnecessary code)
7070 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7072 * Add TLS key material exporter from RFC 5705.
7076 * Add DTLS-SRTP negotiation from RFC 5764.
7080 * Add Next Protocol Negotiation,
7081 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7082 disabled with a no-npn flag to config or Configure. Code donated
7085 *Adam Langley <agl@google.com> and Ben Laurie*
7087 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7088 NIST-P256, NIST-P521, with constant-time single point multiplication on
7089 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7090 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7091 Code made available under Apache License version 2.0.
7093 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7094 line to include this in your build of OpenSSL, and run "make depend" (or
7095 "make update"). This enables the following EC_METHODs:
7097 EC_GFp_nistp224_method()
7098 EC_GFp_nistp256_method()
7099 EC_GFp_nistp521_method()
7101 EC_GROUP_new_by_curve_name() will automatically use these (while
7102 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7105 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7107 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7108 all platforms. Move ssize_t definition from e_os.h to the public
7109 header file e_os2.h as it now appears in public header file cms.h
7113 * New -sigopt option to the ca, req and x509 utilities. Additional
7114 signature parameters can be passed using this option and in
7119 * Add RSA PSS signing function. This will generate and set the
7120 appropriate AlgorithmIdentifiers for PSS based on those in the
7121 corresponding EVP_MD_CTX structure. No application support yet.
7125 * Support for companion algorithm specific ASN1 signing routines.
7126 New function ASN1_item_sign_ctx() signs a pre-initialised
7127 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7128 the appropriate parameters.
7132 * Add new algorithm specific ASN1 verification initialisation function
7133 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7134 handling will be the same no matter what EVP_PKEY_METHOD is used.
7135 Add a PSS handler to support verification of PSS signatures: checked
7136 against a number of sample certificates.
7140 * Add signature printing for PSS. Add PSS OIDs.
7142 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7144 * Add algorithm specific signature printing. An individual ASN1 method
7145 can now print out signatures instead of the standard hex dump.
7147 More complex signatures (e.g. PSS) can print out more meaningful
7148 information. Include DSA version that prints out the signature
7153 * Password based recipient info support for CMS library: implementing
7158 * Split password based encryption into PBES2 and PBKDF2 functions. This
7159 neatly separates the code into cipher and PBE sections and is required
7160 for some algorithms that split PBES2 into separate pieces (such as
7161 password based CMS).
7165 * Session-handling fixes:
7166 - Fix handling of connections that are resuming with a session ID,
7167 but also support Session Tickets.
7168 - Fix a bug that suppressed issuing of a new ticket if the client
7169 presented a ticket with an expired session.
7170 - Try to set the ticket lifetime hint to something reasonable.
7171 - Make tickets shorter by excluding irrelevant information.
7172 - On the client side, don't ignore renewed tickets.
7174 *Adam Langley, Bodo Moeller (Google)*
7176 * Fix PSK session representation.
7180 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7182 This work was sponsored by Intel.
7186 * Add GCM support to TLS library. Some custom code is needed to split
7187 the IV between the fixed (from PRF) and explicit (from TLS record)
7188 portions. This adds all GCM ciphersuites supported by RFC5288 and
7189 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7190 add a special AESGCM string for GCM only.
7194 * Expand range of ctrls for AES GCM. Permit setting invocation
7195 field on decrypt and retrieval of invocation field only on encrypt.
7199 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7200 As required by RFC5289 these ciphersuites cannot be used if for
7201 versions of TLS earlier than 1.2.
7205 * For FIPS capable OpenSSL interpret a NULL default public key method
7206 as unset and return the appropriate default but do *not* set the default.
7207 This means we can return the appropriate method in applications that
7208 switch between FIPS and non-FIPS modes.
7212 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7213 ENGINE is used then we cannot handle that in the FIPS module so we
7214 keep original code iff non-FIPS operations are allowed.
7218 * Add -attime option to openssl utilities.
7220 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7222 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7226 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7227 FIPS EC methods unconditionally for now.
7231 * New build option no-ec2m to disable characteristic 2 code.
7235 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7236 all cases can be covered as some introduce binary incompatibilities.
7240 * Redirect RSA operations to FIPS module including keygen,
7241 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7245 * Add similar low level API blocking to ciphers.
7249 * Low level digest APIs are not approved in FIPS mode: any attempt
7250 to use these will cause a fatal error. Applications that *really* want
7251 to use them can use the `private_*` version instead.
7255 * Redirect cipher operations to FIPS module for FIPS builds.
7259 * Redirect digest operations to FIPS module for FIPS builds.
7263 * Update build system to add "fips" flag which will link in fipscanister.o
7264 for static and shared library builds embedding a signature if needed.
7268 * Output TLS supported curves in preference order instead of numerical
7269 order. This is currently hardcoded for the highest order curves first.
7270 This should be configurable so applications can judge speed vs strength.
7274 * Add TLS v1.2 server support for client authentication.
7278 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7283 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7284 FIPS modules versions.
7288 * Add TLS v1.2 client side support for client authentication. Keep cache
7289 of handshake records longer as we don't know the hash algorithm to use
7290 until after the certificate request message is received.
7294 * Initial TLS v1.2 client support. Add a default signature algorithms
7295 extension including all the algorithms we support. Parse new signature
7296 format in client key exchange. Relax some ECC signing restrictions for
7297 TLS v1.2 as indicated in RFC5246.
7301 * Add server support for TLS v1.2 signature algorithms extension. Switch
7302 to new signature format when needed using client digest preference.
7303 All server ciphersuites should now work correctly in TLS v1.2. No client
7304 support yet and no support for client certificates.
7308 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7309 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7310 ciphersuites. At present only RSA key exchange ciphersuites work with
7311 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7312 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7313 and version checking.
7317 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7318 with this defined it will not be affected by any changes to ssl internal
7319 structures. Add several utility functions to allow openssl application
7320 to work with OPENSSL_NO_SSL_INTERN defined.
7324 * A long standing patch to add support for SRP from EdelWeb (Peter
7325 Sylvester and Christophe Renou) was integrated.
7326 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7327 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7330 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7334 * Permit abbreviated handshakes when renegotiating using the function
7335 SSL_renegotiate_abbreviated().
7337 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7339 * Add call to ENGINE_register_all_complete() to
7340 ENGINE_load_builtin_engines(), so some implementations get used
7341 automatically instead of needing explicit application support.
7345 * Add support for TLS key exporter as described in RFC5705.
7347 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7349 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7350 a few changes are required:
7352 Add SSL_OP_NO_TLSv1_1 flag.
7353 Add TLSv1_1 methods.
7354 Update version checking logic to handle version 1.1.
7355 Add explicit IV handling (ported from DTLS code).
7356 Add command line options to s_client/s_server.
7363 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7365 * X509_ATTRIBUTE memory leak
7367 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7368 memory. This structure is used by the PKCS#7 and CMS routines so any
7369 application which reads PKCS#7 or CMS data from untrusted sources is
7370 affected. SSL/TLS is not affected.
7372 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7378 * Race condition handling PSK identify hint
7380 If PSK identity hints are received by a multi-threaded client then
7381 the values are wrongly updated in the parent SSL_CTX structure. This can
7382 result in a race condition potentially leading to a double free of the
7388 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7390 * Malformed ECParameters causes infinite loop
7392 When processing an ECParameters structure OpenSSL enters an infinite loop
7393 if the curve specified is over a specially malformed binary polynomial
7396 This can be used to perform denial of service against any
7397 system which processes public keys, certificate requests or
7398 certificates. This includes TLS clients and TLS servers with
7399 client authentication enabled.
7401 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7406 * Exploitable out-of-bounds read in X509_cmp_time
7408 X509_cmp_time does not properly check the length of the ASN1_TIME
7409 string and can read a few bytes out of bounds. In addition,
7410 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7413 An attacker can use this to craft malformed certificates and CRLs of
7414 various sizes and potentially cause a segmentation fault, resulting in
7415 a DoS on applications that verify certificates or CRLs. TLS clients
7416 that verify CRLs are affected. TLS clients and servers with client
7417 authentication enabled may be affected if they use custom verification
7420 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7421 independently by Hanno Böck.
7426 * PKCS7 crash with missing EnvelopedContent
7428 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7429 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7430 with missing content and trigger a NULL pointer dereference on parsing.
7432 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7433 structures from untrusted sources are affected. OpenSSL clients and
7434 servers are not affected.
7436 This issue was reported to OpenSSL by Michal Zalewski (Google).
7441 * CMS verify infinite loop with unknown hash function
7443 When verifying a signedData message the CMS code can enter an infinite loop
7444 if presented with an unknown hash function OID. This can be used to perform
7445 denial of service against any system which verifies signedData messages using
7447 This issue was reported to OpenSSL by Johannes Bauer.
7452 * Race condition handling NewSessionTicket
7454 If a NewSessionTicket is received by a multi-threaded client when attempting to
7455 reuse a previous ticket then a race condition can occur potentially leading to
7456 a double free of the ticket data.
7461 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7463 * Segmentation fault in ASN1_TYPE_cmp fix
7465 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7466 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7467 certificate signature algorithm consistency this can be used to crash any
7468 certificate verification operation and exploited in a DoS attack. Any
7469 application which performs certificate verification is vulnerable including
7470 OpenSSL clients and servers which enable client authentication.
7475 * ASN.1 structure reuse memory corruption fix
7477 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7478 memory corruption via an invalid write. Such reuse is and has been
7479 strongly discouraged and is believed to be rare.
7481 Applications that parse structures containing CHOICE or ANY DEFINED BY
7482 components may be affected. Certificate parsing (d2i_X509 and related
7483 functions) are however not affected. OpenSSL clients and servers are
7489 * PKCS7 NULL pointer dereferences fix
7491 The PKCS#7 parsing code does not handle missing outer ContentInfo
7492 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7493 missing content and trigger a NULL pointer dereference on parsing.
7495 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7496 otherwise parse PKCS#7 structures from untrusted sources are
7497 affected. OpenSSL clients and servers are not affected.
7499 This issue was reported to OpenSSL by Michal Zalewski (Google).
7504 * DoS via reachable assert in SSLv2 servers fix
7506 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7507 servers that both support SSLv2 and enable export cipher suites by sending
7508 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7510 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7511 (OpenSSL development team).
7516 * Use After Free following d2i_ECPrivatekey error fix
7518 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7519 could cause a use after free condition. This, in turn, could cause a double
7520 free in several private key parsing functions (such as d2i_PrivateKey
7521 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7522 for applications that receive EC private keys from untrusted
7523 sources. This scenario is considered rare.
7525 This issue was discovered by the BoringSSL project and fixed in their
7531 * X509_to_X509_REQ NULL pointer deref fix
7533 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7534 the certificate key is invalid. This function is rarely used in practice.
7536 This issue was discovered by Brian Carpenter.
7541 * Removed the export ciphers from the DEFAULT ciphers
7545 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7547 * Build fixes for the Windows and OpenVMS platforms
7549 *Matt Caswell and Richard Levitte*
7551 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7553 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7554 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7555 dereference. This could lead to a Denial Of Service attack. Thanks to
7556 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7561 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7562 dtls1_buffer_record function under certain conditions. In particular this
7563 could occur if an attacker sent repeated DTLS records with the same
7564 sequence number but for the next epoch. The memory leak could be exploited
7565 by an attacker in a Denial of Service attack through memory exhaustion.
7566 Thanks to Chris Mueller for reporting this issue.
7571 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7572 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7573 method would be set to NULL which could later result in a NULL pointer
7574 dereference. Thanks to Frank Schmirler for reporting this issue.
7579 * Abort handshake if server key exchange message is omitted for ephemeral
7582 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7583 reporting this issue.
7588 * Remove non-export ephemeral RSA code on client and server. This code
7589 violated the TLS standard by allowing the use of temporary RSA keys in
7590 non-export ciphersuites and could be used by a server to effectively
7591 downgrade the RSA key length used to a value smaller than the server
7592 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7593 INRIA or reporting this issue.
7598 * Fixed issue where DH client certificates are accepted without verification.
7599 An OpenSSL server will accept a DH certificate for client authentication
7600 without the certificate verify message. This effectively allows a client to
7601 authenticate without the use of a private key. This only affects servers
7602 which trust a client certificate authority which issues certificates
7603 containing DH keys: these are extremely rare and hardly ever encountered.
7604 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7610 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7611 results on some platforms, including x86_64. This bug occurs at random
7612 with a very low probability, and is not known to be exploitable in any
7613 way, though its exact impact is difficult to determine. Thanks to Pieter
7614 Wuille (Blockstream) who reported this issue and also suggested an initial
7615 fix. Further analysis was conducted by the OpenSSL development team and
7616 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7617 the OpenSSL core team.
7622 * Fix various certificate fingerprint issues.
7624 By using non-DER or invalid encodings outside the signed portion of a
7625 certificate the fingerprint can be changed without breaking the signature.
7626 Although no details of the signed portion of the certificate can be changed
7627 this can cause problems with some applications: e.g. those using the
7628 certificate fingerprint for blacklists.
7630 1. Reject signatures with non zero unused bits.
7632 If the BIT STRING containing the signature has non zero unused bits reject
7633 the signature. All current signature algorithms require zero unused bits.
7635 2. Check certificate algorithm consistency.
7637 Check the AlgorithmIdentifier inside TBS matches the one in the
7638 certificate signature. NB: this will result in signature failure
7639 errors for some broken certificates.
7641 Thanks to Konrad Kraszewski from Google for reporting this issue.
7643 3. Check DSA/ECDSA signatures use DER.
7645 Reencode DSA/ECDSA signatures and compare with the original received
7646 signature. Return an error if there is a mismatch.
7648 This will reject various cases including garbage after signature
7649 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7650 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7651 (negative or with leading zeroes).
7653 Further analysis was conducted and fixes were developed by Stephen Henson
7654 of the OpenSSL core team.
7660 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7662 * Session Ticket Memory Leak.
7664 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7665 integrity of that ticket is first verified. In the event of a session
7666 ticket integrity check failing, OpenSSL will fail to free memory
7667 causing a memory leak. By sending a large number of invalid session
7668 tickets an attacker could exploit this issue in a Denial Of Service
7674 * Build option no-ssl3 is incomplete.
7676 When OpenSSL is configured with "no-ssl3" as a build option, servers
7677 could accept and complete a SSL 3.0 handshake, and clients could be
7678 configured to send them.
7681 *Akamai and the OpenSSL team*
7683 * Add support for TLS_FALLBACK_SCSV.
7684 Client applications doing fallback retries should call
7685 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7688 *Adam Langley, Bodo Moeller*
7690 * Add additional DigestInfo checks.
7692 Reencode DigestInto in DER and check against the original when
7693 verifying RSA signature: this will reject any improperly encoded
7694 DigestInfo structures.
7696 Note: this is a precautionary measure and no attacks are currently known.
7700 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7702 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7703 to a denial of service attack. A malicious server can crash the client
7704 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7705 ciphersuite and sending carefully crafted handshake messages.
7707 Thanks to Felix Gröbert (Google) for discovering and researching this
7713 * By sending carefully crafted DTLS packets an attacker could cause openssl
7714 to leak memory. This can be exploited through a Denial of Service attack.
7715 Thanks to Adam Langley for discovering and researching this issue.
7720 * An attacker can force openssl to consume large amounts of memory whilst
7721 processing DTLS handshake messages. This can be exploited through a
7722 Denial of Service attack.
7723 Thanks to Adam Langley for discovering and researching this issue.
7728 * An attacker can force an error condition which causes openssl to crash
7729 whilst processing DTLS packets due to memory being freed twice. This
7730 can be exploited through a Denial of Service attack.
7731 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7737 * If a multithreaded client connects to a malicious server using a resumed
7738 session and the server sends an ec point format extension it could write
7739 up to 255 bytes to freed memory.
7741 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7747 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7748 X509_name_oneline, X509_name_print_ex et al. to leak some information
7749 from the stack. Applications may be affected if they echo pretty printing
7750 output to the attacker.
7752 Thanks to Ivan Fratric (Google) for discovering this issue.
7755 *Emilia Käsper, and Steve Henson*
7757 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7758 for corner cases. (Certain input points at infinity could lead to
7759 bogus results, with non-infinity inputs mapped to infinity too.)
7763 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7765 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7766 handshake can force the use of weak keying material in OpenSSL
7767 SSL/TLS clients and servers.
7769 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7770 researching this issue. ([CVE-2014-0224])
7772 *KIKUCHI Masashi, Steve Henson*
7774 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7775 OpenSSL DTLS client the code can be made to recurse eventually crashing
7778 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7781 *Imre Rad, Steve Henson*
7783 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7784 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7785 client or server. This is potentially exploitable to run arbitrary
7786 code on a vulnerable client or server.
7788 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7790 *Jüri Aedla, Steve Henson*
7792 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7793 are subject to a denial of service attack.
7795 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7796 this issue. ([CVE-2014-3470])
7798 *Felix Gröbert, Ivan Fratric, Steve Henson*
7800 * Harmonize version and its documentation. -f flag is used to display
7803 *mancha <mancha1@zoho.com>*
7805 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7806 in i2d_ECPrivateKey.
7808 *mancha <mancha1@zoho.com>*
7810 * Fix some double frees. These are not thought to be exploitable.
7812 *mancha <mancha1@zoho.com>*
7814 * Fix for the attack described in the paper "Recovering OpenSSL
7815 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7816 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7817 <http://eprint.iacr.org/2014/140>
7819 Thanks to Yuval Yarom and Naomi Benger for discovering this
7820 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7822 *Yuval Yarom and Naomi Benger*
7824 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7826 * Keep original DTLS digest and encryption contexts in retransmission
7827 structures so we can use the previous session parameters if they need
7828 to be resent. ([CVE-2013-6450])
7832 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7833 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7834 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7835 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7836 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7837 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7839 *Rob Stradling, Adam Langley*
7841 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7843 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7845 This addresses the flaw in CBC record processing discovered by
7846 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7847 at: <http://www.isg.rhul.ac.uk/tls/>
7849 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7850 Security Group at Royal Holloway, University of London
7851 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7852 Emilia Käsper for the initial patch.
7855 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7857 * Return an error when checking OCSP signatures when key is NULL.
7858 This fixes a DoS attack. ([CVE-2013-0166])
7862 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7863 the right response is stapled. Also change SSL_get_certificate()
7864 so it returns the certificate actually sent.
7865 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7866 (This is a backport)
7868 *Rob Stradling <rob.stradling@comodo.com>*
7870 * Fix possible deadlock when decoding public keys.
7874 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7876 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7879 * Sanity check record length before skipping explicit IV in DTLS
7882 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7883 fuzzing as a service testing platform.
7888 * Initialise tkeylen properly when encrypting CMS messages.
7889 Thanks to Solar Designer of Openwall for reporting this issue.
7893 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7895 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7896 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7897 in CRYPTO_realloc_clean.
7899 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7900 issue and to Adam Langley <agl@chromium.org> for fixing it.
7903 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7905 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7907 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7908 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7909 content decryption and always return the same error. Note: this attack
7910 needs on average 2^20 messages so it only affects automated senders. The
7911 old behaviour can be re-enabled in the CMS code by setting the
7912 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7913 an MMA defence is not necessary.
7914 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7915 this issue. ([CVE-2012-0884])
7919 * Fix CVE-2011-4619: make sure we really are receiving a
7920 client hello before rejecting multiple SGC restarts. Thanks to
7921 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7925 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7927 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7928 Thanks to Antonio Martin, Enterprise Secure Access Research and
7929 Development, Cisco Systems, Inc. for discovering this bug and
7930 preparing a fix. ([CVE-2012-0050])
7934 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7936 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7937 of the Vaudenay padding oracle attack on CBC mode encryption
7938 which enables an efficient plaintext recovery attack against
7939 the OpenSSL implementation of DTLS. Their attack exploits timing
7940 differences arising during decryption processing. A research
7941 paper describing this attack can be found at:
7942 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7943 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7944 Security Group at Royal Holloway, University of London
7945 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7946 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7947 for preparing the fix. ([CVE-2011-4108])
7949 *Robin Seggelmann, Michael Tuexen*
7951 * Clear bytes used for block padding of SSL 3.0 records.
7954 *Adam Langley (Google)*
7956 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7957 Kadianakis <desnacked@gmail.com> for discovering this issue and
7958 Adam Langley for preparing the fix. ([CVE-2011-4619])
7960 *Adam Langley (Google)*
7962 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7964 *Andrey Kulikov <amdeich@gmail.com>*
7966 * Prevent malformed RFC3779 data triggering an assertion failure.
7967 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7968 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7970 *Rob Austein <sra@hactrn.net>*
7972 * Improved PRNG seeding for VOS.
7974 *Paul Green <Paul.Green@stratus.com>*
7976 * Fix ssl_ciph.c set-up race.
7978 *Adam Langley (Google)*
7980 * Fix spurious failures in ecdsatest.c.
7982 *Emilia Käsper (Google)*
7984 * Fix the BIO_f_buffer() implementation (which was mixing different
7985 interpretations of the `..._len` fields).
7987 *Adam Langley (Google)*
7989 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7990 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7991 threads won't reuse the same blinding coefficients.
7993 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7994 lock to call BN_BLINDING_invert_ex, and avoids one use of
7995 BN_BLINDING_update for each BN_BLINDING structure (previously,
7996 the last update always remained unused).
7998 *Emilia Käsper (Google)*
8000 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8002 *Bob Buckholz (Google)*
8004 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8006 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8007 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8009 *Kaspar Brand <ossl@velox.ch>*
8011 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8012 for multi-threaded use of ECDH. ([CVE-2011-3210])
8014 *Adam Langley (Google)*
8016 * Fix x509_name_ex_d2i memory leak on bad inputs.
8020 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8021 signature public key algorithm by using OID xref utilities instead.
8022 Before this you could only use some ECC ciphersuites with SHA1 only.
8026 * Add protection against ECDSA timing attacks as mentioned in the paper
8027 by Billy Bob Brumley and Nicola Tuveri, see:
8028 <http://eprint.iacr.org/2011/232.pdf>
8030 *Billy Bob Brumley and Nicola Tuveri*
8032 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8034 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8036 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8038 * Fix bug in string printing code: if *any* escaping is enabled we must
8039 escape the escape character (backslash) or the resulting string is
8044 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8046 * Disable code workaround for ancient and obsolete Netscape browsers
8047 and servers: an attacker can use it in a ciphersuite downgrade attack.
8048 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8052 * Fixed J-PAKE implementation error, originally discovered by
8053 Sebastien Martini, further info and confirmation from Stefan
8054 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8058 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8060 * Fix extension code to avoid race conditions which can result in a buffer
8061 overrun vulnerability: resumed sessions must not be modified as they can
8062 be shared by multiple threads. CVE-2010-3864
8066 * Fix WIN32 build system to correctly link an ENGINE directory into
8071 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8073 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8076 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8078 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8080 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8081 context. The operation can be customised via the ctrl mechanism in
8082 case ENGINEs want to include additional functionality.
8086 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8090 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8091 output hashes compatible with older versions of OpenSSL.
8093 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8095 * Fix compression algorithm handling: if resuming a session use the
8096 compression algorithm of the resumed session instead of determining
8097 it from client hello again. Don't allow server to change algorithm.
8101 * Add load_crls() function to commands tidying load_certs() too. Add option
8102 to verify utility to allow additional CRLs to be included.
8106 * Update OCSP request code to permit adding custom headers to the request:
8107 some responders need this.
8111 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8114 *Julia Lawall <julia@diku.dk>*
8116 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8117 needlessly dereferenced structures, used obsolete functions and
8118 didn't handle all updated verify codes correctly.
8122 * Disable MD2 in the default configuration.
8126 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8127 indicate the initial BIO being pushed or popped. This makes it possible
8128 to determine whether the BIO is the one explicitly called or as a result
8129 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8130 it handles reference counts correctly and doesn't zero out the I/O bio
8131 when it is not being explicitly popped. WARNING: applications which
8132 included workarounds for the old buggy behaviour will need to be modified
8133 or they could free up already freed BIOs.
8137 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8138 renaming to all platforms (within the 0.9.8 branch, this was
8139 done conditionally on Netware platforms to avoid a name clash).
8141 *Guenter <lists@gknw.net>*
8143 * Add ECDHE and PSK support to DTLS.
8145 *Michael Tuexen <tuexen@fh-muenster.de>*
8147 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8152 * Add "missing" function EVP_MD_flags() (without this the only way to
8153 retrieve a digest flags is by accessing the structure directly. Update
8154 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8155 or cipher is registered as in the "from" argument. Print out all
8156 registered digests in the dgst usage message instead of manually
8157 attempting to work them out.
8161 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8162 this allows the use of compression and extensions. Change default cipher
8163 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8164 by default unless an application cipher string requests it.
8168 * Alter match criteria in PKCS12_parse(). It used to try to use local
8169 key ids to find matching certificates and keys but some PKCS#12 files
8170 don't follow the (somewhat unwritten) rules and this strategy fails.
8171 Now just gather all certificates together and the first private key
8172 then look for the first certificate that matches the key.
8176 * Support use of registered digest and cipher names for dgst and cipher
8177 commands instead of having to add each one as a special case. So now
8184 openssl dgst -sha256 foo
8186 and this works for ENGINE based algorithms too.
8190 * Update Gost ENGINE to support parameter files.
8192 *Victor B. Wagner <vitus@cryptocom.ru>*
8194 * Support GeneralizedTime in ca utility.
8196 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8198 * Enhance the hash format used for certificate directory links. The new
8199 form uses the canonical encoding (meaning equivalent names will work
8200 even if they aren't identical) and uses SHA1 instead of MD5. This form
8201 is incompatible with the older format and as a result c_rehash should
8202 be used to rebuild symbolic links.
8206 * Make PKCS#8 the default write format for private keys, replacing the
8207 traditional format. This form is standardised, more secure and doesn't
8208 include an implicit MD5 dependency.
8212 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8213 committed to OpenSSL should pass this lot as a minimum.
8217 * Add session ticket override functionality for use by EAP-FAST.
8219 *Jouni Malinen <j@w1.fi>*
8221 * Modify HMAC functions to return a value. Since these can be implemented
8222 in an ENGINE errors can occur.
8226 * Type-checked OBJ_bsearch_ex.
8230 * Type-checked OBJ_bsearch. Also some constification necessitated
8231 by type-checking. Still to come: TXT_DB, bsearch(?),
8232 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8237 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8238 seconds to a tm structure directly, instead of going through OS
8239 specific date routines. This avoids any issues with OS routines such
8240 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8241 and X509_time_adj_ex() to cover the extended range. The existing
8242 X509_time_adj() is still usable and will no longer have any date issues.
8246 * Delta CRL support. New use deltas option which will attempt to locate
8247 and search any appropriate delta CRLs available.
8249 This work was sponsored by Google.
8253 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8254 code and add additional score elements. Validate alternate CRL paths
8255 as part of the CRL checking and indicate a new error "CRL path validation
8256 error" in this case. Applications wanting additional details can use
8257 the verify callback and check the new "parent" field. If this is not
8258 NULL CRL path validation is taking place. Existing applications won't
8259 see this because it requires extended CRL support which is off by
8262 This work was sponsored by Google.
8266 * Support for freshest CRL extension.
8268 This work was sponsored by Google.
8272 * Initial indirect CRL support. Currently only supported in the CRLs
8273 passed directly and not via lookup. Process certificate issuer
8274 CRL entry extension and lookup CRL entries by bother issuer name
8275 and serial number. Check and process CRL issuer entry in IDP extension.
8277 This work was sponsored by Google.
8281 * Add support for distinct certificate and CRL paths. The CRL issuer
8282 certificate is validated separately in this case. Only enabled if
8283 an extended CRL support flag is set: this flag will enable additional
8284 CRL functionality in future.
8286 This work was sponsored by Google.
8290 * Add support for policy mappings extension.
8292 This work was sponsored by Google.
8296 * Fixes to pathlength constraint, self issued certificate handling,
8297 policy processing to align with RFC3280 and PKITS tests.
8299 This work was sponsored by Google.
8303 * Support for name constraints certificate extension. DN, email, DNS
8304 and URI types are currently supported.
8306 This work was sponsored by Google.
8310 * To cater for systems that provide a pointer-based thread ID rather
8311 than numeric, deprecate the current numeric thread ID mechanism and
8312 replace it with a structure and associated callback type. This
8313 mechanism allows a numeric "hash" to be extracted from a thread ID in
8314 either case, and on platforms where pointers are larger than 'long',
8315 mixing is done to help ensure the numeric 'hash' is usable even if it
8316 can't be guaranteed unique. The default mechanism is to use "&errno"
8317 as a pointer-based thread ID to distinguish between threads.
8319 Applications that want to provide their own thread IDs should now use
8320 CRYPTO_THREADID_set_callback() to register a callback that will call
8321 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8323 Note that ERR_remove_state() is now deprecated, because it is tied
8324 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8325 to free the current thread's error state should be replaced by
8326 ERR_remove_thread_state(NULL).
8328 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8329 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8330 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8331 application was previously providing a numeric thread callback that
8332 was inappropriate for distinguishing threads, then uniqueness might
8333 have been obtained with &errno that happened immediately in the
8334 intermediate development versions of OpenSSL; this is no longer the
8335 case, the numeric thread callback will now override the automatic use
8338 *Geoff Thorpe, with help from Bodo Moeller*
8340 * Initial support for different CRL issuing certificates. This covers a
8341 simple case where the self issued certificates in the chain exist and
8342 the real CRL issuer is higher in the existing chain.
8344 This work was sponsored by Google.
8348 * Removed effectively defunct crypto/store from the build.
8352 * Revamp of STACK to provide stronger type-checking. Still to come:
8353 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8354 ASN1_STRING, CONF_VALUE.
8358 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8359 RAM on SSL connections. This option can save about 34k per idle SSL.
8363 * Revamp of LHASH to provide stronger type-checking. Still to come:
8364 STACK, TXT_DB, bsearch, qsort.
8368 * Initial support for Cryptographic Message Syntax (aka CMS) based
8369 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8370 support for data, signedData, compressedData, digestedData and
8371 encryptedData, envelopedData types included. Scripts to check against
8372 RFC4134 examples draft and interop and consistency checks of many
8373 content types and variants.
8377 * Add options to enc utility to support use of zlib compression BIO.
8381 * Extend mk1mf to support importing of options and assembly language
8382 files from Configure script, currently only included in VC-WIN32.
8383 The assembly language rules can now optionally generate the source
8384 files from the associated perl scripts.
8388 * Implement remaining functionality needed to support GOST ciphersuites.
8389 Interop testing has been performed using CryptoPro implementations.
8391 *Victor B. Wagner <vitus@cryptocom.ru>*
8393 * s390x assembler pack.
8397 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8402 * Implement Opaque PRF Input TLS extension as specified in
8403 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8404 official specification yet and no extension type assignment by
8405 IANA exists, this extension (for now) will have to be explicitly
8406 enabled when building OpenSSL by providing the extension number
8407 to use. For example, specify an option
8409 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8411 to the "config" or "Configure" script to enable the extension,
8412 assuming extension number 0x9527 (which is a completely arbitrary
8413 and unofficial assignment based on the MD5 hash of the Internet
8414 Draft). Note that by doing so, you potentially lose
8415 interoperability with other TLS implementations since these might
8416 be using the same extension number for other purposes.
8418 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8419 opaque PRF input value to use in the handshake. This will create
8420 an internal copy of the length-'len' string at 'src', and will
8421 return non-zero for success.
8423 To get more control and flexibility, provide a callback function
8426 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8427 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8431 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8434 Callback function 'cb' will be called in handshakes, and is
8435 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8436 Argument 'arg' is for application purposes (the value as given to
8437 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8438 be provided to the callback function). The callback function
8439 has to return non-zero to report success: usually 1 to use opaque
8440 PRF input just if possible, or 2 to enforce use of the opaque PRF
8441 input. In the latter case, the library will abort the handshake
8442 if opaque PRF input is not successfully negotiated.
8444 Arguments 'peerinput' and 'len' given to the callback function
8445 will always be NULL and 0 in the case of a client. A server will
8446 see the client's opaque PRF input through these variables if
8447 available (NULL and 0 otherwise). Note that if the server
8448 provides an opaque PRF input, the length must be the same as the
8449 length of the client's opaque PRF input.
8451 Note that the callback function will only be called when creating
8452 a new session (session resumption can resume whatever was
8453 previously negotiated), and will not be called in SSL 2.0
8454 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8455 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8456 for applications that need to enforce opaque PRF input.
8460 * Update ssl code to support digests other than SHA1+MD5 for handshake
8463 *Victor B. Wagner <vitus@cryptocom.ru>*
8465 * Add RFC4507 support to OpenSSL. This includes the corrections in
8466 RFC4507bis. The encrypted ticket format is an encrypted encoded
8467 SSL_SESSION structure, that way new session features are automatically
8470 If a client application caches session in an SSL_SESSION structure
8471 support is transparent because tickets are now stored in the encoded
8474 The SSL_CTX structure automatically generates keys for ticket
8475 protection in servers so again support should be possible
8476 with no application modification.
8478 If a client or server wishes to disable RFC4507 support then the option
8479 SSL_OP_NO_TICKET can be set.
8481 Add a TLS extension debugging callback to allow the contents of any client
8482 or server extensions to be examined.
8484 This work was sponsored by Google.
8488 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8489 OpenSSL should now compile cleanly on gcc 4.2
8491 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8493 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8494 support including streaming MAC support: this is required for GOST
8495 ciphersuite support.
8497 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8499 * Add option -stream to use PKCS#7 streaming in smime utility. New
8500 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8501 to output in BER and PEM format.
8505 * Experimental support for use of HMAC via EVP_PKEY interface. This
8506 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8507 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8508 ENGINE support for HMAC keys which are unextractable. New -mac and
8509 -macopt options to dgst utility.
8513 * New option -sigopt to dgst utility. Update dgst to use
8514 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8515 alternative signing parameters such as X9.31 or PSS in the dgst
8520 * Change ssl_cipher_apply_rule(), the internal function that does
8521 the work each time a ciphersuite string requests enabling
8522 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8523 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8524 the order of disabled ciphersuites such that those ciphersuites
8525 that most recently went from enabled to disabled not only stay
8526 in order with respect to each other, but also have higher priority
8527 than other disabled ciphersuites the next time ciphersuites are
8530 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8531 the same ciphersuites as with "HIGH" alone, but in a specific
8532 order where the PSK ciphersuites come first (since they are the
8533 most recently disabled ciphersuites when "HIGH" is parsed).
8535 Also, change ssl_create_cipher_list() (using this new
8536 functionality) such that between otherwise identical
8537 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8542 * Change ssl_create_cipher_list() so that it automatically
8543 arranges the ciphersuites in reasonable order before starting
8544 to process the rule string. Thus, the definition for "DEFAULT"
8545 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8546 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8547 This makes it much easier to arrive at a reasonable default order
8548 in applications for which anonymous ciphers are OK (meaning
8549 that you can't actually use DEFAULT).
8551 *Bodo Moeller; suggested by Victor Duchovni*
8553 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8554 processing) into multiple integers instead of setting
8555 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8556 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8557 (These masks as well as the individual bit definitions are hidden
8558 away into the non-exported interface ssl/ssl_locl.h, so this
8559 change to the definition of the SSL_CIPHER structure shouldn't
8560 affect applications.) This give us more bits for each of these
8561 categories, so there is no longer a need to coagulate AES128 and
8562 AES256 into a single algorithm bit, and to coagulate Camellia128
8563 and Camellia256 into a single algorithm bit, which has led to all
8566 Thus, among other things, the kludge introduced in 0.9.7m and
8567 0.9.8e for masking out AES256 independently of AES128 or masking
8568 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8570 With the change, we also introduce new ciphersuite aliases that
8571 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8576 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8577 Use the leftmost N bytes of the signature input if the input is
8578 larger than the prime q (with N being the size in bytes of q).
8582 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8583 it yet and it is largely untested.
8587 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8591 * Initial incomplete changes to avoid need for function casts in OpenSSL
8592 some compilers (gcc 4.2 and later) reject their use. Safestack is
8593 reimplemented. Update ASN1 to avoid use of legacy functions.
8597 * Win32/64 targets are linked with Winsock2.
8601 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8602 to external functions. This can be used to increase CRL handling
8603 efficiency especially when CRLs are very large by (for example) storing
8604 the CRL revoked certificates in a database.
8608 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8609 new CRLs added to a directory can be used. New command line option
8610 -verify_return_error to s_client and s_server. This causes real errors
8611 to be returned by the verify callback instead of carrying on no matter
8612 what. This reflects the way a "real world" verify callback would behave.
8616 * GOST engine, supporting several GOST algorithms and public key formats.
8617 Kindly donated by Cryptocom.
8621 * Partial support for Issuing Distribution Point CRL extension. CRLs
8622 partitioned by DP are handled but no indirect CRL or reason partitioning
8623 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8624 selected via a scoring technique which handles IDP and AKID in CRLs.
8628 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8629 will ultimately be used for all verify operations: this will remove the
8630 X509_STORE dependency on certificate verification and allow alternative
8631 lookup methods. X509_STORE based implementations of these two callbacks.
8635 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8636 Modify get_crl() to find a valid (unexpired) CRL if possible.
8640 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8641 this would be called X509_CRL_cmp() but that name is already used by
8642 a function that just compares CRL issuer names. Cache several CRL
8643 extensions in X509_CRL structure and cache CRLDP in X509.
8647 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8648 this maps equivalent X509_NAME structures into a consistent structure.
8649 Name comparison can then be performed rapidly using memcmp().
8653 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8658 * Allow digests to supply their own micalg string for S/MIME type using
8659 the ctrl EVP_MD_CTRL_MICALG.
8663 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8664 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8665 ctrl. It can then customise the structure before and/or after signing
8670 * New function OBJ_add_sigid() to allow application defined signature OIDs
8671 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8672 to free up any added signature OIDs.
8676 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8677 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8678 digest and cipher tables. New options added to openssl utility:
8679 list-message-digest-algorithms and list-cipher-algorithms.
8683 * Change the array representation of binary polynomials: the list
8684 of degrees of non-zero coefficients is now terminated with -1.
8685 Previously it was terminated with 0, which was also part of the
8686 value; thus, the array representation was not applicable to
8687 polynomials where t^0 has coefficient zero. This change makes
8688 the array representation useful in a more general context.
8692 * Various modifications and fixes to SSL/TLS cipher string
8693 handling. For ECC, the code now distinguishes between fixed ECDH
8694 with RSA certificates on the one hand and with ECDSA certificates
8695 on the other hand, since these are separate ciphersuites. The
8696 unused code for Fortezza ciphersuites has been removed.
8698 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8699 (not "ECDHE"). For consistency with the code for DH
8700 certificates, use of ECDH certificates is now considered ECDH
8701 authentication, not RSA or ECDSA authentication (the latter is
8702 merely the CA's signing algorithm and not actively used in the
8705 The temporary ciphersuite alias "ECCdraft" is no longer
8706 available, and ECC ciphersuites are no longer excluded from "ALL"
8707 and "DEFAULT". The following aliases now exist for RFC 4492
8708 ciphersuites, most of these by analogy with the DH case:
8710 kECDHr - ECDH cert, signed with RSA
8711 kECDHe - ECDH cert, signed with ECDSA
8712 kECDH - ECDH cert (signed with either RSA or ECDSA)
8713 kEECDH - ephemeral ECDH
8714 ECDH - ECDH cert or ephemeral ECDH
8720 AECDH - anonymous ECDH
8721 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8725 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8726 Use correct micalg parameters depending on digest(s) in signed message.
8730 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8731 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8735 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8736 an engine to register a method. Add ENGINE lookups for methods and
8737 functional reference processing.
8741 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8742 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8747 * New -resign option to smime utility. This adds one or more signers
8748 to an existing PKCS#7 signedData structure. Also -md option to use an
8749 alternative message digest algorithm for signing.
8753 * Tidy up PKCS#7 routines and add new functions to make it easier to
8754 create PKCS7 structures containing multiple signers. Update smime
8755 application to support multiple signers.
8759 * New -macalg option to pkcs12 utility to allow setting of an alternative
8764 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8765 Reorganize PBE internals to lookup from a static table using NIDs,
8766 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8767 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8768 PRF which will be automatically used with PBES2.
8772 * Replace the algorithm specific calls to generate keys in "req" with the
8777 * Update PKCS#7 enveloped data routines to use new API. This is now
8778 supported by any public key method supporting the encrypt operation. A
8779 ctrl is added to allow the public key algorithm to examine or modify
8780 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8785 * Add a ctrl to asn1 method to allow a public key algorithm to express
8786 a default digest type to use. In most cases this will be SHA1 but some
8787 algorithms (such as GOST) need to specify an alternative digest. The
8788 return value indicates how strong the preference is 1 means optional and
8789 2 is mandatory (that is it is the only supported type). Modify
8790 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8791 use the default md. Update openssl utilities to use the default digest
8792 type for signing if it is not explicitly indicated.
8796 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8797 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8798 signing method from the key type. This effectively removes the link
8799 between digests and public key types.
8803 * Add an OID cross reference table and utility functions. Its purpose is to
8804 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8805 rsaEncryption. This will allow some of the algorithm specific hackery
8806 needed to use the correct OID to be removed.
8810 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8811 structures for PKCS7_sign(). They are now set up by the relevant public
8816 * Add provisional EC pkey method with support for ECDSA and ECDH.
8820 * Add support for key derivation (agreement) in the API, DH method and
8825 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8826 public and private key formats. As a side effect these add additional
8827 command line functionality not previously available: DSA signatures can be
8828 generated and verified using pkeyutl and DH key support and generation in
8835 *Oliver Tappe <zooey@hirschkaefer.de>*
8837 * New make target "install_html_docs" installs HTML renditions of the
8840 *Oliver Tappe <zooey@hirschkaefer.de>*
8842 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8843 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8844 support key and parameter generation and add initial key generation
8845 functionality for RSA.
8849 * Add functions for main EVP_PKEY_method operations. The undocumented
8850 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8851 `EVP_PKEY_{encrypt,decrypt}_old`.
8855 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8856 key API, doesn't do much yet.
8860 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8861 public key algorithms. New option to openssl utility:
8862 "list-public-key-algorithms" to print out info.
8866 * Implement the Supported Elliptic Curves Extension for
8867 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8871 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8872 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8876 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8877 utilities such as rsa, dsa, dsaparam etc except they process any key
8882 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8883 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8884 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8889 * Initial support for pluggable public key ASN1.
8890 De-spaghettify the public key ASN1 handling. Move public and private
8891 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8892 algorithm specific handling to a single module within the relevant
8893 algorithm directory. Add functions to allow (near) opaque processing
8894 of public and private key structures.
8898 * Implement the Supported Point Formats Extension for
8899 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8903 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8904 for the psk identity [hint] and the psk callback functions to the
8905 SSL_SESSION, SSL and SSL_CTX structure.
8908 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8912 SSL_CTX_use_psk_identity_hint
8913 SSL_get_psk_identity_hint
8914 SSL_get_psk_identity
8915 SSL_use_psk_identity_hint
8917 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8919 * Add RFC 3161 compliant time stamp request creation, response generation
8920 and response verification functionality.
8922 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8924 * Add initial support for TLS extensions, specifically for the server_name
8925 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8926 have new members for a host name. The SSL data structure has an
8927 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8928 stored in that context to allow for session resumption, even after the
8929 SSL has been switched to a new SSL_CTX in reaction to a client's
8930 server_name extension.
8932 New functions (subject to change):
8934 SSL_get_servername()
8935 SSL_get_servername_type()
8938 New CTRL codes and macros (subject to change):
8940 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8941 - SSL_CTX_set_tlsext_servername_callback()
8942 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8943 - SSL_CTX_set_tlsext_servername_arg()
8944 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8946 openssl s_client has a new '-servername ...' option.
8948 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8949 '-key2 ...', '-servername_fatal' (subject to change). This allows
8950 testing the HostName extension for a specific single host name ('-cert'
8951 and '-key' remain fallbacks for handshakes without HostName
8952 negotiation). If the unrecognized_name alert has to be sent, this by
8953 default is a warning; it becomes fatal with the '-servername_fatal'
8956 *Peter Sylvester, Remy Allais, Christophe Renou*
8958 * Whirlpool hash implementation is added.
8962 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8963 bn(64,32). Because of instruction set limitations it doesn't have
8964 any negative impact on performance. This was done mostly in order
8965 to make it possible to share assembler modules, such as bn_mul_mont
8966 implementations, between 32- and 64-bit builds without hassle.
8970 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8971 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8976 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8977 dedicated Montgomery multiplication procedure, is introduced.
8978 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8979 "64-bit" performance on certain 32-bit targets.
8983 * New option SSL_OP_NO_COMP to disable use of compression selectively
8984 in SSL structures. New SSL ctrl to set maximum send fragment size.
8985 Save memory by setting the I/O buffer sizes dynamically instead of
8986 using the maximum available value.
8990 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8991 in addition to the text details.
8995 * Very, very preliminary EXPERIMENTAL support for printing of general
8996 ASN1 structures. This currently produces rather ugly output and doesn't
8997 handle several customised structures at all.
9001 * Integrated support for PVK file format and some related formats such
9002 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9003 these in the 'rsa' and 'dsa' utilities.
9007 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9011 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9012 place for the (very old) "NETSCAPE" format certificates which are now
9013 handled using new ASN1 code equivalents.
9017 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9018 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9019 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9023 * Modify CRL distribution points extension code to print out previously
9024 unsupported fields. Enhance extension setting code to allow setting of
9029 * Add print and set support for Issuing Distribution Point CRL extension.
9033 * Change 'Configure' script to enable Camellia by default.
9040 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9042 * When rejecting SSL/TLS records due to an incorrect version number, never
9043 update s->server with a new major version number. As of
9044 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9045 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9046 the previous behavior could result in a read attempt at NULL when
9047 receiving specific incorrect SSL/TLS records once record payload
9048 protection is active. ([CVE-2010-0740])
9050 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9052 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9053 could be crashed if the relevant tables were not present (e.g. chrooted).
9055 *Tomas Hoger <thoger@redhat.com>*
9057 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9059 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9061 *Martin Olsson, Neel Mehta*
9063 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9064 accommodate for stack sorting, always a write lock!).
9068 * On some versions of WIN32 Heap32Next is very slow. This can cause
9069 excessive delays in the RAND_poll(): over a minute. As a workaround
9070 include a time check in the inner Heap32Next loop too.
9074 * The code that handled flushing of data in SSL/TLS originally used the
9075 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9076 the problem outlined in PR#1949. The fix suggested there however can
9077 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9078 of Apache). So instead simplify the code to flush unconditionally.
9079 This should be fine since flushing with no data to flush is a no op.
9083 * Handle TLS versions 2.0 and later properly and correctly use the
9084 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9085 off ancient servers have a habit of sticking around for a while...
9089 * Modify compression code so it frees up structures without using the
9090 ex_data callbacks. This works around a problem where some applications
9091 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9092 restarting) then use compression (e.g. SSL with compression) later.
9093 This results in significant per-connection memory leaks and
9094 has caused some security issues including CVE-2008-1678 and
9099 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9100 change when encrypting or decrypting.
9104 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9105 connect and renegotiate with servers which do not support RI.
9106 Until RI is more widely deployed this option is enabled by default.
9110 * Add "missing" ssl ctrls to clear options and mode.
9114 * If client attempts to renegotiate and doesn't support RI respond with
9115 a no_renegotiation alert as required by RFC5746. Some renegotiating
9116 TLS clients will continue a connection gracefully when they receive
9117 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9118 waiting for a server hello which it will never receive. Now we treat a
9119 received no_renegotiation alert as a fatal error. This is because
9120 applications requesting a renegotiation might well expect it to succeed
9121 and would have no code in place to handle the server denying it so the
9122 only safe thing to do is to terminate the connection.
9126 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9127 peer supports secure renegotiation and 0 otherwise. Print out peer
9128 renegotiation support in s_client/s_server.
9132 * Replace the highly broken and deprecated SPKAC certification method with
9133 the updated NID creation version. This should correctly handle UTF8.
9137 * Implement RFC5746. Re-enable renegotiation but require the extension
9138 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9139 turns out to be a bad idea. It has been replaced by
9140 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9141 SSL_CTX_set_options(). This is really not recommended unless you
9142 know what you are doing.
9144 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9146 * Fixes to stateless session resumption handling. Use initial_ctx when
9147 issuing and attempting to decrypt tickets in case it has changed during
9148 servername handling. Use a non-zero length session ID when attempting
9149 stateless session resumption: this makes it possible to determine if
9150 a resumption has occurred immediately after receiving server hello
9151 (several places in OpenSSL subtly assume this) instead of later in
9156 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9157 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9158 fixes for a few places where the return code is not checked
9161 *Julia Lawall <julia@diku.dk>*
9163 * Add --strict-warnings option to Configure script to include devteam
9164 warnings in other configurations.
9168 * Add support for --libdir option and LIBDIR variable in makefiles. This
9169 makes it possible to install openssl libraries in locations which
9170 have names other than "lib", for example "/usr/lib64" which some
9173 *Steve Henson, based on patch from Jeremy Utley*
9175 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9176 X690 8.9.12 and can produce some misleading textual output of OIDs.
9178 *Steve Henson, reported by Dan Kaminsky*
9180 * Delete MD2 from algorithm tables. This follows the recommendation in
9181 several standards that it is not used in new applications due to
9182 several cryptographic weaknesses. For binary compatibility reasons
9183 the MD2 API is still compiled in by default.
9187 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9192 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9193 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9196 *Guenter <lists@gknw.net>*
9198 * Fix the server certificate chain building code to use X509_verify_cert(),
9199 it used to have an ad-hoc builder which was unable to cope with anything
9200 other than a simple chain.
9202 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9204 * Don't check self signed certificate signatures in X509_verify_cert()
9205 by default (a flag can override this): it just wastes time without
9206 adding any security. As a useful side effect self signed root CAs
9207 with non-FIPS digests are now usable in FIPS mode.
9211 * In dtls1_process_out_of_seq_message() the check if the current message
9212 is already buffered was missing. For every new message was memory
9213 allocated, allowing an attacker to perform an denial of service attack
9214 with sending out of seq handshake messages until there is no memory
9215 left. Additionally every future message was buffered, even if the
9216 sequence number made no sense and would be part of another handshake.
9217 So only messages with sequence numbers less than 10 in advance will be
9218 buffered. ([CVE-2009-1378])
9220 *Robin Seggelmann, discovered by Daniel Mentz*
9222 * Records are buffered if they arrive with a future epoch to be
9223 processed after finishing the corresponding handshake. There is
9224 currently no limitation to this buffer allowing an attacker to perform
9225 a DOS attack with sending records with future epochs until there is no
9226 memory left. This patch adds the pqueue_size() function to determine
9227 the size of a buffer and limits the record buffer to 100 entries.
9230 *Robin Seggelmann, discovered by Daniel Mentz*
9232 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9233 parent structure is freed. ([CVE-2009-1379])
9237 * Handle non-blocking I/O properly in SSL_shutdown() call.
9239 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9241 * Add `2.5.4.*` OIDs
9243 *Ilya O. <vrghost@gmail.com>*
9245 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9247 * Disable renegotiation completely - this fixes a severe security
9248 problem ([CVE-2009-3555]) at the cost of breaking all
9249 renegotiation. Renegotiation can be re-enabled by setting
9250 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9251 run-time. This is really not recommended unless you know what
9256 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9258 * Don't set val to NULL when freeing up structures, it is freed up by
9259 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9260 zeroing past the valid field. ([CVE-2009-0789])
9262 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9264 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9265 checked correctly. This would allow some invalid signed attributes to
9266 appear to verify correctly. ([CVE-2009-0591])
9268 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9270 * Reject UniversalString and BMPString types with invalid lengths. This
9271 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9272 a legal length. ([CVE-2009-0590])
9276 * Set S/MIME signing as the default purpose rather than setting it
9277 unconditionally. This allows applications to override it at the store
9282 * Permit restricted recursion of ASN1 strings. This is needed in practice
9283 to handle some structures.
9287 * Improve efficiency of mem_gets: don't search whole buffer each time
9290 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9292 * New -hex option for openssl rand.
9296 * Print out UTF8String and NumericString when parsing ASN1.
9300 * Support NumericString type for name components.
9304 * Allow CC in the environment to override the automatically chosen
9305 compiler. Note that nothing is done to ensure flags work with the
9310 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9312 * Properly check EVP_VerifyFinal() and similar return values
9315 *Ben Laurie, Bodo Moeller, Google Security Team*
9317 * Enable TLS extensions by default.
9321 * Allow the CHIL engine to be loaded, whether the application is
9322 multithreaded or not. (This does not release the developer from the
9323 obligation to set up the dynamic locking callbacks.)
9325 *Sander Temme <sander@temme.net>*
9327 * Use correct exit code if there is an error in dgst command.
9329 *Steve Henson; problem pointed out by Roland Dirlewanger*
9331 * Tweak Configure so that you need to say "experimental-jpake" to enable
9332 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9336 * Add experimental JPAKE support, including demo authentication in
9337 s_client and s_server.
9341 * Set the comparison function in v3_addr_canonize().
9343 *Rob Austein <sra@hactrn.net>*
9345 * Add support for XMPP STARTTLS in s_client.
9347 *Philip Paeps <philip@freebsd.org>*
9349 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9350 to ensure that even with this option, only ciphersuites in the
9351 server's preference list will be accepted. (Note that the option
9352 applies only when resuming a session, so the earlier behavior was
9353 just about the algorithm choice for symmetric cryptography.)
9357 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9359 * Fix NULL pointer dereference if a DTLS server received
9360 ChangeCipherSpec as first record ([CVE-2009-1386]).
9364 * Fix a state transition in s3_srvr.c and d1_srvr.c
9365 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9369 * The fix in 0.9.8c that supposedly got rid of unsafe
9370 double-checked locking was incomplete for RSA blinding,
9371 addressing just one layer of what turns out to have been
9372 doubly unsafe triple-checked locking.
9374 So now fix this for real by retiring the MONT_HELPER macro
9375 in crypto/rsa/rsa_eay.c.
9377 *Bodo Moeller; problem pointed out by Marius Schilder*
9379 * Various precautionary measures:
9381 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9383 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9384 (NB: This would require knowledge of the secret session ticket key
9385 to exploit, in which case you'd be SOL either way.)
9387 - Change bn_nist.c so that it will properly handle input BIGNUMs
9388 outside the expected range.
9390 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9393 *Neel Mehta, Bodo Moeller*
9395 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9396 the load fails. Useful for distros.
9398 *Ben Laurie and the FreeBSD team*
9400 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9404 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9408 * Expand ENGINE to support engine supplied SSL client certificate functions.
9410 This work was sponsored by Logica.
9414 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9415 keystores. Support for SSL/TLS client authentication too.
9416 Not compiled unless enable-capieng specified to Configure.
9418 This work was sponsored by Logica.
9422 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9423 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9424 attribute creation routines such as certificate requests and PKCS#12
9429 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9431 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9432 handshake which could lead to a client crash as found using the
9433 Codenomicon TLS test suite ([CVE-2008-1672])
9435 *Steve Henson, Mark Cox*
9437 * Fix double free in TLS server name extensions which could lead to
9438 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9442 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9444 Clear the error queue to ensure that error entries left from
9445 older function calls do not interfere with the correct operation.
9447 *Lutz Jaenicke, Erik de Castro Lopo*
9449 * Remove root CA certificates of commercial CAs:
9451 The OpenSSL project does not recommend any specific CA and does not
9452 have any policy with respect to including or excluding any CA.
9453 Therefore it does not make any sense to ship an arbitrary selection
9454 of root CA certificates with the OpenSSL software.
9458 * RSA OAEP patches to fix two separate invalid memory reads.
9459 The first one involves inputs when 'lzero' is greater than
9460 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9461 before the beginning of from). The second one involves inputs where
9462 the 'db' section contains nothing but zeroes (there is a one-byte
9463 invalid read after the end of 'db').
9465 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9467 * Partial backport from 0.9.9-dev:
9469 Introduce bn_mul_mont (dedicated Montgomery multiplication
9470 procedure) as a candidate for BIGNUM assembler implementation.
9471 While 0.9.9-dev uses assembler for various architectures, only
9472 x86_64 is available by default here in the 0.9.8 branch, and
9473 32-bit x86 is available through a compile-time setting.
9475 To try the 32-bit x86 assembler implementation, use Configure
9476 option "enable-montasm" (which exists only for this backport).
9478 As "enable-montasm" for 32-bit x86 disclaims code stability
9479 anyway, in this constellation we activate additional code
9480 backported from 0.9.9-dev for further performance improvements,
9481 namely BN_from_montgomery_word. (To enable this otherwise,
9482 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9484 *Andy Polyakov (backport partially by Bodo Moeller)*
9486 * Add TLS session ticket callback. This allows an application to set
9487 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9488 values. This is useful for key rollover for example where several key
9489 sets may exist with different names.
9493 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9494 This was broken until now in 0.9.8 releases, such that the only way
9495 a registered ENGINE could be used (assuming it initialises
9496 successfully on the host) was to explicitly set it as the default
9497 for the relevant algorithms. This is in contradiction with 0.9.7
9498 behaviour and the documentation. With this fix, when an ENGINE is
9499 registered into a given algorithm's table of implementations, the
9500 'uptodate' flag is reset so that auto-discovery will be used next
9501 time a new context for that algorithm attempts to select an
9504 *Ian Lister (tweaked by Geoff Thorpe)*
9506 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9507 implementation in the following ways:
9509 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9512 Lack of BER streaming support means one pass streaming processing is
9513 only supported if data is detached: setting the streaming flag is
9514 ignored for embedded content.
9516 CMS support is disabled by default and must be explicitly enabled
9517 with the enable-cms configuration option.
9521 * Update the GMP engine glue to do direct copies between BIGNUM and
9522 mpz_t when openssl and GMP use the same limb size. Otherwise the
9523 existing "conversion via a text string export" trick is still used.
9525 *Paul Sheer <paulsheer@gmail.com>*
9527 * Zlib compression BIO. This is a filter BIO which compressed and
9528 uncompresses any data passed through it.
9532 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9533 RFC3394 compatible AES key wrapping.
9537 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9538 sets string data without copying. X509_ALGOR_set0() and
9539 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9540 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9541 from an X509_ATTRIBUTE structure optionally checking it occurs only
9542 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9547 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9548 to get the expected BN_FLG_CONSTTIME behavior.
9550 *Bodo Moeller (Google)*
9554 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9555 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9556 - added some more tests to do_tests.pl
9557 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9558 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9559 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9560 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9561 - various changes to netware.pl to enable gcc-cross builds on Win32
9563 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9564 - various changes to fix missing prototype warnings
9565 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9566 - added AES, WHIRLPOOL and CPUID assembler code to build files
9567 - added missing AES assembler make rules to mk1mf.pl
9568 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9570 *Guenter Knauf <eflash@gmx.net>*
9572 * Implement certificate status request TLS extension defined in RFC3546.
9573 A client can set the appropriate parameters and receive the encoded
9574 OCSP response via a callback. A server can query the supplied parameters
9575 and set the encoded OCSP response in the callback. Add simplified examples
9576 to s_client and s_server.
9580 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9583 + Binary incompatibility of ssl_ctx_st structure
9584 + DTLS interoperation with non-compliant servers
9585 + Don't call get_session_cb() without proposed session
9586 + Fix ia64 assembler code
9588 *Andy Polyakov, Steve Henson*
9590 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9592 * DTLS Handshake overhaul. There were longstanding issues with
9593 OpenSSL DTLS implementation, which were making it impossible for
9594 RFC 4347 compliant client to communicate with OpenSSL server.
9595 Unfortunately just fixing these incompatibilities would "cut off"
9596 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9597 server keeps tolerating non RFC compliant syntax. The opposite is
9598 not true, 0.9.8f client can not communicate with earlier server.
9599 This update even addresses CVE-2007-4995.
9603 * Changes to avoid need for function casts in OpenSSL: some compilers
9604 (gcc 4.2 and later) reject their use.
9605 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9608 * Add RFC4507 support to OpenSSL. This includes the corrections in
9609 RFC4507bis. The encrypted ticket format is an encrypted encoded
9610 SSL_SESSION structure, that way new session features are automatically
9613 If a client application caches session in an SSL_SESSION structure
9614 support is transparent because tickets are now stored in the encoded
9617 The SSL_CTX structure automatically generates keys for ticket
9618 protection in servers so again support should be possible
9619 with no application modification.
9621 If a client or server wishes to disable RFC4507 support then the option
9622 SSL_OP_NO_TICKET can be set.
9624 Add a TLS extension debugging callback to allow the contents of any client
9625 or server extensions to be examined.
9627 This work was sponsored by Google.
9631 * Add initial support for TLS extensions, specifically for the server_name
9632 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9633 have new members for a host name. The SSL data structure has an
9634 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9635 stored in that context to allow for session resumption, even after the
9636 SSL has been switched to a new SSL_CTX in reaction to a client's
9637 server_name extension.
9639 New functions (subject to change):
9641 SSL_get_servername()
9642 SSL_get_servername_type()
9645 New CTRL codes and macros (subject to change):
9647 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9648 - SSL_CTX_set_tlsext_servername_callback()
9649 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9650 - SSL_CTX_set_tlsext_servername_arg()
9651 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9653 openssl s_client has a new '-servername ...' option.
9655 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9656 '-key2 ...', '-servername_fatal' (subject to change). This allows
9657 testing the HostName extension for a specific single host name ('-cert'
9658 and '-key' remain fallbacks for handshakes without HostName
9659 negotiation). If the unrecognized_name alert has to be sent, this by
9660 default is a warning; it becomes fatal with the '-servername_fatal'
9663 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9665 * Add AES and SSE2 assembly language support to VC++ build.
9669 * Mitigate attack on final subtraction in Montgomery reduction.
9673 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9674 (which previously caused an internal error).
9678 * Squeeze another 10% out of IGE mode when in != out.
9682 * AES IGE mode speedup.
9684 *Dean Gaudet (Google)*
9686 * Add the Korean symmetric 128-bit cipher SEED (see
9687 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9688 add SEED ciphersuites from RFC 4162:
9690 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9691 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9692 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9693 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9695 To minimize changes between patchlevels in the OpenSSL 0.9.8
9696 series, SEED remains excluded from compilation unless OpenSSL
9697 is configured with 'enable-seed'.
9699 *KISA, Bodo Moeller*
9701 * Mitigate branch prediction attacks, which can be practical if a
9702 single processor is shared, allowing a spy process to extract
9703 information. For detailed background information, see
9704 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9705 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9706 and Necessary Software Countermeasures"). The core of the change
9707 are new versions BN_div_no_branch() and
9708 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9709 respectively, which are slower, but avoid the security-relevant
9710 conditional branches. These are automatically called by BN_div()
9711 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9712 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9713 remove a conditional branch.
9715 BN_FLG_CONSTTIME is the new name for the previous
9716 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9717 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9718 in the exponent causes BN_mod_exp_mont() to use the alternative
9719 implementation in BN_mod_exp_mont_consttime().) The old name
9720 remains as a deprecated alias.
9722 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9723 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9724 constant-time implementations for more than just exponentiation.
9725 Here too the old name is kept as a deprecated alias.
9727 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9728 the BN_BLINDING structure gets an independent copy of the
9729 modulus. This means that the previous `BIGNUM *m` argument to
9730 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9731 essentially becomes `const BIGNUM *m`, although we can't actually
9732 change this in the header file before 0.9.9. It allows
9733 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9734 enable BN_FLG_CONSTTIME.
9736 *Matthew D Wood (Intel Corp)*
9738 * In the SSL/TLS server implementation, be strict about session ID
9739 context matching (which matters if an application uses a single
9740 external cache for different purposes). Previously,
9741 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9742 set. This did ensure strict client verification, but meant that,
9743 with applications using a single external cache for quite
9744 different requirements, clients could circumvent ciphersuite
9745 restrictions for a given session ID context by starting a session
9746 in a different context.
9750 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9751 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9752 authentication-only ciphersuites.
9756 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9757 not complete and could lead to a possible single byte overflow
9758 ([CVE-2007-5135]) [Ben Laurie]
9760 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9762 * Since AES128 and AES256 (and similarly Camellia128 and
9763 Camellia256) share a single mask bit in the logic of
9764 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9765 kludge to work properly if AES128 is available and AES256 isn't
9766 (or if Camellia128 is available and Camellia256 isn't).
9770 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9771 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9772 When a point or a seed is encoded in a BIT STRING, we need to
9773 prevent the removal of trailing zero bits to get the proper DER
9774 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9775 of a NamedBitList, for which trailing 0 bits need to be removed.)
9779 * Have SSL/TLS server implementation tolerate "mismatched" record
9780 protocol version while receiving ClientHello even if the
9781 ClientHello is fragmented. (The server can't insist on the
9782 particular protocol version it has chosen before the ServerHello
9783 message has informed the client about his choice.)
9787 * Add RFC 3779 support.
9789 *Rob Austein for ARIN, Ben Laurie*
9791 * Load error codes if they are not already present instead of using a
9792 static variable. This allows them to be cleanly unloaded and reloaded.
9793 Improve header file function name parsing.
9797 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9798 or CAPABILITY handshake as required by RFCs.
9802 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9804 * Introduce limits to prevent malicious keys being able to
9805 cause a denial of service. ([CVE-2006-2940])
9807 *Steve Henson, Bodo Moeller*
9809 * Fix ASN.1 parsing of certain invalid structures that can result
9810 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9812 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9813 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9815 * Fix SSL client code which could crash if connecting to a
9816 malicious SSLv2 server. ([CVE-2006-4343])
9818 *Tavis Ormandy and Will Drewry, Google Security Team*
9820 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9821 match only those. Before that, "AES256-SHA" would be interpreted
9822 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9823 the same strength classification in 0.9.7h) as we currently only
9824 have a single AES bit in the ciphersuite description bitmap.
9825 That change, however, also applied to ciphersuite strings such as
9826 "RC4-MD5" that intentionally matched multiple ciphersuites --
9827 namely, SSL 2.0 ciphersuites in addition to the more common ones
9828 from SSL 3.0/TLS 1.0.
9830 So we change the selection algorithm again: Naming an explicit
9831 ciphersuite selects this one ciphersuite, and any other similar
9832 ciphersuite (same bitmap) from *other* protocol versions.
9833 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9834 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9836 Since SSL 2.0 does not have any ciphersuites for which the
9837 128/256 bit distinction would be relevant, this works for now.
9838 The proper fix will be to use different bits for AES128 and
9839 AES256, which would have avoided the problems from the beginning;
9840 however, bits are scarce, so we can only do this in a new release
9841 (not just a patchlevel) when we can change the SSL_CIPHER
9842 definition to split the single 'unsigned long mask' bitmap into
9843 multiple values to extend the available space.
9847 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9849 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9850 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9852 * Add AES IGE and biIGE modes.
9856 * Change the Unix randomness entropy gathering to use poll() when
9857 possible instead of select(), since the latter has some
9858 undesirable limitations.
9860 *Darryl Miles via Richard Levitte and Bodo Moeller*
9862 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9863 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9864 cannot be implicitly activated as part of, e.g., the "AES" alias.
9865 However, please upgrade to OpenSSL 0.9.9[-dev] for
9866 non-experimental use of the ECC ciphersuites to get TLS extension
9867 support, which is required for curve and point format negotiation
9868 to avoid potential handshake problems.
9872 * Disable rogue ciphersuites:
9874 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9875 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9876 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9878 The latter two were purportedly from
9879 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9882 Also deactivate the remaining ciphersuites from
9883 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9884 unofficial, and the ID has long expired.
9888 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9889 dual-core machines) and other potential thread-safety issues.
9893 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9894 versions), which is now available for royalty-free use
9895 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9896 Also, add Camellia TLS ciphersuites from RFC 4132.
9898 To minimize changes between patchlevels in the OpenSSL 0.9.8
9899 series, Camellia remains excluded from compilation unless OpenSSL
9900 is configured with 'enable-camellia'.
9904 * Disable the padding bug check when compression is in use. The padding
9905 bug check assumes the first packet is of even length, this is not
9906 necessarily true if compression is enabled and can result in false
9907 positives causing handshake failure. The actual bug test is ancient
9908 code so it is hoped that implementations will either have fixed it by
9909 now or any which still have the bug do not support compression.
9913 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9915 * When applying a cipher rule check to see if string match is an explicit
9916 cipher suite and only match that one cipher suite if it is.
9920 * Link in manifests for VC++ if needed.
9922 *Austin Ziegler <halostatue@gmail.com>*
9924 * Update support for ECC-based TLS ciphersuites according to
9925 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9926 TLS extensions, which are supported starting with the 0.9.9
9927 branch, not in the OpenSSL 0.9.8 branch).
9931 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9932 opaque EVP_CIPHER_CTX handling.
9936 * Fixes and enhancements to zlib compression code. We now only use
9937 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9938 to conform with the standards mentioned here:
9939 <http://www.zlib.net/DLL_FAQ.txt>
9940 Static zlib linking now works on Windows and the new --with-zlib-include
9941 --with-zlib-lib options to Configure can be used to supply the location
9942 of the headers and library. Gracefully handle case where zlib library
9947 * Several fixes and enhancements to the OID generation code. The old code
9948 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9949 handle numbers larger than ULONG_MAX, truncated printing and had a
9950 non standard OBJ_obj2txt() behaviour.
9954 * Add support for building of engines under engine/ as shared libraries
9955 under VC++ build system.
9959 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9960 Hopefully, we will not see any false combination of paths any more.
9964 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9966 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9967 (part of SSL_OP_ALL). This option used to disable the
9968 countermeasure against man-in-the-middle protocol-version
9969 rollback in the SSL 2.0 server implementation, which is a bad
9970 idea. ([CVE-2005-2969])
9972 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9973 for Information Security, National Institute of Advanced Industrial
9974 Science and Technology [AIST], Japan)*
9976 * Add two function to clear and return the verify parameter flags.
9980 * Keep cipherlists sorted in the source instead of sorting them at
9981 runtime, thus removing the need for a lock.
9985 * Avoid some small subgroup attacks in Diffie-Hellman.
9987 *Nick Mathewson and Ben Laurie*
9989 * Add functions for well-known primes.
9993 * Extended Windows CE support.
9995 *Satoshi Nakamura and Andy Polyakov*
9997 * Initialize SSL_METHOD structures at compile time instead of during
9998 runtime, thus removing the need for a lock.
10002 * Make PKCS7_decrypt() work even if no certificate is supplied by
10003 attempting to decrypt each encrypted key in turn. Add support to
10008 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10010 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10013 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10017 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10018 key into the same file any more.
10022 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10026 * Add -utf8 command line and config file option to 'ca'.
10028 *Stefan <stf@udoma.org*
10030 * Removed the macro des_crypt(), as it seems to conflict with some
10031 libraries. Use DES_crypt().
10035 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10036 involves renaming the source and generated shared-libs for
10037 both. The engines will accept the corrected or legacy ids
10038 ('ncipher' and '4758_cca' respectively) when binding. NB,
10039 this only applies when building 'shared'.
10041 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10043 * Add attribute functions to EVP_PKEY structure. Modify
10044 PKCS12_create() to recognize a CSP name attribute and
10045 use it. Make -CSP option work again in pkcs12 utility.
10049 * Add new functionality to the bn blinding code:
10050 - automatic re-creation of the BN_BLINDING parameters after
10051 a fixed number of uses (currently 32)
10052 - add new function for parameter creation
10053 - introduce flags to control the update behaviour of the
10054 BN_BLINDING parameters
10055 - hide BN_BLINDING structure
10056 Add a second BN_BLINDING slot to the RSA structure to improve
10057 performance when a single RSA object is shared among several
10062 * Add support for DTLS.
10064 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10066 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10067 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10071 * Remove buggy and incomplete DH cert support from
10072 ssl/ssl_rsa.c and ssl/s3_both.c
10076 * Use SHA-1 instead of MD5 as the default digest algorithm for
10077 the `apps/openssl` commands.
10081 * Compile clean with "-Wall -Wmissing-prototypes
10082 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10083 DEBUG_SAFESTACK must also be set.
10087 * Change ./Configure so that certain algorithms can be disabled by default.
10088 The new counterpiece to "no-xxx" is "enable-xxx".
10090 The patented RC5 and MDC2 algorithms will now be disabled unless
10091 "enable-rc5" and "enable-mdc2", respectively, are specified.
10093 (IDEA remains enabled despite being patented. This is because IDEA
10094 is frequently required for interoperability, and there is no license
10095 fee for non-commercial use. As before, "no-idea" can be used to
10096 avoid this algorithm.)
10100 * Add processing of proxy certificates (see RFC 3820). This work was
10101 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10102 EGEE (Enabling Grids for E-science in Europe).
10106 * RC4 performance overhaul on modern architectures/implementations, such
10107 as Intel P4, IA-64 and AMD64.
10111 * New utility extract-section.pl. This can be used specify an alternative
10112 section number in a pod file instead of having to treat each file as
10113 a separate case in Makefile. This can be done by adding two lines to the
10116 =for comment openssl_section:XXX
10118 The blank line is mandatory.
10122 * New arguments -certform, -keyform and -pass for s_client and s_server
10123 to allow alternative format key and certificate files and passphrase
10128 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10129 update associated structures and add various utility functions.
10131 Add new policy related verify parameters, include policy checking in
10132 standard verify code. Enhance 'smime' application with extra parameters
10133 to support policy checking and print out.
10137 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10138 Nehemiah processors. These extensions support AES encryption in hardware
10139 as well as RNG (though RNG support is currently disabled).
10141 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10143 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10147 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10149 *Andy Polyakov and a number of other people*
10151 * Improved PowerPC platform support. Most notably BIGNUM assembler
10152 implementation contributed by IBM.
10154 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10156 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10157 exponent rather than 'unsigned long'. There is a corresponding change to
10158 the new 'rsa_keygen' element of the RSA_METHOD structure.
10160 *Jelte Jansen, Geoff Thorpe*
10162 * Functionality for creating the initial serial number file is now
10163 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10165 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10166 number file to 1, which is bound to cause problems. To avoid
10167 the problems while respecting compatibility between different 0.9.7
10168 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10169 CA.pl for serial number initialization. With the new release 0.9.8,
10170 we can fix the problem directly in the 'ca' utility.)
10174 * Reduced header interdependencies by declaring more opaque objects in
10175 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10176 give fewer recursive includes, which could break lazy source code - so
10177 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10178 developers should define this symbol when building and using openssl to
10179 ensure they track the recommended behaviour, interfaces, [etc], but
10180 backwards-compatible behaviour prevails when this isn't defined.
10184 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10188 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10189 This will generate a random key of the appropriate length based on the
10190 cipher context. The EVP_CIPHER can provide its own random key generation
10191 routine to support keys of a specific form. This is used in the des and
10192 3des routines to generate a key of the correct parity. Update S/MIME
10193 code to use new functions and hence generate correct parity DES keys.
10194 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10195 valid (weak or incorrect parity).
10199 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10200 as looking them up. This is useful when the verified structure may contain
10201 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10202 present unless the new PKCS7_NO_CRL flag is asserted.
10206 * Extend ASN1 oid configuration module. It now additionally accepts the
10209 shortName = some long name, 1.2.3.4
10213 * Reimplemented the BN_CTX implementation. There is now no more static
10214 limitation on the number of variables it can handle nor the depth of the
10215 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10216 information can now expand as required, and rather than having a single
10217 static array of bignums, BN_CTX now uses a linked-list of such arrays
10218 allowing it to expand on demand whilst maintaining the usefulness of
10219 BN_CTX's "bundling".
10223 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10224 to allow all RSA operations to function using a single BN_CTX.
10228 * Preliminary support for certificate policy evaluation and checking. This
10229 is initially intended to pass the tests outlined in "Conformance Testing
10230 of Relying Party Client Certificate Path Processing Logic" v1.07.
10234 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10235 remained unused and not that useful. A variety of other little bignum
10236 tweaks and fixes have also been made continuing on from the audit (see
10241 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10242 associated ASN1, EVP and SSL functions and old ASN1 macros.
10246 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10247 and this should never fail. So the return value from the use of
10248 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10249 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10253 * BN_CTX_get() should return zero-valued bignums, providing the same
10254 initialised value as BN_new().
10256 *Geoff Thorpe, suggested by Ulf Möller*
10258 * Support for inhibitAnyPolicy certificate extension.
10262 * An audit of the BIGNUM code is underway, for which debugging code is
10263 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10264 is considered valid when processing BIGNUMs, and causes execution to
10265 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10266 further steps are taken to deliberately pollute unused data in BIGNUM
10267 structures to try and expose faulty code further on. For now, openssl will
10268 (in its default mode of operation) continue to tolerate the inconsistent
10269 forms that it has tolerated in the past, but authors and packagers should
10270 consider trying openssl and their own applications when compiled with
10271 these debugging symbols defined. It will help highlight potential bugs in
10272 their own code, and will improve the test coverage for OpenSSL itself. At
10273 some point, these tighter rules will become openssl's default to improve
10274 maintainability, though the assert()s and other overheads will remain only
10275 in debugging configurations. See bn.h for more details.
10277 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10279 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10280 that can only be obtained through BN_CTX_new() (which implicitly
10281 initialises it). The presence of this function only made it possible
10282 to overwrite an existing structure (and cause memory leaks).
10286 * Because of the callback-based approach for implementing LHASH as a
10287 template type, lh_insert() adds opaque objects to hash-tables and
10288 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10289 to clean up those corresponding objects before destroying the hash table
10290 (and losing the object pointers). So some over-zealous constifications in
10291 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10292 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10293 prototyped to have "const" restrictions on the object pointers they are
10294 given (and so aren't required to cast them away any more).
10298 * The tmdiff.h API was so ugly and minimal that our own timing utility
10299 (speed) prefers to use its own implementation. The two implementations
10300 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10301 its object type properly exposed (MS_TM) instead of casting to/from
10302 `char *`. This may still change yet if someone realises MS_TM and
10304 aren't necessarily the greatest nomenclatures - but this is what was used
10305 internally to the implementation so I've used that for now.
10309 * Ensure that deprecated functions do not get compiled when
10310 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10311 the self-tests were still using deprecated key-generation functions so
10312 these have been updated also.
10316 * Reorganise PKCS#7 code to separate the digest location functionality
10317 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10318 New function PKCS7_set_digest() to set the digest type for PKCS#7
10319 digestedData type. Add additional code to correctly generate the
10320 digestedData type and add support for this type in PKCS7 initialization
10325 * New function PKCS7_set0_type_other() this initializes a PKCS7
10326 structure of type "other".
10330 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10331 sure the loop does correctly stop and breaking ("division by zero")
10332 modulus operations are not performed. The (pre-generated) prime
10333 table crypto/bn/bn_prime.h was already correct, but it could not be
10334 re-generated on some platforms because of the "division by zero"
10335 situation in the script.
10337 *Ralf S. Engelschall*
10339 * Update support for ECC-based TLS ciphersuites according to
10340 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10341 SHA-1 now is only used for "small" curves (where the
10342 representation of a field element takes up to 24 bytes); for
10343 larger curves, the field element resulting from ECDH is directly
10344 used as premaster secret.
10346 *Douglas Stebila (Sun Microsystems Laboratories)*
10348 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10349 curve secp160r1 to the tests.
10351 *Douglas Stebila (Sun Microsystems Laboratories)*
10353 * Add the possibility to load symbols globally with DSO.
10355 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10357 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10358 control of the error stack.
10362 * Add support for STORE in ENGINE.
10366 * Add the STORE type. The intention is to provide a common interface
10367 to certificate and key stores, be they simple file-based stores, or
10368 HSM-type store, or LDAP stores, or...
10369 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10373 * Add a generic structure called OPENSSL_ITEM. This can be used to
10374 pass a list of arguments to any function as well as provide a way
10375 for a function to pass data back to the caller.
10379 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10380 works like BUF_strdup() but can be used to duplicate a portion of
10381 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10386 * Add the function sk_find_ex() which works like sk_find(), but will
10387 return an index to an element even if an exact match couldn't be
10388 found. The index is guaranteed to point at the element where the
10389 searched-for key would be inserted to preserve sorting order.
10393 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10394 takes an extra flags argument for optional functionality. Currently,
10395 the following flags are defined:
10397 OBJ_BSEARCH_VALUE_ON_NOMATCH
10398 This one gets OBJ_bsearch_ex() to return a pointer to the first
10399 element where the comparing function returns a negative or zero
10402 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10403 This one gets OBJ_bsearch_ex() to return a pointer to the first
10404 element where the comparing function returns zero. This is useful
10405 if there are more than one element where the comparing function
10410 * Make it possible to create self-signed certificates with 'openssl ca'
10411 in such a way that the self-signed certificate becomes part of the
10412 CA database and uses the same mechanisms for serial number generation
10413 as all other certificate signing. The new flag '-selfsign' enables
10414 this functionality. Adapt CA.sh and CA.pl.in.
10418 * Add functionality to check the public key of a certificate request
10419 against a given private. This is useful to check that a certificate
10420 request can be signed by that key (self-signing).
10424 * Make it possible to have multiple active certificates with the same
10425 subject in the CA index file. This is done only if the keyword
10426 'unique_subject' is set to 'no' in the main CA section (default
10427 if 'CA_default') of the configuration file. The value is saved
10428 with the database itself in a separate index attribute file,
10429 named like the index file with '.attr' appended to the name.
10433 * Generate multi-valued AVAs using '+' notation in config files for
10438 * Support for nameConstraints certificate extension.
10442 * Support for policyConstraints certificate extension.
10446 * Support for policyMappings certificate extension.
10450 * Make sure the default DSA_METHOD implementation only uses its
10451 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10452 and change its own handlers to be NULL so as to remove unnecessary
10453 indirection. This lets alternative implementations fallback to the
10454 default implementation more easily.
10458 * Support for directoryName in GeneralName related extensions
10463 * Make it possible to link applications using Makefile.shared.
10464 Make that possible even when linking against static libraries!
10468 * Support for single pass processing for S/MIME signing. This now
10469 means that S/MIME signing can be done from a pipe, in addition
10470 cleartext signing (multipart/signed type) is effectively streaming
10471 and the signed data does not need to be all held in memory.
10473 This is done with a new flag PKCS7_STREAM. When this flag is set
10474 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10475 is done after the data is output (and digests calculated) in
10476 SMIME_write_PKCS7().
10480 * Add full support for -rpath/-R, both in shared libraries and
10481 applications, at least on the platforms where it's known how
10486 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10487 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10488 will now compute a table of multiples of the generator that
10489 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10490 faster (notably in the case of a single point multiplication,
10491 scalar * generator).
10493 *Nils Larsch, Bodo Moeller*
10495 * IPv6 support for certificate extensions. The various extensions
10496 which use the IP:a.b.c.d can now take IPv6 addresses using the
10497 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10502 * Added an ENGINE that implements RSA by performing private key
10503 exponentiations with the GMP library. The conversions to and from
10504 GMP's mpz_t format aren't optimised nor are any montgomery forms
10505 cached, and on x86 it appears OpenSSL's own performance has caught up.
10506 However there are likely to be other architectures where GMP could
10507 provide a boost. This ENGINE is not built in by default, but it can be
10508 specified at Configure time and should be accompanied by the necessary
10509 linker additions, eg;
10510 ./config -DOPENSSL_USE_GMP -lgmp
10514 * "openssl engine" will not display ENGINE/DSO load failure errors when
10515 testing availability of engines with "-t" - the old behaviour is
10516 produced by increasing the feature's verbosity with "-tt".
10520 * ECDSA routines: under certain error conditions uninitialized BN objects
10521 could be freed. Solution: make sure initialization is performed early
10522 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10527 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10528 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10529 software implementations. For DSA and DH, parameter generation can
10530 also be overridden by providing the appropriate method callbacks.
10534 * Change the "progress" mechanism used in key-generation and
10535 primality testing to functions that take a new BN_GENCB pointer in
10536 place of callback/argument pairs. The new API functions have `_ex`
10537 postfixes and the older functions are reimplemented as wrappers for
10538 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10539 declarations of the old functions to help (graceful) attempts to
10540 migrate to the new functions. Also, the new key-generation API
10541 functions operate on a caller-supplied key-structure and return
10542 success/failure rather than returning a key or NULL - this is to
10543 help make "keygen" another member function of RSA_METHOD etc.
10545 Example for using the new callback interface:
10547 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10548 void *my_arg = ...;
10551 BN_GENCB_set(&my_cb, my_callback, my_arg);
10553 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10554 /* For the meaning of a, b in calls to my_callback(), see the
10555 * documentation of the function that calls the callback.
10556 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10557 * my_callback should return 1 if it wants BN_is_prime_ex()
10558 * to continue, or 0 to stop.
10563 * Change the ZLIB compression method to be stateful, and make it
10564 available to TLS with the number defined in
10565 draft-ietf-tls-compression-04.txt.
10569 * Add the ASN.1 structures and functions for CertificatePair, which
10570 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10572 CertificatePair ::= SEQUENCE {
10573 forward [0] Certificate OPTIONAL,
10574 reverse [1] Certificate OPTIONAL,
10575 -- at least one of the pair shall be present -- }
10577 Also implement the PEM functions to read and write certificate
10578 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10580 This needed to be defined, mostly for the sake of the LDAP
10581 attribute crossCertificatePair, but may prove useful elsewhere as
10586 * Make it possible to inhibit symlinking of shared libraries in
10587 Makefile.shared, for Cygwin's sake.
10591 * Extend the BIGNUM API by creating a function
10592 void BN_set_negative(BIGNUM *a, int neg);
10593 and a macro that behave like
10594 int BN_is_negative(const BIGNUM *a);
10596 to avoid the need to access 'a->neg' directly in applications.
10600 * Implement fast modular reduction for pseudo-Mersenne primes
10601 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10602 EC_GROUP_new_curve_GFp() will now automatically use this
10605 *Nils Larsch <nla@trustcenter.de>*
10607 * Add new lock type (CRYPTO_LOCK_BN).
10611 * Change the ENGINE framework to automatically load engines
10612 dynamically from specific directories unless they could be
10613 found to already be built in or loaded. Move all the
10614 current engines except for the cryptodev one to a new
10615 directory engines/.
10616 The engines in engines/ are built as shared libraries if
10617 the "shared" options was given to ./Configure or ./config.
10618 Otherwise, they are inserted in libcrypto.a.
10619 /usr/local/ssl/engines is the default directory for dynamic
10620 engines, but that can be overridden at configure time through
10621 the usual use of --prefix and/or --openssldir, and at run
10622 time with the environment variable OPENSSL_ENGINES.
10624 *Geoff Thorpe and Richard Levitte*
10626 * Add Makefile.shared, a helper makefile to build shared
10627 libraries. Adapt Makefile.org.
10631 * Add version info to Win32 DLLs.
10633 *Peter 'Luna' Runestig" <peter@runestig.com>*
10635 * Add new 'medium level' PKCS#12 API. Certificates and keys
10636 can be added using this API to created arbitrary PKCS#12
10637 files while avoiding the low level API.
10639 New options to PKCS12_create(), key or cert can be NULL and
10640 will then be omitted from the output file. The encryption
10641 algorithm NIDs can be set to -1 for no encryption, the mac
10642 iteration count can be set to 0 to omit the mac.
10644 Enhance pkcs12 utility by making the -nokeys and -nocerts
10645 options work when creating a PKCS#12 file. New option -nomac
10646 to omit the mac, NONE can be set for an encryption algorithm.
10647 New code is modified to use the enhanced PKCS12_create()
10648 instead of the low level API.
10652 * Extend ASN1 encoder to support indefinite length constructed
10653 encoding. This can output sequences tags and octet strings in
10654 this form. Modify pk7_asn1.c to support indefinite length
10655 encoding. This is experimental and needs additional code to
10656 be useful, such as an ASN1 bio and some enhanced streaming
10659 Extend template encode functionality so that tagging is passed
10660 down to the template encoder.
10664 * Let 'openssl req' fail if an argument to '-newkey' is not
10665 recognized instead of using RSA as a default.
10669 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10670 As these are not official, they are not included in "ALL";
10671 the "ECCdraft" ciphersuite group alias can be used to select them.
10673 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10675 * Add ECDH engine support.
10677 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10679 * Add ECDH in new directory crypto/ecdh/.
10681 *Douglas Stebila (Sun Microsystems Laboratories)*
10683 * Let BN_rand_range() abort with an error after 100 iterations
10684 without success (which indicates a broken PRNG).
10688 * Change BN_mod_sqrt() so that it verifies that the input value
10689 is really the square of the return value. (Previously,
10690 BN_mod_sqrt would show GIGO behaviour.)
10694 * Add named elliptic curves over binary fields from X9.62, SECG,
10695 and WAP/WTLS; add OIDs that were still missing.
10697 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10699 * Extend the EC library for elliptic curves over binary fields
10700 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10703 EC_GF2m_simple_method
10707 EC_GROUP_new_curve_GF2m
10708 EC_GROUP_set_curve_GF2m
10709 EC_GROUP_get_curve_GF2m
10710 EC_POINT_set_affine_coordinates_GF2m
10711 EC_POINT_get_affine_coordinates_GF2m
10712 EC_POINT_set_compressed_coordinates_GF2m
10714 Point compression for binary fields is disabled by default for
10715 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10718 As binary polynomials are represented as BIGNUMs, various members
10719 of the EC_GROUP and EC_POINT data structures can be shared
10720 between the implementations for prime fields and binary fields;
10721 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10722 are essentially identical to their `..._GFp` counterparts.
10723 (For simplicity, the `..._GFp` prefix has been dropped from
10724 various internal method names.)
10726 An internal 'field_div' method (similar to 'field_mul' and
10727 'field_sqr') has been added; this is used only for binary fields.
10729 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10731 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10732 through methods ('mul', 'precompute_mult').
10734 The generic implementations (now internally called 'ec_wNAF_mul'
10735 and 'ec_wNAF_precomputed_mult') remain the default if these
10736 methods are undefined.
10738 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10740 * New function EC_GROUP_get_degree, which is defined through
10741 EC_METHOD. For curves over prime fields, this returns the bit
10742 length of the modulus.
10744 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10746 * New functions EC_GROUP_dup, EC_POINT_dup.
10747 (These simply call ..._new and ..._copy).
10749 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10751 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10752 Polynomials are represented as BIGNUMs (where the sign bit is not
10753 used) in the following functions [macros]:
10756 BN_GF2m_sub [= BN_GF2m_add]
10757 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10758 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10759 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10761 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10762 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10763 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10764 BN_GF2m_cmp [= BN_ucmp]
10766 (Note that only the 'mod' functions are actually for fields GF(2^m).
10767 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10769 For some functions, an the irreducible polynomial defining a
10770 field can be given as an 'unsigned int[]' with strictly
10771 decreasing elements giving the indices of those bits that are set;
10772 i.e., p[] represents the polynomial
10773 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10775 p[0] > p[1] > ... > p[k] = 0.
10776 This applies to the following functions:
10779 BN_GF2m_mod_mul_arr
10780 BN_GF2m_mod_sqr_arr
10781 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10782 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10783 BN_GF2m_mod_exp_arr
10784 BN_GF2m_mod_sqrt_arr
10785 BN_GF2m_mod_solve_quad_arr
10789 Conversion can be performed by the following functions:
10794 bntest.c has additional tests for binary polynomial arithmetic.
10796 Two implementations for BN_GF2m_mod_div() are available.
10797 The default algorithm simply uses BN_GF2m_mod_inv() and
10798 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10799 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10800 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10802 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10804 * Add new error code 'ERR_R_DISABLED' that can be used when some
10805 functionality is disabled at compile-time.
10807 *Douglas Stebila <douglas.stebila@sun.com>*
10809 * Change default behaviour of 'openssl asn1parse' so that more
10810 information is visible when viewing, e.g., a certificate:
10812 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10813 mode the content of non-printable OCTET STRINGs is output in a
10814 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10815 avoid the appearance of a printable string.
10817 *Nils Larsch <nla@trustcenter.de>*
10819 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10821 EC_GROUP_set_asn1_flag()
10822 EC_GROUP_get_asn1_flag()
10823 EC_GROUP_set_point_conversion_form()
10824 EC_GROUP_get_point_conversion_form()
10825 These control ASN1 encoding details:
10826 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10827 has been set to OPENSSL_EC_NAMED_CURVE.
10828 - Points are encoded in uncompressed form by default; options for
10829 asn1_for are as for point2oct, namely
10830 POINT_CONVERSION_COMPRESSED
10831 POINT_CONVERSION_UNCOMPRESSED
10832 POINT_CONVERSION_HYBRID
10834 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10836 EC_GROUP_set_seed()
10837 EC_GROUP_get0_seed()
10838 EC_GROUP_get_seed_len()
10839 This is used only for ASN1 purposes (so far).
10841 *Nils Larsch <nla@trustcenter.de>*
10843 * Add 'field_type' member to EC_METHOD, which holds the NID
10844 of the appropriate field type OID. The new function
10845 EC_METHOD_get_field_type() returns this value.
10847 *Nils Larsch <nla@trustcenter.de>*
10850 EC_POINT_point2bn()
10851 EC_POINT_bn2point()
10852 EC_POINT_point2hex()
10853 EC_POINT_hex2point()
10854 providing useful interfaces to EC_POINT_point2oct() and
10855 EC_POINT_oct2point().
10857 *Nils Larsch <nla@trustcenter.de>*
10859 * Change internals of the EC library so that the functions
10860 EC_GROUP_set_generator()
10861 EC_GROUP_get_generator()
10862 EC_GROUP_get_order()
10863 EC_GROUP_get_cofactor()
10864 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10865 to methods, which would lead to unnecessary code duplication when
10866 adding different types of curves.
10868 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10870 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10871 arithmetic, and such that modified wNAFs are generated
10872 (which avoid length expansion in many cases).
10876 * Add a function EC_GROUP_check_discriminant() (defined via
10877 EC_METHOD) that verifies that the curve discriminant is non-zero.
10879 Add a function EC_GROUP_check() that makes some sanity tests
10880 on a EC_GROUP, its generator and order. This includes
10881 EC_GROUP_check_discriminant().
10883 *Nils Larsch <nla@trustcenter.de>*
10885 * Add ECDSA in new directory crypto/ecdsa/.
10887 Add applications 'openssl ecparam' and 'openssl ecdsa'
10888 (these are based on 'openssl dsaparam' and 'openssl dsa').
10890 ECDSA support is also included in various other files across the
10891 library. Most notably,
10892 - 'openssl req' now has a '-newkey ecdsa:file' option;
10893 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10894 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10895 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10896 them suitable for ECDSA where domain parameters must be
10897 extracted before the specific public key;
10898 - ECDSA engine support has been added.
10900 *Nils Larsch <nla@trustcenter.de>*
10902 * Include some named elliptic curves, and add OIDs from X9.62,
10903 SECG, and WAP/WTLS. Each curve can be obtained from the new
10905 EC_GROUP_new_by_curve_name(),
10906 and the list of available named curves can be obtained with
10907 EC_get_builtin_curves().
10908 Also add a 'curve_name' member to EC_GROUP objects, which can be
10910 EC_GROUP_set_curve_name()
10911 EC_GROUP_get_curve_name()
10913 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10915 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10916 was actually never needed) and in BN_mul(). The removal in BN_mul()
10917 required a small change in bn_mul_part_recursive() and the addition
10918 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10919 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10920 bn_sub_words() and bn_add_words() except they take arrays with
10925 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10927 * Cleanse PEM buffers before freeing them since they may contain
10930 *Benjamin Bennett <ben@psc.edu>*
10932 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10933 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10934 authentication-only ciphersuites.
10938 * Since AES128 and AES256 share a single mask bit in the logic of
10939 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10940 kludge to work properly if AES128 is available and AES256 isn't.
10944 * Expand security boundary to match 1.1.1 module.
10948 * Remove redundant features: hash file source, editing of test vectors
10949 modify fipsld to use external fips_premain.c signature.
10953 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10954 run algorithm test programs.
10958 * Make algorithm test programs more tolerant of whitespace.
10962 * Have SSL/TLS server implementation tolerate "mismatched" record
10963 protocol version while receiving ClientHello even if the
10964 ClientHello is fragmented. (The server can't insist on the
10965 particular protocol version it has chosen before the ServerHello
10966 message has informed the client about his choice.)
10970 * Load error codes if they are not already present instead of using a
10971 static variable. This allows them to be cleanly unloaded and reloaded.
10975 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10977 * Introduce limits to prevent malicious keys being able to
10978 cause a denial of service. ([CVE-2006-2940])
10980 *Steve Henson, Bodo Moeller*
10982 * Fix ASN.1 parsing of certain invalid structures that can result
10983 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10985 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10986 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10988 * Fix SSL client code which could crash if connecting to a
10989 malicious SSLv2 server. ([CVE-2006-4343])
10991 *Tavis Ormandy and Will Drewry, Google Security Team*
10993 * Change ciphersuite string processing so that an explicit
10994 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10995 will no longer include "AES128-SHA"), and any other similar
10996 ciphersuite (same bitmap) from *other* protocol versions (so that
10997 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10998 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10999 changes from 0.9.8b and 0.9.8d.
11003 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11005 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11006 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11008 * Change the Unix randomness entropy gathering to use poll() when
11009 possible instead of select(), since the latter has some
11010 undesirable limitations.
11012 *Darryl Miles via Richard Levitte and Bodo Moeller*
11014 * Disable rogue ciphersuites:
11016 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11017 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11018 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11020 The latter two were purportedly from
11021 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11024 Also deactivate the remaining ciphersuites from
11025 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11026 unofficial, and the ID has long expired.
11030 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11031 dual-core machines) and other potential thread-safety issues.
11035 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11037 * Adapt fipsld and the build system to link against the validated FIPS
11038 module in FIPS mode.
11042 * Fixes for VC++ 2005 build under Windows.
11046 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11047 from a Windows bash shell such as MSYS. It is autodetected from the
11048 "config" script when run from a VC++ environment. Modify standard VC++
11049 build to use fipscanister.o from the GNU make build.
11053 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11055 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11056 The value now differs depending on if you build for FIPS or not.
11057 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11058 safely run with a non-FIPSed libcrypto, as it may crash because of
11059 the difference induced by this change.
11063 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11065 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11066 (part of SSL_OP_ALL). This option used to disable the
11067 countermeasure against man-in-the-middle protocol-version
11068 rollback in the SSL 2.0 server implementation, which is a bad
11069 idea. ([CVE-2005-2969])
11071 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11072 for Information Security, National Institute of Advanced Industrial
11073 Science and Technology [AIST, Japan)]*
11075 * Minimal support for X9.31 signatures and PSS padding modes. This is
11076 mainly for FIPS compliance and not fully integrated at this stage.
11080 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11081 the exponentiation using a fixed-length exponent. (Otherwise,
11082 the information leaked through timing could expose the secret key
11083 after many signatures; cf. Bleichenbacher's attack on DSA with
11088 * Make a new fixed-window mod_exp implementation the default for
11089 RSA, DSA, and DH private-key operations so that the sequence of
11090 squares and multiplies and the memory access pattern are
11091 independent of the particular secret key. This will mitigate
11092 cache-timing and potential related attacks.
11094 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11095 and this is automatically used by BN_mod_exp_mont() if the new flag
11096 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11097 will use this BN flag for private exponents unless the flag
11098 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11099 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11101 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11103 * Change the client implementation for SSLv23_method() and
11104 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11105 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11106 (Previously, the SSL 2.0 backwards compatible Client Hello
11107 message format would be used even with SSL_OP_NO_SSLv2.)
11111 * Add support for smime-type MIME parameter in S/MIME messages which some
11116 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11117 a threadsafe manner. Modify rsa code to use new function and add calls
11118 to dsa and dh code (which had race conditions before).
11122 * Include the fixed error library code in the C error file definitions
11123 instead of fixing them up at runtime. This keeps the error code
11124 structures constant.
11128 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11130 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11133 * Fixes for newer kerberos headers. NB: the casts are needed because
11134 the 'length' field is signed on one version and unsigned on another
11135 with no (?) obvious way to tell the difference, without these VC++
11136 complains. Also the "definition" of FAR (blank) is no longer included
11137 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11138 some needed definitions.
11142 * Undo Cygwin change.
11146 * Added support for proxy certificates according to RFC 3820.
11147 Because they may be a security thread to unaware applications,
11148 they must be explicitly allowed in run-time. See
11149 docs/HOWTO/proxy_certificates.txt for further information.
11153 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11155 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11156 server and client random values. Previously
11157 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11158 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11160 This change has negligible security impact because:
11162 1. Server and client random values still have 24 bytes of pseudo random
11165 2. Server and client random values are sent in the clear in the initial
11168 3. The master secret is derived using the premaster secret (48 bytes in
11169 size for static RSA ciphersuites) as well as client server and random
11172 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11175 *Stephen Henson, reported by UK NISCC*
11177 * Use Windows randomness collection on Cygwin.
11181 * Fix hang in EGD/PRNGD query when communication socket is closed
11182 prematurely by EGD/PRNGD.
11184 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11186 * Prompt for pass phrases when appropriate for PKCS12 input format.
11190 * Back-port of selected performance improvements from development
11191 branch, as well as improved support for PowerPC platforms.
11195 * Add lots of checks for memory allocation failure, error codes to indicate
11196 failure and freeing up memory if a failure occurs.
11198 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11200 * Add new -passin argument to dgst.
11204 * Perform some character comparisons of different types in X509_NAME_cmp:
11205 this is needed for some certificates that re-encode DNs into UTF8Strings
11206 (in violation of RFC3280) and can't or won't issue name rollover
11211 * Make an explicit check during certificate validation to see that
11212 the CA setting in each certificate on the chain is correct. As a
11213 side effect always do the following basic checks on extensions,
11214 not just when there's an associated purpose to the check:
11216 - if there is an unhandled critical extension (unless the user
11217 has chosen to ignore this fault)
11218 - if the path length has been exceeded (if one is set at all)
11219 - that certain extensions fit the associated purpose (if one has
11224 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11226 * Avoid a race condition when CRLs are checked in a multi threaded
11227 environment. This would happen due to the reordering of the revoked
11228 entries during signature checking and serial number lookup. Now the
11229 encoding is cached and the serial number sort performed under a lock.
11230 Add new STACK function sk_is_sorted().
11234 * Add Delta CRL to the extension code.
11238 * Various fixes to s3_pkt.c so alerts are sent properly.
11240 *David Holmes <d.holmes@f5.com>*
11242 * Reduce the chances of duplicate issuer name and serial numbers (in
11243 violation of RFC3280) using the OpenSSL certificate creation utilities.
11244 This is done by creating a random 64 bit value for the initial serial
11245 number when a serial number file is created or when a self signed
11246 certificate is created using 'openssl req -x509'. The initial serial
11247 number file is created using 'openssl x509 -next_serial' in CA.pl
11248 rather than being initialized to 1.
11252 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11254 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11255 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11257 *Joe Orton, Steve Henson*
11259 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11262 *Joe Orton, Steve Henson*
11264 * Make it possible to have multiple active certificates with the same
11265 subject in the CA index file. This is done only if the keyword
11266 'unique_subject' is set to 'no' in the main CA section (default
11267 if 'CA_default') of the configuration file. The value is saved
11268 with the database itself in a separate index attribute file,
11269 named like the index file with '.attr' appended to the name.
11273 * X509 verify fixes. Disable broken certificate workarounds when
11274 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11275 keyUsage extension present. Don't accept CRLs with unhandled critical
11276 extensions: since verify currently doesn't process CRL extensions this
11277 rejects a CRL with *any* critical extensions. Add new verify error codes
11282 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11283 A clarification of RFC2560 will require the use of OCTET STRINGs and
11284 some implementations cannot handle the current raw format. Since OpenSSL
11285 copies and compares OCSP nonces as opaque blobs without any attempt at
11286 parsing them this should not create any compatibility issues.
11290 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11291 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11292 this HMAC (and other) operations are several times slower than OpenSSL
11297 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11299 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11301 * Use the correct content when signing type "other".
11305 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11307 * Fix various bugs revealed by running the NISCC test suite:
11309 Stop out of bounds reads in the ASN1 code when presented with
11310 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11312 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11314 If verify callback ignores invalid public key errors don't try to check
11315 certificate signature with the NULL public key.
11319 * New -ignore_err option in ocsp application to stop the server
11320 exiting on the first error in a request.
11324 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11325 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11330 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11331 extra data after the compression methods not only for TLS 1.0
11332 but also for SSL 3.0 (as required by the specification).
11334 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11336 * Change X509_certificate_type() to mark the key as exported/exportable
11337 when it's 512 *bits* long, not 512 bytes.
11341 * Change AES_cbc_encrypt() so it outputs exact multiple of
11342 blocks during encryption.
11346 * Various fixes to base64 BIO and non blocking I/O. On write
11347 flushes were not handled properly if the BIO retried. On read
11348 data was not being buffered properly and had various logic bugs.
11349 This also affects blocking I/O when the data being decoded is a
11354 * Various S/MIME bugfixes and compatibility changes:
11355 output correct application/pkcs7 MIME type if
11356 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11357 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11358 of files as .eml work). Correctly handle very long lines in MIME
11363 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11365 * Countermeasure against the Klima-Pokorny-Rosa extension of
11366 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11367 a protocol version number mismatch like a decryption error
11368 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11372 * Turn on RSA blinding by default in the default implementation
11373 to avoid a timing attack. Applications that don't want it can call
11374 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11375 They would be ill-advised to do so in most cases.
11377 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11379 * Change RSA blinding code so that it works when the PRNG is not
11380 seeded (in this case, the secret RSA exponent is abused as
11381 an unpredictable seed -- if it is not unpredictable, there
11382 is no point in blinding anyway). Make RSA blinding thread-safe
11383 by remembering the creator's thread ID in rsa->blinding and
11384 having all other threads use local one-time blinding factors
11385 (this requires more computation than sharing rsa->blinding, but
11386 avoids excessive locking; and if an RSA object is not shared
11387 between threads, blinding will still be very fast).
11391 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11392 ENGINE as defaults for all supported algorithms irrespective of
11393 the 'flags' parameter. 'flags' is now honoured, so applications
11394 should make sure they are passing it correctly.
11398 * Target "mingw" now allows native Windows code to be generated in
11399 the Cygwin environment as well as with the MinGW compiler.
11403 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11405 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11406 via timing by performing a MAC computation even if incorrect
11407 block cipher padding has been found. This is a countermeasure
11408 against active attacks where the attacker has to distinguish
11409 between bad padding and a MAC verification error. ([CVE-2003-0078])
11411 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11412 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11413 Martin Vuagnoux (EPFL, Ilion)*
11415 * Make the no-err option work as intended. The intention with no-err
11416 is not to have the whole error stack handling routines removed from
11417 libcrypto, it's only intended to remove all the function name and
11418 reason texts, thereby removing some of the footprint that may not
11419 be interesting if those errors aren't displayed anyway.
11421 NOTE: it's still possible for any application or module to have its
11422 own set of error texts inserted. The routines are there, just not
11423 used by default when no-err is given.
11427 * Add support for FreeBSD on IA64.
11429 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11431 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11432 Kerberos function mit_des_cbc_cksum(). Before this change,
11433 the value returned by DES_cbc_cksum() was like the one from
11434 mit_des_cbc_cksum(), except the bytes were swapped.
11436 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11438 * Allow an application to disable the automatic SSL chain building.
11439 Before this a rather primitive chain build was always performed in
11440 ssl3_output_cert_chain(): an application had no way to send the
11441 correct chain if the automatic operation produced an incorrect result.
11443 Now the chain builder is disabled if either:
11445 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11447 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11449 The reasoning behind this is that an application would not want the
11450 auto chain building to take place if extra chain certificates are
11451 present and it might also want a means of sending no additional
11452 certificates (for example the chain has two certificates and the
11457 * Add the possibility to build without the ENGINE framework.
11459 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11461 * Under Win32 gmtime() can return NULL: check return value in
11462 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11466 * DSA routines: under certain error conditions uninitialized BN objects
11467 could be freed. Solution: make sure initialization is performed early
11468 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11469 Nils Larsch <nla@trustcenter.de> via PR#459)
11473 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11474 checked on reconnect on the client side, therefore session resumption
11475 could still fail with a "ssl session id is different" error. This
11476 behaviour is masked when SSL_OP_ALL is used due to
11477 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11478 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11479 followup to PR #377.
11483 * IA-32 assembler support enhancements: unified ELF targets, support
11484 for SCO/Caldera platforms, fix for Cygwin shared build.
11488 * Add support for FreeBSD on sparc64. As a consequence, support for
11489 FreeBSD on non-x86 processors is separate from x86 processors on
11490 the config script, much like the NetBSD support.
11492 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11494 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11496 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11499 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11500 code (06) was taken as the first octet of the session ID and the last
11501 octet was ignored consequently. As a result SSLv2 client side session
11502 caching could not have worked due to the session ID mismatch between
11504 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11509 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11510 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11515 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11516 seems that in spite of existing for more than a year, many application
11517 author have done nothing to provide the necessary callbacks, which
11518 means that this particular engine will not work properly anywhere.
11519 This is a very unfortunate situation which forces us, in the name
11520 of usability, to give the hw_ncipher.c a static lock, which is part
11522 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11523 appear in 0.9.8 or later. We EXPECT application authors to have
11524 dealt properly with this when 0.9.8 is released (unless we actually
11525 make such changes in the libcrypto locking code that changes will
11526 have to be made anyway).
11530 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11531 octets have been read, EOF or an error occurs. Without this change
11532 some truncated ASN1 structures will not produce an error.
11536 * Disable Heimdal support, since it hasn't been fully implemented.
11537 Still give the possibility to force the use of Heimdal, but with
11538 warnings and a request that patches get sent to openssl-dev.
11542 * Add the VC-CE target, introduce the WINCE sysname, and add
11543 INSTALL.WCE and appropriate conditionals to make it build.
11545 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11547 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11548 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11549 edit numbers of the version.
11551 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11553 * Introduce safe string copy and catenation functions
11554 (BUF_strlcpy() and BUF_strlcat()).
11556 *Ben Laurie (CHATS) and Richard Levitte*
11558 * Avoid using fixed-size buffers for one-line DNs.
11560 *Ben Laurie (CHATS)*
11562 * Add BUF_MEM_grow_clean() to avoid information leakage when
11563 resizing buffers containing secrets, and use where appropriate.
11565 *Ben Laurie (CHATS)*
11567 * Avoid using fixed size buffers for configuration file location.
11569 *Ben Laurie (CHATS)*
11571 * Avoid filename truncation for various CA files.
11573 *Ben Laurie (CHATS)*
11575 * Use sizeof in preference to magic numbers.
11577 *Ben Laurie (CHATS)*
11579 * Avoid filename truncation in cert requests.
11581 *Ben Laurie (CHATS)*
11583 * Add assertions to check for (supposedly impossible) buffer
11586 *Ben Laurie (CHATS)*
11588 * Don't cache truncated DNS entries in the local cache (this could
11589 potentially lead to a spoofing attack).
11591 *Ben Laurie (CHATS)*
11593 * Fix various buffers to be large enough for hex/decimal
11594 representations in a platform independent manner.
11596 *Ben Laurie (CHATS)*
11598 * Add CRYPTO_realloc_clean() to avoid information leakage when
11599 resizing buffers containing secrets, and use where appropriate.
11601 *Ben Laurie (CHATS)*
11603 * Add BIO_indent() to avoid much slightly worrying code to do
11606 *Ben Laurie (CHATS)*
11608 * Convert sprintf()/BIO_puts() to BIO_printf().
11610 *Ben Laurie (CHATS)*
11612 * buffer_gets() could terminate with the buffer only half
11615 *Ben Laurie (CHATS)*
11617 * Add assertions to prevent user-supplied crypto functions from
11618 overflowing internal buffers by having large block sizes, etc.
11620 *Ben Laurie (CHATS)*
11622 * New OPENSSL_assert() macro (similar to assert(), but enabled
11625 *Ben Laurie (CHATS)*
11627 * Eliminate unused copy of key in RC4.
11629 *Ben Laurie (CHATS)*
11631 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11633 *Ben Laurie (CHATS)*
11635 * Fix off-by-one error in EGD path.
11637 *Ben Laurie (CHATS)*
11639 * If RANDFILE path is too long, ignore instead of truncating.
11641 *Ben Laurie (CHATS)*
11643 * Eliminate unused and incorrectly sized X.509 structure
11646 *Ben Laurie (CHATS)*
11648 * Eliminate unused and dangerous function knumber().
11650 *Ben Laurie (CHATS)*
11652 * Eliminate unused and dangerous structure, KSSL_ERR.
11654 *Ben Laurie (CHATS)*
11656 * Protect against overlong session ID context length in an encoded
11657 session object. Since these are local, this does not appear to be
11660 *Ben Laurie (CHATS)*
11662 * Change from security patch (see 0.9.6e below) that did not affect
11663 the 0.9.6 release series:
11665 Remote buffer overflow in SSL3 protocol - an attacker could
11666 supply an oversized master key in Kerberos-enabled versions.
11669 *Ben Laurie (CHATS)*
11671 * Change the SSL kerb5 codes to match RFC 2712.
11675 * Make -nameopt work fully for req and add -reqopt switch.
11677 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11679 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11681 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11683 * Make sure tests can be performed even if the corresponding algorithms
11684 have been removed entirely. This was also the last step to make
11685 OpenSSL compilable with DJGPP under all reasonable conditions.
11687 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11689 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11690 to allow version independent disabling of normally unselected ciphers,
11691 which may be activated as a side-effect of selecting a single cipher.
11693 (E.g., cipher list string "RSA" enables ciphersuites that are left
11694 out of "ALL" because they do not provide symmetric encryption.
11695 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11697 *Lutz Jaenicke, Bodo Moeller*
11699 * Add appropriate support for separate platform-dependent build
11700 directories. The recommended way to make a platform-dependent
11701 build directory is the following (tested on Linux), maybe with
11704 # Place yourself outside of the OpenSSL source tree. In
11705 # this example, the environment variable OPENSSL_SOURCE
11706 # is assumed to contain the absolute OpenSSL source directory.
11707 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11708 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11709 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11710 mkdir -p `dirname $F`
11711 ln -s $OPENSSL_SOURCE/$F $F
11714 To be absolutely sure not to disturb the source tree, a "make clean"
11715 is a good thing. If it isn't successful, don't worry about it,
11716 it probably means the source directory is very clean.
11720 * Make sure any ENGINE control commands make local copies of string
11721 pointers passed to them whenever necessary. Otherwise it is possible
11722 the caller may have overwritten (or deallocated) the original string
11723 data when a later ENGINE operation tries to use the stored values.
11725 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11727 * Improve diagnostics in file reading and command-line digests.
11729 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11731 * Add AES modes CFB and OFB to the object database. Correct an
11732 error in AES-CFB decryption.
11736 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11737 allows existing EVP_CIPHER_CTX structures to be reused after
11738 calling `EVP_*Final()`. This behaviour is used by encryption
11739 BIOs and some applications. This has the side effect that
11740 applications must explicitly clean up cipher contexts with
11741 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11745 * Check the values of dna and dnb in bn_mul_recursive before calling
11746 bn_mul_comba (a non zero value means the a or b arrays do not contain
11747 n2 elements) and fallback to bn_mul_normal if either is not zero.
11751 * Fix escaping of non-ASCII characters when using the -subj option
11752 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11756 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11757 form for "surname", serialNumber has no short form.
11758 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11759 therefore remove "mail" short name for "internet 7".
11760 The OID for unique identifiers in X509 certificates is
11761 x500UniqueIdentifier, not uniqueIdentifier.
11762 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11766 * Add an "init" command to the ENGINE config module and auto initialize
11767 ENGINEs. Without any "init" command the ENGINE will be initialized
11768 after all ctrl commands have been executed on it. If init=1 the
11769 ENGINE is initialized at that point (ctrls before that point are run
11770 on the uninitialized ENGINE and after on the initialized one). If
11771 init=0 then the ENGINE will not be initialized at all.
11775 * Fix the 'app_verify_callback' interface so that the user-defined
11776 argument is actually passed to the callback: In the
11777 SSL_CTX_set_cert_verify_callback() prototype, the callback
11778 declaration has been changed from
11781 int (*cb)(X509_STORE_CTX *,void *);
11782 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11783 i=s->ctx->app_verify_callback(&ctx)
11784 has been changed into
11785 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11787 To update applications using SSL_CTX_set_cert_verify_callback(),
11788 a dummy argument can be added to their callback functions.
11790 *D. K. Smetters <smetters@parc.xerox.com>*
11792 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11794 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11796 * Add and OPENSSL_LOAD_CONF define which will cause
11797 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11798 This allows older applications to transparently support certain
11799 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11800 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11801 load the config file and OPENSSL_add_all_algorithms_conf() which will
11802 always load it have also been added.
11806 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11807 Adjust NIDs and EVP layer.
11809 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11811 * Config modules support in openssl utility.
11813 Most commands now load modules from the config file,
11814 though in a few (such as version) this isn't done
11815 because it couldn't be used for anything.
11817 In the case of ca and req the config file used is
11818 the same as the utility itself: that is the -config
11819 command line option can be used to specify an
11824 * Move default behaviour from OPENSSL_config(). If appname is NULL
11825 use "openssl_conf" if filename is NULL use default openssl config file.
11829 * Add an argument to OPENSSL_config() to allow the use of an alternative
11830 config section name. Add a new flag to tolerate a missing config file
11831 and move code to CONF_modules_load_file().
11835 * Support for crypto accelerator cards from Accelerated Encryption
11836 Processing, www.aep.ie. (Use engine 'aep')
11837 The support was copied from 0.9.6c [engine] and adapted/corrected
11838 to work with the new engine framework.
11840 *AEP Inc. and Richard Levitte*
11842 * Support for SureWare crypto accelerator cards from Baltimore
11843 Technologies. (Use engine 'sureware')
11844 The support was copied from 0.9.6c [engine] and adapted
11845 to work with the new engine framework.
11849 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11850 make the newer ENGINE framework commands for the CHIL engine work.
11852 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11854 * Make it possible to produce shared libraries on ReliantUNIX.
11856 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11858 * Add the configuration target debug-linux-ppro.
11859 Make 'openssl rsa' use the general key loading routines
11860 implemented in `apps.c`, and make those routines able to
11861 handle the key format FORMAT_NETSCAPE and the variant
11864 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11866 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11868 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11870 * Add -keyform to rsautl, and document -engine.
11872 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11874 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11875 BIO_R_NO_SUCH_FILE error code rather than the generic
11876 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11880 * Add new functions
11881 ERR_peek_last_error
11882 ERR_peek_last_error_line
11883 ERR_peek_last_error_line_data.
11884 These are similar to
11886 ERR_peek_error_line
11887 ERR_peek_error_line_data,
11888 but report on the latest error recorded rather than the first one
11889 still in the error queue.
11891 *Ben Laurie, Bodo Moeller*
11893 * default_algorithms option in ENGINE config module. This allows things
11895 default_algorithms = ALL
11896 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11900 * Preliminary ENGINE config module.
11904 * New experimental application configuration code.
11908 * Change the AES code to follow the same name structure as all other
11909 symmetric ciphers, and behave the same way. Move everything to
11910 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11912 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11914 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11916 *Ben Laurie and Theo de Raadt*
11918 * Add option to output public keys in req command.
11920 *Massimiliano Pala madwolf@openca.org*
11922 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11923 (up to about 10% better than before for P-192 and P-224).
11927 * New functions/macros
11929 SSL_CTX_set_msg_callback(ctx, cb)
11930 SSL_CTX_set_msg_callback_arg(ctx, arg)
11931 SSL_set_msg_callback(ssl, cb)
11932 SSL_set_msg_callback_arg(ssl, arg)
11934 to request calling a callback function
11936 void cb(int write_p, int version, int content_type,
11937 const void *buf, size_t len, SSL *ssl, void *arg)
11939 whenever a protocol message has been completely received
11940 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11941 protocol version according to which the SSL library interprets
11942 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11943 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11944 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11945 specification (change_cipher_spec(20), alert(21), handshake(22)).
11946 'buf' and 'len' point to the actual message, 'ssl' to the
11947 SSL object, and 'arg' is the application-defined value set by
11948 SSL[_CTX]_set_msg_callback_arg().
11950 'openssl s_client' and 'openssl s_server' have new '-msg' options
11951 to enable a callback that displays all protocol messages.
11955 * Change the shared library support so shared libraries are built as
11956 soon as the corresponding static library is finished, and thereby get
11957 openssl and the test programs linked against the shared library.
11958 This still only happens when the keyword "shard" has been given to
11959 the configuration scripts.
11961 NOTE: shared library support is still an experimental thing, and
11962 backward binary compatibility is still not guaranteed.
11964 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11966 * Add support for Subject Information Access extension.
11968 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11970 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11971 additional bytes when new memory had to be allocated, not just
11972 when reusing an existing buffer.
11976 * New command line and configuration option 'utf8' for the req command.
11977 This allows field values to be specified as UTF8 strings.
11981 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11982 runs for the former and machine-readable output for the latter.
11986 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11987 of the e-mail address in the DN (i.e., it will go into a certificate
11988 extension only). The new configuration file option 'email_in_dn = no'
11989 has the same effect.
11991 *Massimiliano Pala madwolf@openca.org*
11993 * Change all functions with names starting with `des_` to be starting
11994 with `DES_` instead. Add wrappers that are compatible with libdes,
11995 but are named `_ossl_old_des_*`. Finally, add macros that map the
11996 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11997 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11998 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12001 Since we provide two compatibility mappings, the user needs to
12002 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12003 compatibility is desired. The default (i.e., when that macro
12004 isn't defined) is OpenSSL 0.9.6c compatibility.
12006 There are also macros that enable and disable the support of old
12007 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12008 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12009 are defined, the default will apply: to support the old des routines.
12011 In either case, one must include openssl/des.h to get the correct
12012 definitions. Do not try to just include openssl/des_old.h, that
12015 NOTE: This is a major break of an old API into a new one. Software
12016 authors are encouraged to switch to the `DES_` style functions. Some
12017 time in the future, des_old.h and the libdes compatibility functions
12018 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12019 default), and then completely removed.
12023 * Test for certificates which contain unsupported critical extensions.
12024 If such a certificate is found during a verify operation it is
12025 rejected by default: this behaviour can be overridden by either
12026 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12027 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12028 X509_supported_extension() has also been added which returns 1 if a
12029 particular extension is supported.
12033 * Modify the behaviour of EVP cipher functions in similar way to digests
12034 to retain compatibility with existing code.
12038 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12039 compatibility with existing code. In particular the 'ctx' parameter does
12040 not have to be to be initialized before the call to EVP_DigestInit() and
12041 it is tidied up after a call to EVP_DigestFinal(). New function
12042 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12043 EVP_MD_CTX_copy() changed to not require the destination to be
12044 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12045 requires the destination to be valid.
12047 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12048 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12052 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12053 so that complete 'Handshake' protocol structures are kept in memory
12054 instead of overwriting 'msg_type' and 'length' with 'body' data.
12058 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12060 *Massimo Santin via Richard Levitte*
12062 * Major restructuring to the underlying ENGINE code. This includes
12063 reduction of linker bloat, separation of pure "ENGINE" manipulation
12064 (initialisation, etc) from functionality dealing with implementations
12065 of specific crypto interfaces. This change also introduces integrated
12066 support for symmetric ciphers and digest implementations - so ENGINEs
12067 can now accelerate these by providing EVP_CIPHER and EVP_MD
12068 implementations of their own. This is detailed in
12069 [crypto/engine/README.md](crypto/engine/README.md)
12070 as it couldn't be adequately described here. However, there are a few
12071 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12072 were changed in the original introduction of ENGINE code have now
12073 reverted back - the hooking from this code to ENGINE is now a good
12074 deal more passive and at run-time, operations deal directly with
12075 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12076 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12077 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12078 they were not being used by the framework as there is no concept of a
12079 BIGNUM_METHOD and they could not be generalised to the new
12080 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12081 ENGINE_cpy() has been removed as it cannot be consistently defined in
12086 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12090 * Change mkdef.pl to sort symbols that get the same entry number,
12091 and make sure the automatically generated functions `ERR_load_*`
12092 become part of libeay.num as well.
12096 * New function SSL_renegotiate_pending(). This returns true once
12097 renegotiation has been requested (either SSL_renegotiate() call
12098 or HelloRequest/ClientHello received from the peer) and becomes
12099 false once a handshake has been completed.
12100 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12101 sends a HelloRequest, but does not ensure that a handshake takes
12102 place. SSL_renegotiate_pending() is useful for checking if the
12103 client has followed the request.)
12107 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12108 By default, clients may request session resumption even during
12109 renegotiation (if session ID contexts permit); with this option,
12110 session resumption is possible only in the first handshake.
12112 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12113 more bits available for options that should not be part of
12114 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12118 * Add some demos for certificate and certificate request creation.
12122 * Make maximum certificate chain size accepted from the peer application
12123 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12124 "Douglas E. Engert" <deengert@anl.gov>.
12128 * Add support for shared libraries for Unixware-7
12129 (Boyd Lynn Gerber <gerberb@zenez.com>).
12133 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12134 be done prior to destruction. Use this to unload error strings from
12135 ENGINEs that load their own error strings. NB: This adds two new API
12136 functions to "get" and "set" this destroy handler in an ENGINE.
12140 * Alter all existing ENGINE implementations (except "openssl" and
12141 "openbsd") to dynamically instantiate their own error strings. This
12142 makes them more flexible to be built both as statically-linked ENGINEs
12143 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12144 Also, add stub code to each that makes building them as self-contained
12145 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12149 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12150 implementations into applications that are completely implemented in
12151 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12152 commands that can be used to configure what shared-library to load and
12153 to control aspects of the way it is handled. Also, made an update to
12154 the [README-Engine.md](README-Engine.md) file
12155 that brings its information up-to-date and
12156 provides some information and instructions on the "dynamic" ENGINE
12157 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12161 * Make it possible to unload ranges of ERR strings with a new
12162 "ERR_unload_strings" function.
12166 * Add a copy() function to EVP_MD.
12170 * Make EVP_MD routines take a context pointer instead of just the
12171 md_data void pointer.
12175 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12176 that the digest can only process a single chunk of data
12177 (typically because it is provided by a piece of
12178 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12179 is only going to provide a single chunk of data, and hence the
12180 framework needn't accumulate the data for oneshot drivers.
12184 * As with "ERR", make it possible to replace the underlying "ex_data"
12185 functions. This change also alters the storage and management of global
12186 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12187 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12188 index counters. The API functions that use this state have been changed
12189 to take a "class_index" rather than pointers to the class's local STACK
12190 and counter, and there is now an API function to dynamically create new
12191 classes. This centralisation allows us to (a) plug a lot of the
12192 thread-safety problems that existed, and (b) makes it possible to clean
12193 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12194 such data would previously have always leaked in application code and
12195 workarounds were in place to make the memory debugging turn a blind eye
12196 to it. Application code that doesn't use this new function will still
12197 leak as before, but their memory debugging output will announce it now
12198 rather than letting it slide.
12200 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12201 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12202 has a return value to indicate success or failure.
12206 * Make it possible to replace the underlying "ERR" functions such that the
12207 global state (2 LHASH tables and 2 locks) is only used by the "default"
12208 implementation. This change also adds two functions to "get" and "set"
12209 the implementation prior to it being automatically set the first time
12210 any other ERR function takes place. Ie. an application can call "get",
12211 pass the return value to a module it has just loaded, and that module
12212 can call its own "set" function using that value. This means the
12213 module's "ERR" operations will use (and modify) the error state in the
12214 application and not in its own statically linked copy of OpenSSL code.
12218 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12219 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12220 the operation, and provides a more encapsulated way for external code
12221 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12222 to use these functions rather than manually incrementing the counts.
12224 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12228 * Add EVP test program.
12232 * Add symmetric cipher support to ENGINE. Expect the API to change!
12236 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12237 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12238 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12239 These allow a CRL to be built without having to access X509_CRL fields
12240 directly. Modify 'ca' application to use new functions.
12244 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12245 bug workarounds. Rollback attack detection is a security feature.
12246 The problem will only arise on OpenSSL servers when TLSv1 is not
12247 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12248 Software authors not wanting to support TLSv1 will have special reasons
12249 for their choice and can explicitly enable this option.
12251 *Bodo Moeller, Lutz Jaenicke*
12253 * Rationalise EVP so it can be extended: don't include a union of
12254 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12255 (similar to those existing for EVP_CIPHER_CTX).
12260 EVP_MD_CTX_init(&md); /* new function call */
12261 EVP_DigestInit(&md, EVP_sha1());
12262 EVP_DigestUpdate(&md, in, len);
12263 EVP_DigestFinal(&md, out, NULL);
12264 EVP_MD_CTX_cleanup(&md); /* new function call */
12268 * Make DES key schedule conform to the usual scheme, as well as
12269 correcting its structure. This means that calls to DES functions
12270 now have to pass a pointer to a des_key_schedule instead of a
12271 plain des_key_schedule (which was actually always a pointer
12274 des_key_schedule ks;
12276 des_set_key_checked(..., &ks);
12277 des_ncbc_encrypt(..., &ks, ...);
12279 (Note that a later change renames 'des_...' into 'DES_...'.)
12283 * Initial reduction of linker bloat: the use of some functions, such as
12284 PEM causes large amounts of unused functions to be linked in due to
12285 poor organisation. For example pem_all.c contains every PEM function
12286 which has a knock on effect of linking in large amounts of (unused)
12287 ASN1 code. Grouping together similar functions and splitting unrelated
12288 functions prevents this.
12292 * Cleanup of EVP macros.
12296 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12297 correct `_ecb suffix`.
12301 * Add initial OCSP responder support to ocsp application. The
12302 revocation information is handled using the text based index
12303 use by the ca application. The responder can either handle
12304 requests generated internally, supplied in files (for example
12305 via a CGI script) or using an internal minimal server.
12309 * Add configuration choices to get zlib compression for TLS.
12313 * Changes to Kerberos SSL for RFC 2712 compliance:
12314 1. Implemented real KerberosWrapper, instead of just using
12315 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12316 2. Implemented optional authenticator field of KerberosWrapper.
12318 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12319 and authenticator structs; see crypto/krb5/.
12321 Generalized Kerberos calls to support multiple Kerberos libraries.
12322 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12323 via Richard Levitte*
12325 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12326 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12327 values for each of the key sizes rather than having just
12328 parameters (and 'speed' generating keys each time).
12332 * Speed up EVP routines.
12335 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12336 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12337 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12338 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12340 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12341 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12342 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12345 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12347 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12351 * Added the OS2-EMX target.
12353 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12355 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12356 New functions to support `NCONF `routines in extension code.
12357 New function `CONF_set_nconf()`
12358 to allow functions which take an `NCONF` to also handle the old `LHASH`
12359 structure: this means that the old `CONF` compatible routines can be
12360 retained (in particular w.rt. extensions) without having to duplicate the
12361 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12365 * Enhance the general user interface with mechanisms for inner control
12366 and with possibilities to have yes/no kind of prompts.
12370 * Change all calls to low level digest routines in the library and
12371 applications to use EVP. Add missing calls to HMAC_cleanup() and
12372 don't assume HMAC_CTX can be copied using memcpy().
12374 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12376 * Add the possibility to control engines through control names but with
12377 arbitrary arguments instead of just a string.
12378 Change the key loaders to take a UI_METHOD instead of a callback
12379 function pointer. NOTE: this breaks binary compatibility with earlier
12380 versions of OpenSSL [engine].
12381 Adapt the nCipher code for these new conditions and add a card insertion
12386 * Enhance the general user interface with mechanisms to better support
12387 dialog box interfaces, application-defined prompts, the possibility
12388 to use defaults (for example default passwords from somewhere else)
12389 and interrupts/cancellations.
12393 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12394 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12398 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12399 tidy up some unnecessarily weird code in 'sk_new()').
12401 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12403 * Change the key loading routines for ENGINEs to use the same kind
12404 callback (pem_password_cb) as all other routines that need this
12409 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12410 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12411 than this minimum value is recommended.
12415 * New random seeder for OpenVMS, using the system process statistics
12416 that are easily reachable.
12420 * Windows apparently can't transparently handle global
12421 variables defined in DLLs. Initialisations such as:
12423 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12425 won't compile. This is used by the any applications that need to
12426 declare their own ASN1 modules. This was fixed by adding the option
12427 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12428 needed for static libraries under Win32.
12432 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12433 setting of purpose and trust fields. New X509_STORE trust and
12434 purpose functions and tidy up setting in other SSL functions.
12438 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12439 structure. These are inherited by X509_STORE_CTX when it is
12440 initialised. This allows various defaults to be set in the
12441 X509_STORE structure (such as flags for CRL checking and custom
12442 purpose or trust settings) for functions which only use X509_STORE_CTX
12443 internally such as S/MIME.
12445 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12446 trust settings if they are not set in X509_STORE. This allows X509_STORE
12447 purposes and trust (in S/MIME for example) to override any set by default.
12449 Add command line options for CRL checking to smime, s_client and s_server
12454 * Initial CRL based revocation checking. If the CRL checking flag(s)
12455 are set then the CRL is looked up in the X509_STORE structure and
12456 its validity and signature checked, then if the certificate is found
12457 in the CRL the verify fails with a revoked error.
12459 Various new CRL related callbacks added to X509_STORE_CTX structure.
12461 Command line options added to 'verify' application to support this.
12463 This needs some additional work, such as being able to handle multiple
12464 CRLs with different times, extension based lookup (rather than just
12465 by subject name) and ultimately more complete V2 CRL extension
12470 * Add a general user interface API (crypto/ui/). This is designed
12471 to replace things like des_read_password and friends (backward
12472 compatibility functions using this new API are provided).
12473 The purpose is to remove prompting functions from the DES code
12474 section as well as provide for prompting through dialog boxes in
12475 a window system and the like.
12479 * Add "ex_data" support to ENGINE so implementations can add state at a
12480 per-structure level rather than having to store it globally.
12484 * Make it possible for ENGINE structures to be copied when retrieved by
12485 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12486 This causes the "original" ENGINE structure to act like a template,
12487 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12488 operational state can be localised to each ENGINE structure, despite the
12489 fact they all share the same "methods". New ENGINE structures returned in
12490 this case have no functional references and the return value is the single
12491 structural reference. This matches the single structural reference returned
12492 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12497 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12498 needs to match any other type at all we need to manually clear the
12503 * Changes to the "openssl engine" utility to include;
12504 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12505 about an ENGINE's available control commands.
12506 - executing control commands from command line arguments using the
12507 '-pre' and '-post' switches. '-post' is only used if '-t' is
12508 specified and the ENGINE is successfully initialised. The syntax for
12509 the individual commands are colon-separated, for example;
12510 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12514 * New dynamic control command support for ENGINEs. ENGINEs can now
12515 declare their own commands (numbers), names (strings), descriptions,
12516 and input types for run-time discovery by calling applications. A
12517 subset of these commands are implicitly classed as "executable"
12518 depending on their input type, and only these can be invoked through
12519 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12520 can be based on user input, config files, etc). The distinction is
12521 that "executable" commands cannot return anything other than a boolean
12522 result and can only support numeric or string input, whereas some
12523 discoverable commands may only be for direct use through
12524 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12525 pointers, or other custom uses. The "executable" commands are to
12526 support parameterisations of ENGINE behaviour that can be
12527 unambiguously defined by ENGINEs and used consistently across any
12528 OpenSSL-based application. Commands have been added to all the
12529 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12530 control over shared-library paths without source code alterations.
12534 * Changed all ENGINE implementations to dynamically allocate their
12535 ENGINEs rather than declaring them statically. Apart from this being
12536 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12537 this also allows the implementations to compile without using the
12538 internal engine_int.h header.
12542 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12543 'const' value. Any code that should be able to modify a RAND_METHOD
12544 should already have non-const pointers to it (ie. they should only
12545 modify their own ones).
12549 * Made a variety of little tweaks to the ENGINE code.
12550 - "atalla" and "ubsec" string definitions were moved from header files
12551 to C code. "nuron" string definitions were placed in variables
12552 rather than hard-coded - allowing parameterisation of these values
12553 later on via ctrl() commands.
12554 - Removed unused "#if 0"'d code.
12555 - Fixed engine list iteration code so it uses ENGINE_free() to release
12556 structural references.
12557 - Constified the RAND_METHOD element of ENGINE structures.
12558 - Constified various get/set functions as appropriate and added
12559 missing functions (including a catch-all ENGINE_cpy that duplicates
12560 all ENGINE values onto a new ENGINE except reference counts/state).
12561 - Removed NULL parameter checks in get/set functions. Setting a method
12562 or function to NULL is a way of cancelling out a previously set
12563 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12564 and doesn't justify the extra error symbols and code.
12565 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12566 flags from engine_int.h to engine.h.
12567 - Changed prototypes for ENGINE handler functions (init(), finish(),
12568 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12572 * Implement binary inversion algorithm for BN_mod_inverse in addition
12573 to the algorithm using long division. The binary algorithm can be
12574 used only if the modulus is odd. On 32-bit systems, it is faster
12575 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12576 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12577 up to 450 bits. In 64-bit environments, the binary algorithm
12578 appears to be advantageous for much longer moduli; here we use it
12579 for moduli up to 2048 bits.
12583 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12584 could not support the combine flag in choice fields.
12588 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12589 extensions from a certificate request to the certificate.
12593 * Allow multiple 'certopt' and 'nameopt' options to be separated
12594 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12595 file: this allows the display of the certificate about to be
12596 signed to be customised, to allow certain fields to be included
12597 or excluded and extension details. The old system didn't display
12598 multicharacter strings properly, omitted fields not in the policy
12599 and couldn't display additional details such as extensions.
12603 * Function EC_POINTs_mul for multiple scalar multiplication
12604 of an arbitrary number of elliptic curve points
12605 \sum scalars[i]*points[i],
12606 optionally including the generator defined for the EC_GROUP:
12607 scalar*generator + \sum scalars[i]*points[i].
12609 EC_POINT_mul is a simple wrapper function for the typical case
12610 that the point list has just one item (besides the optional
12615 * First EC_METHODs for curves over GF(p):
12617 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12618 operations and provides various method functions that can also
12619 operate with faster implementations of modular arithmetic.
12621 EC_GFp_mont_method() reuses most functions that are part of
12622 EC_GFp_simple_method, but uses Montgomery arithmetic.
12624 *Bodo Moeller; point addition and point doubling
12625 implementation directly derived from source code provided by
12626 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12628 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12629 crypto/ec/ec_lib.c):
12631 Curves are EC_GROUP objects (with an optional group generator)
12632 based on EC_METHODs that are built into the library.
12634 Points are EC_POINT objects based on EC_GROUP objects.
12636 Most of the framework would be able to handle curves over arbitrary
12637 finite fields, but as there are no obvious types for fields other
12638 than GF(p), some functions are limited to that for now.
12642 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12643 that the file contains a complete HTTP response.
12647 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12648 change the def and num file printf format specifier from "%-40sXXX"
12649 to "%-39s XXX". The latter will always guarantee a space after the
12650 field while the former will cause them to run together if the field
12651 is 40 of more characters long.
12655 * Constify the cipher and digest 'method' functions and structures
12656 and modify related functions to take constant EVP_MD and EVP_CIPHER
12661 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12662 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12666 * Modify `EVP_Digest*()` routines so they now return values. Although the
12667 internal software routines can never fail additional hardware versions
12672 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12674 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12675 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12678 ERR_R_NESTED_ASN1_ERROR
12680 ERR_R_MISSING_ASN1_EOS
12681 were 4 .. 9, conflicting with
12682 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12684 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12685 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12687 Add new error code 'ERR_R_INTERNAL_ERROR'.
12691 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12696 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12697 sets the subject name for a new request or supersedes the
12698 subject name in a given request. Formats that can be parsed are
12699 'CN=Some Name, OU=myOU, C=IT'
12701 'CN=Some Name/OU=myOU/C=IT'.
12703 Add options '-batch' and '-verbose' to 'openssl req'.
12705 *Massimiliano Pala <madwolf@hackmasters.net>*
12707 * Introduce the possibility to access global variables through
12708 functions on platform were that's the best way to handle exporting
12709 global variables in shared libraries. To enable this functionality,
12710 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12711 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12712 is normally done by Configure or something similar).
12714 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12715 in the source file (foo.c) like this:
12717 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12718 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12720 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12721 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12723 OPENSSL_DECLARE_GLOBAL(int,foo);
12724 #define foo OPENSSL_GLOBAL_REF(foo)
12725 OPENSSL_DECLARE_GLOBAL(double,bar);
12726 #define bar OPENSSL_GLOBAL_REF(bar)
12728 The #defines are very important, and therefore so is including the
12729 header file everywhere where the defined globals are used.
12731 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12732 of ASN.1 items, but that structure is a bit different.
12734 The largest change is in util/mkdef.pl which has been enhanced with
12735 better and easier to understand logic to choose which symbols should
12736 go into the Windows .def files as well as a number of fixes and code
12737 cleanup (among others, algorithm keywords are now sorted
12738 lexicographically to avoid constant rewrites).
12742 * In BN_div() keep a copy of the sign of 'num' before writing the
12743 result to 'rm' because if rm==num the value will be overwritten
12744 and produce the wrong result if 'num' is negative: this caused
12745 problems with BN_mod() and BN_nnmod().
12749 * Function OCSP_request_verify(). This checks the signature on an
12750 OCSP request and verifies the signer certificate. The signer
12751 certificate is just checked for a generic purpose and OCSP request
12756 * Add OCSP_check_validity() function to check the validity of OCSP
12757 responses. OCSP responses are prepared in real time and may only
12758 be a few seconds old. Simply checking that the current time lies
12759 between thisUpdate and nextUpdate max reject otherwise valid responses
12760 caused by either OCSP responder or client clock inaccuracy. Instead
12761 we allow thisUpdate and nextUpdate to fall within a certain period of
12762 the current time. The age of the response can also optionally be
12763 checked. Two new options -validity_period and -status_age added to
12768 * If signature or public key algorithm is unrecognized print out its
12769 OID rather that just UNKNOWN.
12773 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12774 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12775 ID to be generated from the issuer certificate alone which can then be
12776 passed to OCSP_id_issuer_cmp().
12780 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12781 ASN1 modules to export functions returning ASN1_ITEM pointers
12782 instead of the ASN1_ITEM structures themselves. This adds several
12783 new macros which allow the underlying ASN1 function/structure to
12784 be accessed transparently. As a result code should not use ASN1_ITEM
12785 references directly (such as &X509_it) but instead use the relevant
12786 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12787 use of the new ASN1 code on platforms where exporting structures
12788 is problematical (for example in shared libraries) but exporting
12789 functions returning pointers to structures is not.
12793 * Add support for overriding the generation of SSL/TLS session IDs.
12794 These callbacks can be registered either in an SSL_CTX or per SSL.
12795 The purpose of this is to allow applications to control, if they wish,
12796 the arbitrary values chosen for use as session IDs, particularly as it
12797 can be useful for session caching in multiple-server environments. A
12798 command-line switch for testing this (and any client code that wishes
12799 to use such a feature) has been added to "s_server".
12801 *Geoff Thorpe, Lutz Jaenicke*
12803 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12804 of the form `#if defined(...) || defined(...) || ...` and
12805 `#if !defined(...) && !defined(...) && ...`. This also avoids
12806 the growing number of special cases it was previously handling.
12810 * Make all configuration macros available for application by making
12811 sure they are available in opensslconf.h, by giving them names starting
12812 with `OPENSSL_` to avoid conflicts with other packages and by making
12813 sure e_os2.h will cover all platform-specific cases together with
12815 Additionally, it is now possible to define configuration/platform-
12816 specific names (called "system identities"). In the C code, these
12817 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12818 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12819 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12824 * New option -set_serial to 'req' and 'x509' this allows the serial
12825 number to use to be specified on the command line. Previously self
12826 signed certificates were hard coded with serial number 0 and the
12827 CA options of 'x509' had to use a serial number in a file which was
12832 * New options to 'ca' utility to support V2 CRL entry extensions.
12833 Currently CRL reason, invalidity date and hold instruction are
12834 supported. Add new CRL extensions to V3 code and some new objects.
12838 * New function EVP_CIPHER_CTX_set_padding() this is used to
12839 disable standard block padding (aka PKCS#5 padding) in the EVP
12840 API, which was previously mandatory. This means that the data is
12841 not padded in any way and so the total length much be a multiple
12842 of the block size, otherwise an error occurs.
12846 * Initial (incomplete) OCSP SSL support.
12850 * New function OCSP_parse_url(). This splits up a URL into its host,
12851 port and path components: primarily to parse OCSP URLs. New -url
12852 option to ocsp utility.
12856 * New nonce behavior. The return value of OCSP_check_nonce() now
12857 reflects the various checks performed. Applications can decide
12858 whether to tolerate certain situations such as an absent nonce
12859 in a response when one was present in a request: the ocsp application
12860 just prints out a warning. New function OCSP_add1_basic_nonce()
12861 this is to allow responders to include a nonce in a response even if
12862 the request is nonce-less.
12866 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12867 skipped when using openssl x509 multiple times on a single input file,
12868 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12872 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12873 set string type: to handle setting ASN1_TIME structures. Fix ca
12874 utility to correctly initialize revocation date of CRLs.
12878 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12879 the clients preferred ciphersuites and rather use its own preferences.
12880 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12881 Internet Explorer by ensuring unchanged hash method during stepup.
12882 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12886 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12887 to aes and add a new 'exist' option to print out symbols that don't
12892 * Additional options to ocsp utility to allow flags to be set and
12893 additional certificates supplied.
12897 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12898 OCSP client a number of certificate to only verify the response
12903 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12904 handle the new API. Currently only ECB, CBC modes supported. Add new
12907 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12908 Encryption Standard (AES) Ciphersuites for Transport Layer
12909 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12910 not enabled by default and were not part of the "ALL" ciphersuite
12911 alias because they were not yet official; they could be
12912 explicitly requested by specifying the "AESdraft" ciphersuite
12913 group alias. In the final release of OpenSSL 0.9.7, the group
12914 alias is called "AES" and is part of "ALL".)
12916 *Ben Laurie, Steve Henson, Bodo Moeller*
12918 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12919 request to response.
12923 * Functions for OCSP responders. OCSP_request_onereq_count(),
12924 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12925 extract information from a certificate request. OCSP_response_create()
12926 creates a response and optionally adds a basic response structure.
12927 OCSP_basic_add1_status() adds a complete single response to a basic
12928 response and returns the OCSP_SINGLERESP structure just added (to allow
12929 extensions to be included for example). OCSP_basic_add1_cert() adds a
12930 certificate to a basic response and OCSP_basic_sign() signs a basic
12931 response with various flags. New helper functions ASN1_TIME_check()
12932 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12933 (converts ASN1_TIME to GeneralizedTime).
12937 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12938 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12939 structure from a certificate. X509_pubkey_digest() digests the public_key
12940 contents: this is used in various key identifiers.
12944 * Make sk_sort() tolerate a NULL argument.
12946 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12948 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12949 passed by the function are trusted implicitly. If any of them signed the
12950 response then it is assumed to be valid and is not verified.
12954 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12955 to data. This was previously part of the PKCS7 ASN1 code. This
12956 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12957 *Steve Henson, reported by Kenneth R. Robinette
12958 <support@securenetterm.com>*
12960 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12961 routines: without these tracing memory leaks is very painful.
12962 Fix leaks in PKCS12 and PKCS7 routines.
12966 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12967 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12968 effectively meant GeneralizedTime would never be used. Now it
12969 is initialised to -1 but X509_time_adj() now has to check the value
12970 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12971 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12972 *Steve Henson, reported by Kenneth R. Robinette
12973 <support@securenetterm.com>*
12975 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12976 result in a zero length in the ASN1_INTEGER structure which was
12977 not consistent with the structure when d2i_ASN1_INTEGER() was used
12978 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12979 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12980 where it did not print out a minus for negative ASN1_INTEGER.
12984 * Add summary printout to ocsp utility. The various functions which
12985 convert status values to strings have been renamed to:
12986 OCSP_response_status_str(), OCSP_cert_status_str() and
12987 OCSP_crl_reason_str() and are no longer static. New options
12988 to verify nonce values and to disable verification. OCSP response
12989 printout format cleaned up.
12993 * Add additional OCSP certificate checks. These are those specified
12994 in RFC2560. This consists of two separate checks: the CA of the
12995 certificate being checked must either be the OCSP signer certificate
12996 or the issuer of the OCSP signer certificate. In the latter case the
12997 OCSP signer certificate must contain the OCSP signing extended key
12998 usage. This check is performed by attempting to match the OCSP
12999 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13000 in the OCSP_CERTID structures of the response.
13004 * Initial OCSP certificate verification added to OCSP_basic_verify()
13005 and related routines. This uses the standard OpenSSL certificate
13006 verify routines to perform initial checks (just CA validity) and
13007 to obtain the certificate chain. Then additional checks will be
13008 performed on the chain. Currently the root CA is checked to see
13009 if it is explicitly trusted for OCSP signing. This is used to set
13010 a root CA as a global signing root: that is any certificate that
13011 chains to that CA is an acceptable OCSP signing certificate.
13015 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13016 extensions from a separate configuration file.
13017 As when reading extensions from the main configuration file,
13018 the '-extensions ...' option may be used for specifying the
13021 *Massimiliano Pala <madwolf@comune.modena.it>*
13023 * New OCSP utility. Allows OCSP requests to be generated or
13024 read. The request can be sent to a responder and the output
13025 parsed, outputted or printed in text form. Not complete yet:
13026 still needs to check the OCSP response validity.
13030 * New subcommands for 'openssl ca':
13031 `openssl ca -status <serial>` prints the status of the cert with
13032 the given serial number (according to the index file).
13033 `openssl ca -updatedb` updates the expiry status of certificates
13036 *Massimiliano Pala <madwolf@comune.modena.it>*
13038 * New '-newreq-nodes' command option to CA.pl. This is like
13039 '-newreq', but calls 'openssl req' with the '-nodes' option
13040 so that the resulting key is not encrypted.
13042 *Damien Miller <djm@mindrot.org>*
13044 * New configuration for the GNU Hurd.
13046 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13048 * Initial code to implement OCSP basic response verify. This
13049 is currently incomplete. Currently just finds the signer's
13050 certificate and verifies the signature on the response.
13054 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13055 value of OPENSSLDIR. This is available via the new '-d' option
13056 to 'openssl version', and is also included in 'openssl version -a'.
13060 * Allowing defining memory allocation callbacks that will be given
13061 file name and line number information in additional arguments
13062 (a `const char*` and an int). The basic functionality remains, as
13063 well as the original possibility to just replace malloc(),
13064 realloc() and free() by functions that do not know about these
13065 additional arguments. To register and find out the current
13066 settings for extended allocation functions, the following
13067 functions are provided:
13069 CRYPTO_set_mem_ex_functions
13070 CRYPTO_set_locked_mem_ex_functions
13071 CRYPTO_get_mem_ex_functions
13072 CRYPTO_get_locked_mem_ex_functions
13074 These work the same way as CRYPTO_set_mem_functions and friends.
13075 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13076 extended allocation function is enabled.
13077 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13078 a conventional allocation function is enabled.
13080 *Richard Levitte, Bodo Moeller*
13082 * Finish off removing the remaining LHASH function pointer casts.
13083 There should no longer be any prototype-casting required when using
13084 the LHASH abstraction, and any casts that remain are "bugs". See
13085 the callback types and macros at the head of lhash.h for details
13086 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13090 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13091 If /dev/[u]random devices are not available or do not return enough
13092 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13094 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13095 /etc/entropy will be queried once each in this sequence, querying stops
13096 when enough entropy was collected without querying more sockets.
13100 * Change the Unix RAND_poll() variant to be able to poll several
13101 random devices, as specified by DEVRANDOM, until a sufficient amount
13102 of data has been collected. We spend at most 10 ms on each file
13103 (select timeout) and read in non-blocking mode. DEVRANDOM now
13104 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13105 (previously it was just the string "/dev/urandom"), so on typical
13106 platforms the 10 ms delay will never occur.
13107 Also separate out the Unix variant to its own file, rand_unix.c.
13108 For VMS, there's a currently-empty rand_vms.c.
13112 * Move OCSP client related routines to ocsp_cl.c. These
13113 provide utility functions which an application needing
13114 to issue a request to an OCSP responder and analyse the
13115 response will typically need: as opposed to those which an
13116 OCSP responder itself would need which will be added later.
13118 OCSP_request_sign() signs an OCSP request with an API similar
13119 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13120 response. OCSP_response_get1_basic() extracts basic response
13121 from response. OCSP_resp_find_status(): finds and extracts status
13122 information from an OCSP_CERTID structure (which will be created
13123 when the request structure is built). These are built from lower
13124 level functions which work on OCSP_SINGLERESP structures but
13125 won't normally be used unless the application wishes to examine
13126 extensions in the OCSP response for example.
13128 Replace nonce routines with a pair of functions.
13129 OCSP_request_add1_nonce() adds a nonce value and optionally
13130 generates a random value. OCSP_check_nonce() checks the
13131 validity of the nonce in an OCSP response.
13135 * Change function OCSP_request_add() to OCSP_request_add0_id().
13136 This doesn't copy the supplied OCSP_CERTID and avoids the
13137 need to free up the newly created id. Change return type
13138 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13139 This can then be used to add extensions to the request.
13140 Deleted OCSP_request_new(), since most of its functionality
13141 is now in OCSP_REQUEST_new() (and the case insensitive name
13142 clash) apart from the ability to set the request name which
13143 will be added elsewhere.
13147 * Update OCSP API. Remove obsolete extensions argument from
13148 various functions. Extensions are now handled using the new
13149 OCSP extension code. New simple OCSP HTTP function which
13150 can be used to send requests and parse the response.
13154 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13155 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13156 uses the special reorder version of SET OF to sort the attributes
13157 and reorder them to match the encoded order. This resolves a long
13158 standing problem: a verify on a PKCS7 structure just after signing
13159 it used to fail because the attribute order did not match the
13160 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13161 it uses the received order. This is necessary to tolerate some broken
13162 software that does not order SET OF. This is handled by encoding
13163 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13164 to produce the required SET OF.
13168 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13169 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13170 files to get correct declarations of the ASN.1 item variables.
13174 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13175 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13176 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13177 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13178 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13179 ASN1_ITEM and no wrapper functions.
13183 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13184 replace the old function pointer based I/O routines. Change most of
13185 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13189 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13190 lines, recognize more "algorithms" that can be deselected, and make
13191 it complain about algorithm deselection that isn't recognised.
13195 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13196 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13197 to use new functions. Add NO_ASN1_OLD which can be set to remove
13198 some old style ASN1 functions: this can be used to determine if old
13199 code will still work when these eventually go away.
13203 * New extension functions for OCSP structures, these follow the
13204 same conventions as certificates and CRLs.
13208 * New function X509V3_add1_i2d(). This automatically encodes and
13209 adds an extension. Its behaviour can be customised with various
13210 flags to append, replace or delete. Various wrappers added for
13211 certificates and CRLs.
13215 * Fix to avoid calling the underlying ASN1 print routine when
13216 an extension cannot be parsed. Correct a typo in the
13217 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13221 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13222 entries for variables.
13226 * Add functionality to `apps/openssl.c` for detecting locking
13227 problems: As the program is single-threaded, all we have
13228 to do is register a locking callback using an array for
13229 storing which locks are currently held by the program.
13233 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13234 SSL_get_ex_data_X509_STORE_idx(), which is used in
13235 ssl_verify_cert_chain() and thus can be called at any time
13236 during TLS/SSL handshakes so that thread-safety is essential.
13237 Unfortunately, the ex_data design is not at all suited
13238 for multi-threaded use, so it probably should be abolished.
13242 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13244 *Broadcom, tweaked and integrated by Geoff Thorpe*
13246 * Move common extension printing code to new function
13247 X509V3_print_extensions(). Reorganise OCSP print routines and
13248 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13252 * New function X509_signature_print() to remove duplication in some
13257 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13258 set (this was treated exactly the same as SET OF previously). This
13259 is used to reorder the STACK representing the structure to match the
13260 encoding. This will be used to get round a problem where a PKCS7
13261 structure which was signed could not be verified because the STACK
13262 order did not reflect the encoded order.
13266 * Reimplement the OCSP ASN1 module using the new code.
13270 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13271 for its ASN1 operations. The old style function pointers still exist
13272 for now but they will eventually go away.
13276 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13277 completely replaces the old ASN1 functionality with a table driven
13278 encoder and decoder which interprets an ASN1_ITEM structure describing
13279 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13280 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13281 has also been converted to the new form.
13285 * Change BN_mod_exp_recp so that negative moduli are tolerated
13286 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13287 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13288 for negative moduli.
13292 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13293 of not touching the result's sign bit.
13297 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13302 * Changed the LHASH code to use prototypes for callbacks, and created
13303 macros to declare and implement thin (optionally static) functions
13304 that provide type-safety and avoid function pointer casting for the
13305 type-specific callbacks.
13309 * Added Kerberos Cipher Suites to be used with TLS, as written in
13311 *Veers Staats <staatsvr@asc.hpc.mil>,
13312 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13314 * Reformat the FAQ so the different questions and answers can be divided
13315 in sections depending on the subject.
13319 * Have the zlib compression code load ZLIB.DLL dynamically under
13324 * New function BN_mod_sqrt for computing square roots modulo a prime
13325 (using the probabilistic Tonelli-Shanks algorithm unless
13326 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13327 be handled deterministically).
13329 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13331 * Make BN_mod_inverse faster by explicitly handling small quotients
13332 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13333 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13337 * New function BN_kronecker.
13341 * Fix BN_gcd so that it works on negative inputs; the result is
13342 positive unless both parameters are zero.
13343 Previously something reasonably close to an infinite loop was
13344 possible because numbers could be growing instead of shrinking
13345 in the implementation of Euclid's algorithm.
13349 * Fix BN_is_word() and BN_is_one() macros to take into account the
13350 sign of the number in question.
13352 Fix BN_is_word(a,w) to work correctly for w == 0.
13354 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13355 because its test if the absolute value of 'a' equals 'w'.
13356 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13357 it exists mostly for use in the implementations of BN_is_zero(),
13358 BN_is_one(), and BN_is_word().
13362 * New function BN_swap.
13366 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13367 the exponentiation functions are more likely to produce reasonable
13368 results on negative inputs.
13372 * Change BN_mod_mul so that the result is always non-negative.
13373 Previously, it could be negative if one of the factors was negative;
13374 I don't think anyone really wanted that behaviour.
13378 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13379 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13380 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13381 and add new functions:
13390 BN_mod_lshift1_quick
13392 BN_mod_lshift_quick
13394 These functions always generate non-negative results.
13396 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13397 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13399 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13400 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13401 be reduced modulo `m`.
13403 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13406 The following entry accidentally appeared in the CHANGES file
13407 distributed with OpenSSL 0.9.7. The modifications described in
13408 it do *not* apply to OpenSSL 0.9.7.
13410 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13411 was actually never needed) and in BN_mul(). The removal in BN_mul()
13412 required a small change in bn_mul_part_recursive() and the addition
13413 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13414 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13415 bn_sub_words() and bn_add_words() except they take arrays with
13421 * In 'openssl passwd', verify passwords read from the terminal
13422 unless the '-salt' option is used (which usually means that
13423 verification would just waste user's time since the resulting
13424 hash is going to be compared with some given password hash)
13425 or the new '-noverify' option is used.
13427 This is an incompatible change, but it does not affect
13428 non-interactive use of 'openssl passwd' (passwords on the command
13429 line, '-stdin' option, '-in ...' option) and thus should not
13430 cause any problems.
13434 * Remove all references to RSAref, since there's no more need for it.
13438 * Make DSO load along a path given through an environment variable
13439 (SHLIB_PATH) with shl_load().
13443 * Constify the ENGINE code as a result of BIGNUM constification.
13444 Also constify the RSA code and most things related to it. In a
13445 few places, most notable in the depth of the ASN.1 code, ugly
13446 casts back to non-const were required (to be solved at a later
13451 * Make it so the openssl application has all engines loaded by default.
13455 * Constify the BIGNUM routines a little more.
13459 * Add the following functions:
13461 ENGINE_load_cswift()
13463 ENGINE_load_atalla()
13464 ENGINE_load_nuron()
13465 ENGINE_load_builtin_engines()
13467 That way, an application can itself choose if external engines that
13468 are built-in in OpenSSL shall ever be used or not. The benefit is
13469 that applications won't have to be linked with libdl or other dso
13470 libraries unless it's really needed.
13472 Changed 'openssl engine' to load all engines on demand.
13473 Changed the engine header files to avoid the duplication of some
13474 declarations (they differed!).
13478 * 'openssl engine' can now list capabilities.
13482 * Better error reporting in 'openssl engine'.
13486 * Never call load_dh_param(NULL) in s_server.
13490 * Add engine application. It can currently list engines by name and
13491 identity, and test if they are actually available.
13495 * Improve RPM specification file by forcing symbolic linking and making
13496 sure the installed documentation is also owned by root.root.
13498 *Damien Miller <djm@mindrot.org>*
13500 * Give the OpenSSL applications more possibilities to make use of
13501 keys (public as well as private) handled by engines.
13505 * Add OCSP code that comes from CertCo.
13509 * Add VMS support for the Rijndael code.
13513 * Added untested support for Nuron crypto accelerator.
13517 * Add support for external cryptographic devices. This code was
13518 previously distributed separately as the "engine" branch.
13520 *Geoff Thorpe, Richard Levitte*
13522 * Rework the filename-translation in the DSO code. It is now possible to
13523 have far greater control over how a "name" is turned into a filename
13524 depending on the operating environment and any oddities about the
13525 different shared library filenames on each system.
13529 * Support threads on FreeBSD-elf in Configure.
13533 * Fix for SHA1 assembly problem with MASM: it produces
13534 warnings about corrupt line number information when assembling
13535 with debugging information. This is caused by the overlapping
13538 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13541 NCONF_get_number() has no error checking at all. As a replacement,
13542 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13543 promoted strongly. The old NCONF_get_number is kept around for
13544 binary backward compatibility.
13545 Make it possible for methods to load from something other than a BIO,
13546 by providing a function pointer that is given a name instead of a BIO.
13547 For example, this could be used to load configuration data from an
13552 * Fix for non blocking accept BIOs. Added new I/O special reason
13553 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13554 with non blocking I/O was not possible because no retry code was
13555 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13560 * Added the beginnings of Rijndael support.
13564 * Fix for bug in DirectoryString mask setting. Add support for
13565 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13566 to allow certificate printing to more controllable, additional
13567 'certopt' option to 'x509' to allow new printing options to be
13572 * Clean old EAY MD5 hack from e_os.h.
13576 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13578 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13579 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13581 *Joe Orton, Steve Henson*
13583 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13585 * Fix additional bug revealed by the NISCC test suite:
13587 Stop bug triggering large recursion when presented with
13588 certain ASN.1 tags ([CVE-2003-0851])
13592 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13594 * Fix various bugs revealed by running the NISCC test suite:
13596 Stop out of bounds reads in the ASN1 code when presented with
13597 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13599 If verify callback ignores invalid public key errors don't try to check
13600 certificate signature with the NULL public key.
13604 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13605 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13610 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13611 extra data after the compression methods not only for TLS 1.0
13612 but also for SSL 3.0 (as required by the specification).
13614 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13616 * Change X509_certificate_type() to mark the key as exported/exportable
13617 when it's 512 *bits* long, not 512 bytes.
13621 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13623 * Countermeasure against the Klima-Pokorny-Rosa extension of
13624 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13625 a protocol version number mismatch like a decryption error
13626 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13630 * Turn on RSA blinding by default in the default implementation
13631 to avoid a timing attack. Applications that don't want it can call
13632 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13633 They would be ill-advised to do so in most cases.
13635 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13637 * Change RSA blinding code so that it works when the PRNG is not
13638 seeded (in this case, the secret RSA exponent is abused as
13639 an unpredictable seed -- if it is not unpredictable, there
13640 is no point in blinding anyway). Make RSA blinding thread-safe
13641 by remembering the creator's thread ID in rsa->blinding and
13642 having all other threads use local one-time blinding factors
13643 (this requires more computation than sharing rsa->blinding, but
13644 avoids excessive locking; and if an RSA object is not shared
13645 between threads, blinding will still be very fast).
13649 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13651 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13652 via timing by performing a MAC computation even if incorrect
13653 block cipher padding has been found. This is a countermeasure
13654 against active attacks where the attacker has to distinguish
13655 between bad padding and a MAC verification error. ([CVE-2003-0078])
13657 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13658 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13659 Martin Vuagnoux (EPFL, Ilion)*
13661 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13663 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13664 memory from its contents. This is done with a counter that will
13665 place alternating values in each byte. This can be used to solve
13666 two issues: 1) the removal of calls to memset() by highly optimizing
13667 compilers, and 2) cleansing with other values than 0, since those can
13668 be read through on certain media, for example a swap space on disk.
13672 * Bugfix: client side session caching did not work with external caching,
13673 because the session->cipher setting was not restored when reloading
13674 from the external cache. This problem was masked, when
13675 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13676 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13680 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13681 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13683 *Zeev Lieber <zeev-l@yahoo.com>*
13685 * Undo an undocumented change introduced in 0.9.6e which caused
13686 repeated calls to OpenSSL_add_all_ciphers() and
13687 OpenSSL_add_all_digests() to be ignored, even after calling
13692 * Change the default configuration reader to deal with last line not
13693 being properly terminated.
13697 * Change X509_NAME_cmp() so it applies the special rules on handling
13698 DN values that are of type PrintableString, as well as RDNs of type
13699 emailAddress where the value has the type ia5String.
13701 *stefank@valicert.com via Richard Levitte*
13703 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13704 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13705 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13706 the bitwise-OR of the two for use by the majority of applications
13707 wanting this behaviour, and update the docs. The documented
13708 behaviour and actual behaviour were inconsistent and had been
13709 changing anyway, so this is more a bug-fix than a behavioural
13712 *Geoff Thorpe, diagnosed by Nadav Har'El*
13714 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13715 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13719 * Fix initialization code race conditions in
13720 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13721 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13722 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13723 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13724 ssl2_get_cipher_by_char(),
13725 ssl3_get_cipher_by_char().
13727 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13729 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13730 the cached sessions are flushed, as the remove_cb() might use ex_data
13731 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13732 (see [openssl.org #212]).
13734 *Geoff Thorpe, Lutz Jaenicke*
13736 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13737 length, instead of the encoding length to d2i_ASN1_OBJECT.
13741 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13743 * [In 0.9.6g-engine release:]
13744 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13746 *Lynn Gazis <lgazis@rainbow.com>*
13748 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13750 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13751 and get fix the header length calculation.
13752 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13753 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13755 * Use proper error handling instead of 'assertions' in buffer
13756 overflow checks added in 0.9.6e. This prevents DoS (the
13757 assertions could call abort()).
13759 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13761 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13763 * Add various sanity checks to asn1_get_length() to reject
13764 the ASN1 length bytes if they exceed sizeof(long), will appear
13765 negative or the content length exceeds the length of the
13768 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13770 * Fix cipher selection routines: ciphers without encryption had no flags
13771 for the cipher strength set and where therefore not handled correctly
13772 by the selection routines (PR #130).
13776 * Fix EVP_dsa_sha macro.
13781 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13782 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13783 that was added in OpenSSL 0.9.6d.
13785 As the countermeasure turned out to be incompatible with some
13786 broken SSL implementations, the new option is part of SSL_OP_ALL.
13787 SSL_OP_ALL is usually employed when compatibility with weird SSL
13788 implementations is desired (e.g. '-bugs' option to 's_client' and
13789 's_server'), so the new option is automatically set in many
13794 * Changes in security patch:
13796 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13797 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13798 Air Force Materiel Command, USAF, under agreement number
13801 * Add various sanity checks to asn1_get_length() to reject
13802 the ASN1 length bytes if they exceed sizeof(long), will appear
13803 negative or the content length exceeds the length of the
13804 supplied buffer. ([CVE-2002-0659])
13806 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13808 * Assertions for various potential buffer overflows, not known to
13809 happen in practice.
13811 *Ben Laurie (CHATS)*
13813 * Various temporary buffers to hold ASCII versions of integers were
13814 too small for 64 bit platforms. ([CVE-2002-0655])
13815 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13817 * Remote buffer overflow in SSL3 protocol - an attacker could
13818 supply an oversized session ID to a client. ([CVE-2002-0656])
13820 *Ben Laurie (CHATS)*
13822 * Remote buffer overflow in SSL2 protocol - an attacker could
13823 supply an oversized client master key. ([CVE-2002-0656])
13825 *Ben Laurie (CHATS)*
13827 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13829 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13830 encoded as NULL) with id-dsa-with-sha1.
13832 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13834 * Check various `X509_...()` return values in `apps/req.c`.
13836 *Nils Larsch <nla@trustcenter.de>*
13838 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13839 an end-of-file condition would erroneously be flagged, when the CRLF
13840 was just at the end of a processed block. The bug was discovered when
13841 processing data through a buffering memory BIO handing the data to a
13842 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13843 <ptsekov@syntrex.com> and Nedelcho Stanev.
13847 * Implement a countermeasure against a vulnerability recently found
13848 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13849 before application data chunks to avoid the use of known IVs
13850 with data potentially chosen by the attacker.
13854 * Fix length checks in ssl3_get_client_hello().
13858 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13859 to prevent ssl3_read_internal() from incorrectly assuming that
13860 ssl3_read_bytes() found application data while handshake
13861 processing was enabled when in fact s->s3->in_read_app_data was
13862 merely automatically cleared during the initial handshake.
13864 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13866 * Fix object definitions for Private and Enterprise: they were not
13867 recognized in their shortname (=lowercase) representation. Extend
13868 obj_dat.pl to issue an error when using undefined keywords instead
13869 of silently ignoring the problem (Svenning Sorensen
13870 <sss@sss.dnsalias.net>).
13874 * Fix DH_generate_parameters() so that it works for 'non-standard'
13875 generators, i.e. generators other than 2 and 5. (Previously, the
13876 code did not properly initialise the 'add' and 'rem' values to
13877 BN_generate_prime().)
13879 In the new general case, we do not insist that 'generator' is
13880 actually a primitive root: This requirement is rather pointless;
13881 a generator of the order-q subgroup is just as good, if not
13886 * Map new X509 verification errors to alerts. Discovered and submitted by
13887 Tom Wu <tom@arcot.com>.
13891 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13892 returning non-zero before the data has been completely received
13893 when using non-blocking I/O.
13895 *Bodo Moeller; problem pointed out by John Hughes*
13897 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13899 *Ben Laurie, Lutz Jaenicke*
13901 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13902 Yoram Zahavi <YoramZ@gilian.com>).
13906 * Add information about CygWin 1.3 and on, and preserve proper
13907 configuration for the versions before that.
13909 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13911 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13912 check whether we deal with a copy of a session and do not delete from
13913 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13914 <izhar@checkpoint.com>.
13918 * Do not store session data into the internal session cache, if it
13919 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13920 flag is set). Proposed by Aslam <aslam@funk.com>.
13924 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13929 * [In 0.9.6d-engine release:]
13930 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13932 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13934 * Add the configuration target linux-s390x.
13936 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13938 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13939 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13940 variable as an indication that a ClientHello message has been
13941 received. As the flag value will be lost between multiple
13942 invocations of ssl3_accept when using non-blocking I/O, the
13943 function may not be aware that a handshake has actually taken
13944 place, thus preventing a new session from being added to the
13947 To avoid this problem, we now set s->new_session to 2 instead of
13948 using a local variable.
13950 *Lutz Jaenicke, Bodo Moeller*
13952 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13953 if the SSL_R_LENGTH_MISMATCH error is detected.
13955 *Geoff Thorpe, Bodo Moeller*
13957 * New 'shared_ldflag' column in Configure platform table.
13961 * Fix EVP_CIPHER_mode macro.
13963 *"Dan S. Camper" <dan@bti.net>*
13965 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13966 type, we must throw them away by setting rr->length to 0.
13968 *D P Chang <dpc@qualys.com>*
13970 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13972 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13973 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13974 worked incorrectly for those cases where range = `10..._2` and
13975 `3*range` is two bits longer than range.)
13979 * Only add signing time to PKCS7 structures if it is not already
13984 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13985 OBJ_ld_ce should be OBJ_id_ce.
13986 Also some ip-pda OIDs in crypto/objects/objects.txt were
13987 incorrect (cf. RFC 3039).
13989 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13991 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13992 returns early because it has nothing to do.
13994 *Andy Schneider <andy.schneider@bjss.co.uk>*
13996 * [In 0.9.6c-engine release:]
13997 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13999 *Andy Schneider <andy.schneider@bjss.co.uk>*
14001 * [In 0.9.6c-engine release:]
14002 Add support for Cryptographic Appliance's keyserver technology.
14003 (Use engine 'keyclient')
14005 *Cryptographic Appliances and Geoff Thorpe*
14007 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14008 is called via tools/c89.sh because arguments have to be
14009 rearranged (all '-L' options must appear before the first object
14012 *Richard Shapiro <rshapiro@abinitio.com>*
14014 * [In 0.9.6c-engine release:]
14015 Add support for Broadcom crypto accelerator cards, backported
14018 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14020 * [In 0.9.6c-engine release:]
14021 Add support for SureWare crypto accelerator cards from
14022 Baltimore Technologies. (Use engine 'sureware')
14024 *Baltimore Technologies and Mark Cox*
14026 * [In 0.9.6c-engine release:]
14027 Add support for crypto accelerator cards from Accelerated
14028 Encryption Processing, www.aep.ie. (Use engine 'aep')
14030 *AEP Inc. and Mark Cox*
14032 * Add a configuration entry for gcc on UnixWare.
14034 *Gary Benson <gbenson@redhat.com>*
14036 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14037 messages are stored in a single piece (fixed-length part and
14038 variable-length part combined) and fix various bugs found on the way.
14042 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14043 instead. BIO_gethostbyname() does not know what timeouts are
14044 appropriate, so entries would stay in cache even when they have
14046 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14048 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14049 faced with a pathologically small ClientHello fragment that does
14050 not contain client_version: Instead of aborting with an error,
14051 simply choose the highest available protocol version (i.e.,
14052 TLS 1.0 unless it is disabled). In practice, ClientHello
14053 messages are never sent like this, but this change gives us
14054 strictly correct behaviour at least for TLS.
14058 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14059 never resets s->method to s->ctx->method when called from within
14060 one of the SSL handshake functions.
14062 *Bodo Moeller; problem pointed out by Niko Baric*
14064 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14065 (sent using the client's version number) if client_version is
14066 smaller than the protocol version in use. Also change
14067 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14068 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14069 the client will at least see that alert.
14073 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14078 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14079 client receives HelloRequest while in a handshake.
14081 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14083 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14084 should end in 'break', not 'goto end' which circumvents various
14085 cleanups done in state SSL_ST_OK. But session related stuff
14086 must be disabled for SSL_ST_OK in the case that we just sent a
14089 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14090 before just sending a HelloRequest.
14092 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14094 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14095 reveal whether illegal block cipher padding was found or a MAC
14096 verification error occurred. (Neither SSLerr() codes nor alerts
14097 are directly visible to potential attackers, but the information
14098 may leak via logfiles.)
14100 Similar changes are not required for the SSL 2.0 implementation
14101 because the number of padding bytes is sent in clear for SSL 2.0,
14102 and the extra bytes are just ignored. However ssl/s2_pkt.c
14103 failed to verify that the purported number of padding bytes is in
14108 * Add OpenUNIX-8 support including shared libraries
14109 (Boyd Lynn Gerber <gerberb@zenez.com>).
14113 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14114 'wristwatch attack' using huge encoding parameters (cf.
14115 James H. Manger's CRYPTO 2001 paper). Note that the
14116 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14117 encoding parameters and hence was not vulnerable.
14121 * BN_sqr() bug fix.
14123 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14125 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14126 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14127 followed by modular reduction.
14129 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14131 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14132 equivalent based on BN_pseudo_rand() instead of BN_rand().
14136 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14137 This function was broken, as the check for a new client hello message
14138 to handle SGC did not allow these large messages.
14139 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14143 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14147 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14148 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14152 * Rework the configuration and shared library support for Tru64 Unix.
14153 The configuration part makes use of modern compiler features and
14154 still retains old compiler behavior for those that run older versions
14155 of the OS. The shared library support part includes a variant that
14156 uses the RPATH feature, and is available through the special
14157 configuration target "alpha-cc-rpath", which will never be selected
14160 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14162 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14163 with the same message size as in ssl3_get_certificate_request().
14164 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14165 messages might inadvertently be reject as too long.
14167 *Petr Lampa <lampa@fee.vutbr.cz>*
14169 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14173 * Modified SSL library such that the verify_callback that has been set
14174 specifically for an SSL object with SSL_set_verify() is actually being
14175 used. Before the change, a verify_callback set with this function was
14176 ignored and the verify_callback() set in the SSL_CTX at the time of
14177 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14178 to allow the necessary settings.
14182 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14183 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14184 done automatically (in contradiction to the requirements of the C
14185 standard). This made problems when used from OpenSSH.
14189 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14190 dh->length and always used
14192 BN_rand_range(priv_key, dh->p).
14194 BN_rand_range() is not necessary for Diffie-Hellman, and this
14195 specific range makes Diffie-Hellman unnecessarily inefficient if
14196 dh->length (recommended exponent length) is much smaller than the
14197 length of dh->p. We could use BN_rand_range() if the order of
14198 the subgroup was stored in the DH structure, but we only have
14203 BN_rand(priv_key, l, ...)
14205 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14212 RSA_eay_public_encrypt
14213 RSA_eay_private_decrypt
14214 RSA_eay_private_encrypt (signing)
14215 RSA_eay_public_decrypt (signature verification)
14217 (default implementations for RSA_public_encrypt,
14218 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14219 always reject numbers >= n.
14223 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14224 to synchronize access to 'locking_thread'. This is necessary on
14225 systems where access to 'locking_thread' (an 'unsigned long'
14226 variable) is not atomic.
14230 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14231 *before* setting the 'crypto_lock_rand' flag. The previous code had
14232 a race condition if 0 is a valid thread ID.
14234 *Travis Vitek <vitek@roguewave.com>*
14236 * Add support for shared libraries under Irix.
14238 *Albert Chin-A-Young <china@thewrittenword.com>*
14240 * Add configuration option to build on Linux on both big-endian and
14241 little-endian MIPS.
14243 *Ralf Baechle <ralf@uni-koblenz.de>*
14245 * Add the possibility to create shared libraries on HP-UX.
14249 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14251 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14252 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14253 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14254 PRNG state recovery was possible based on the output of
14255 one PRNG request appropriately sized to gain knowledge on
14256 'md' followed by enough consecutive 1-byte PRNG requests
14257 to traverse all of 'state'.
14259 1. When updating 'md_local' (the current thread's copy of 'md')
14260 during PRNG output generation, hash all of the previous
14261 'md_local' value, not just the half used for PRNG output.
14263 2. Make the number of bytes from 'state' included into the hash
14264 independent from the number of PRNG bytes requested.
14266 The first measure alone would be sufficient to avoid
14267 Markku-Juhani's attack. (Actually it had never occurred
14268 to me that the half of 'md_local' used for chaining was the
14269 half from which PRNG output bytes were taken -- I had always
14270 assumed that the secret half would be used.) The second
14271 measure makes sure that additional data from 'state' is never
14272 mixed into 'md_local' in small portions; this heuristically
14273 further strengthens the PRNG.
14277 * Fix crypto/bn/asm/mips3.s.
14281 * When only the key is given to "enc", the IV is undefined. Print out
14282 an error message in this case.
14286 * Handle special case when X509_NAME is empty in X509 printing routines.
14290 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14291 positive and less than q.
14295 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14296 used: it isn't thread safe and the add_lock_callback should handle
14299 *Paul Rose <Paul.Rose@bridge.com>*
14301 * Verify that incoming data obeys the block size in
14302 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14308 *Ulf Möller, Bodo Möller*
14310 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14311 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14312 when fixing the server behaviour for backwards-compatible 'client
14313 hello' messages. (Note that the attack is impractical against
14314 SSL 3.0 and TLS 1.0 anyway because length and version checking
14315 means that the probability of guessing a valid ciphertext is
14316 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14319 Before 0.9.5, the countermeasure (hide the error by generating a
14320 random 'decryption result') did not work properly because
14321 ERR_clear_error() was missing, meaning that SSL_get_error() would
14322 detect the supposedly ignored error.
14324 Both problems are now fixed.
14328 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14329 (previously it was 1024).
14333 * Fix for compatibility mode trust settings: ignore trust settings
14334 unless some valid trust or reject settings are present.
14338 * Fix for blowfish EVP: its a variable length cipher.
14342 * Fix various bugs related to DSA S/MIME verification. Handle missing
14343 parameters in DSA public key structures and return an error in the
14344 DSA routines if parameters are absent.
14348 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14349 in the current directory if neither $RANDFILE nor $HOME was set.
14350 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14351 caused some confusion to Windows users who haven't defined $HOME.
14352 Thus RAND_file_name() is changed again: e_os.h can define a
14353 DEFAULT_HOME, which will be used if $HOME is not set.
14354 For Windows, we use "C:"; on other platforms, we still require
14355 environment variables.
14357 * Move 'if (!initialized) RAND_poll()' into regions protected by
14358 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14359 having multiple threads call RAND_poll() concurrently.
14363 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14364 combination of a flag and a thread ID variable.
14365 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14366 flag), *other* threads can enter ssleay_add_bytes without obeying
14367 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14368 that they do not hold after the first thread unsets add_do_not_lock).
14372 * Change bctest again: '-x' expressions are not available in all
14373 versions of 'test'.
14377 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14379 * Fix a couple of memory leaks in PKCS7_dataDecode()
14381 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14383 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14384 the default extension for executables, if any. Also, make the perl
14385 scripts that use symlink() to test if it really exists and use "cp"
14386 if it doesn't. All this made OpenSSL compilable and installable in
14391 * Fix for asn1_GetSequence() for indefinite length constructed data.
14392 If SEQUENCE is length is indefinite just set c->slen to the total
14393 amount of data available.
14395 *Steve Henson, reported by shige@FreeBSD.org*
14397 *This change does not apply to 0.9.7.*
14399 * Change bctest to avoid here-documents inside command substitution
14400 (workaround for FreeBSD /bin/sh bug).
14401 For compatibility with Ultrix, avoid shell functions (introduced
14402 in the bctest version that searches along $PATH).
14406 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14407 with des_encrypt() defined on some operating systems, like Solaris
14412 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14413 On the Importance of Eliminating Errors in Cryptographic
14414 Computations, J. Cryptology 14 (2001) 2, 101-119,
14415 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14419 * MIPS assembler BIGNUM division bug fix.
14423 * Disabled incorrect Alpha assembler code.
14427 * Fix PKCS#7 decode routines so they correctly update the length
14428 after reading an EOC for the EXPLICIT tag.
14432 *This change does not apply to 0.9.7.*
14434 * Fix bug in PKCS#12 key generation routines. This was triggered
14435 if a 3DES key was generated with a 0 initial byte. Include
14436 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14437 (but broken) behaviour.
14441 * Enhance bctest to search for a working bc along $PATH and print
14444 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14446 * Fix memory leaks in err.c: free err_data string if necessary;
14447 don't write to the wrong index in ERR_set_error_data.
14451 * Implement ssl23_peek (analogous to ssl23_read), which previously
14456 * Replace rdtsc with `_emit` statements for VC++ version 5.
14458 *Jeremy Cooper <jeremy@baymoo.org>*
14460 * Make it possible to reuse SSLv2 sessions.
14464 * In copy_email() check for >= 0 as a return value for
14465 X509_NAME_get_index_by_NID() since 0 is a valid index.
14467 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14469 * Avoid coredump with unsupported or invalid public keys by checking if
14470 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14471 PKCS7_verify() fails with non detached data.
14475 * Don't use getenv in library functions when run as setuid/setgid.
14476 New function OPENSSL_issetugid().
14480 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14481 due to incorrect handling of multi-threading:
14483 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14485 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14487 3. Count how many times MemCheck_off() has been called so that
14488 nested use can be treated correctly. This also avoids
14489 inband-signalling in the previous code (which relied on the
14490 assumption that thread ID 0 is impossible).
14494 * Add "-rand" option also to s_client and s_server.
14498 * Fix CPU detection on Irix 6.x.
14499 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14500 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14502 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14507 *This change does not apply to 0.9.7.*
14509 * Use the cached encoding of an X509_NAME structure rather than
14510 copying it. This is apparently the reason for the libsafe "errors"
14511 but the code is actually correct.
14515 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14516 Bleichenbacher's DSA attack.
14517 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14518 to be set and top=0 forces the highest bit to be set; top=-1 is new
14519 and leaves the highest bit random.
14521 *Ulf Moeller, Bodo Moeller*
14523 * In the `NCONF_...`-based implementations for `CONF_...` queries
14524 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14525 a temporary CONF structure with the data component set to NULL
14526 (which gives segmentation faults in lh_retrieve).
14527 Instead, use NULL for the CONF pointer in CONF_get_string and
14528 CONF_get_number (which may use environment variables) and directly
14529 return NULL from CONF_get_section.
14533 * Fix potential buffer overrun for EBCDIC.
14537 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14538 keyUsage if basicConstraints absent for a CA.
14542 * Make SMIME_write_PKCS7() write mail header values with a format that
14543 is more generally accepted (no spaces before the semicolon), since
14544 some programs can't parse those values properly otherwise. Also make
14545 sure BIO's that break lines after each write do not create invalid
14550 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14551 macros previously used would not encode an empty SEQUENCE OF
14552 and break the signature.
14556 *This change does not apply to 0.9.7.*
14558 * Zero the premaster secret after deriving the master secret in
14563 * Add some EVP_add_digest_alias registrations (as found in
14564 OpenSSL_add_all_digests()) to SSL_library_init()
14565 aka OpenSSL_add_ssl_algorithms(). This provides improved
14566 compatibility with peers using X.509 certificates
14567 with unconventional AlgorithmIdentifier OIDs.
14571 * Fix for Irix with NO_ASM.
14573 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14575 * ./config script fixes.
14577 *Ulf Moeller, Richard Levitte*
14579 * Fix 'openssl passwd -1'.
14583 * Change PKCS12_key_gen_asc() so it can cope with non null
14584 terminated strings whose length is passed in the passlen
14585 parameter, for example from PEM callbacks. This was done
14586 by adding an extra length parameter to asc2uni().
14588 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14590 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14591 call failed, free the DSA structure.
14595 * Fix to uni2asc() to cope with zero length Unicode strings.
14596 These are present in some PKCS#12 files.
14600 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14601 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14602 when writing a 32767 byte record.
14604 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14606 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14607 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14609 (RSA objects have a reference count access to which is protected
14610 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14611 so they are meant to be shared between threads.)
14612 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14613 "Reddie, Steven" <Steven.Reddie@ca.com>*
14615 * Fix a deadlock in CRYPTO_mem_leaks().
14619 * Use better test patterns in bntest.
14623 * rand_win.c fix for Borland C.
14627 * BN_rshift bugfix for n == 0.
14631 * Add a 'bctest' script that checks for some known 'bc' bugs
14632 so that 'make test' does not abort just because 'bc' is broken.
14636 * Store verify_result within SSL_SESSION also for client side to
14637 avoid potential security hole. (Re-used sessions on the client side
14638 always resulted in verify_result==X509_V_OK, not using the original
14639 result of the server certificate verification.)
14643 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14644 SSL3_RT_APPLICATION_DATA, return 0.
14645 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14650 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14651 releases, have been re-implemented by renaming the previous
14652 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14653 and ssl3_read_internal, respectively, and adding 'peek' parameters
14654 to them. The new ssl[23]_{read,peek} functions are calls to
14655 ssl[23]_read_internal with the 'peek' flag set appropriately.
14656 A 'peek' parameter has also been added to ssl3_read_bytes, which
14657 does the actual work for ssl3_read_internal.
14661 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14662 the method-specific "init()" handler. Also clean up ex_data after
14663 calling the method-specific "finish()" handler. Previously, this was
14664 happening the other way round.
14668 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14669 The previous value, 12, was not always sufficient for BN_mod_exp().
14673 * Make sure that shared libraries get the internal name engine with
14674 the full version number and not just 0. This should mark the
14675 shared libraries as not backward compatible. Of course, this should
14676 be changed again when we can guarantee backward binary compatibility.
14680 * Fix typo in get_cert_by_subject() in by_dir.c
14682 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14684 * Rework the system to generate shared libraries:
14686 - Make note of the expected extension for the shared libraries and
14687 if there is a need for symbolic links from for example libcrypto.so.0
14688 to libcrypto.so.0.9.7. There is extended info in Configure for
14691 - Make as few rebuilds of the shared libraries as possible.
14693 - Still avoid linking the OpenSSL programs with the shared libraries.
14695 - When installing, install the shared libraries separately from the
14700 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14702 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14703 and not in SSL_clear because the latter is also used by the
14704 accept/connect functions; previously, the settings made by
14705 SSL_set_read_ahead would be lost during the handshake.
14707 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14709 * Correct util/mkdef.pl to be selective about disabled algorithms.
14710 Previously, it would create entries for disabled algorithms no
14715 * Added several new manual pages for SSL_* function.
14719 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14721 * In ssl23_get_client_hello, generate an error message when faced
14722 with an initial SSL 3.0/TLS record that is too small to contain the
14723 first two bytes of the ClientHello message, i.e. client_version.
14724 (Note that this is a pathologic case that probably has never happened
14725 in real life.) The previous approach was to use the version number
14726 from the record header as a substitute; but our protocol choice
14727 should not depend on that one because it is not authenticated
14728 by the Finished messages.
14732 * More robust randomness gathering functions for Windows.
14734 *Jeffrey Altman <jaltman@columbia.edu>*
14736 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14737 not set then we don't setup the error code for issuer check errors
14738 to avoid possibly overwriting other errors which the callback does
14739 handle. If an application does set the flag then we assume it knows
14740 what it is doing and can handle the new informational codes
14745 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14746 a general "ANY" type, as such it should be able to decode anything
14747 including tagged types. However it didn't check the class so it would
14748 wrongly interpret tagged types in the same way as their universal
14749 counterpart and unknown types were just rejected. Changed so that the
14750 tagged and unknown types are handled in the same way as a SEQUENCE:
14751 that is the encoding is stored intact. There is also a new type
14752 "V_ASN1_OTHER" which is used when the class is not universal, in this
14753 case we have no idea what the actual type is so we just lump them all
14758 * On VMS, stdout may very well lead to a file that is written to
14759 in a record-oriented fashion. That means that every write() will
14760 write a separate record, which will be read separately by the
14761 programs trying to read from it. This can be very confusing.
14763 The solution is to put a BIO filter in the way that will buffer
14764 text until a linefeed is reached, and then write everything a
14765 line at a time, so every record written will be an actual line,
14766 not chunks of lines and not (usually doesn't happen, but I've
14767 seen it once) several lines in one record. BIO_f_linebuffer() is
14770 Currently, it's a VMS-only method, because that's where it has
14771 been tested well enough.
14775 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14776 it can return incorrect results.
14777 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14778 but it was in 0.9.6-beta[12].)
14782 * Disable the check for content being present when verifying detached
14783 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14784 include zero length content when signing messages.
14788 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14789 BIO_ctrl (for BIO pairs).
14793 * Add DSO method for VMS.
14797 * Bug fix: Montgomery multiplication could produce results with the
14802 * Add RPM specification openssl.spec and modify it to build three
14803 packages. The default package contains applications, application
14804 documentation and run-time libraries. The devel package contains
14805 include files, static libraries and function documentation. The
14806 doc package contains the contents of the doc directory. The original
14807 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14811 * Add a large number of documentation files for many SSL routines.
14813 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14815 * Add a configuration entry for Sony News 4.
14817 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14819 * Don't set the two most significant bits to one when generating a
14820 random number < q in the DSA library.
14824 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14825 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14826 the underlying transport is blocking) if a handshake took place.
14827 (The default behaviour is needed by applications such as s_client
14828 and s_server that use select() to determine when to use SSL_read;
14829 but for applications that know in advance when to expect data, it
14830 just makes things more complicated.)
14834 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14839 * Add a few more EBCDIC conditionals that make `req` and `x509`
14840 work better on such systems.
14842 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14844 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14845 Update PKCS12_parse() so it copies the friendlyName and the
14846 keyid to the certificates aux info.
14850 * Fix bug in PKCS7_verify() which caused an infinite loop
14851 if there was more than one signature.
14853 *Sven Uszpelkat <su@celocom.de>*
14855 * Major change in util/mkdef.pl to include extra information
14856 about each symbol, as well as presenting variables as well
14857 as functions. This change means that there's n more need
14858 to rebuild the .num files when some algorithms are excluded.
14862 * Allow the verify time to be set by an application,
14863 rather than always using the current time.
14867 * Phase 2 verify code reorganisation. The certificate
14868 verify code now looks up an issuer certificate by a
14869 number of criteria: subject name, authority key id
14870 and key usage. It also verifies self signed certificates
14871 by the same criteria. The main comparison function is
14872 X509_check_issued() which performs these checks.
14874 Lot of changes were necessary in order to support this
14875 without completely rewriting the lookup code.
14877 Authority and subject key identifier are now cached.
14879 The LHASH 'certs' is X509_STORE has now been replaced
14880 by a STACK_OF(X509_OBJECT). This is mainly because an
14881 LHASH can't store or retrieve multiple objects with
14882 the same hash value.
14884 As a result various functions (which were all internal
14885 use only) have changed to handle the new X509_STORE
14886 structure. This will break anything that messed round
14887 with X509_STORE internally.
14889 The functions X509_STORE_add_cert() now checks for an
14890 exact match, rather than just subject name.
14892 The X509_STORE API doesn't directly support the retrieval
14893 of multiple certificates matching a given criteria, however
14894 this can be worked round by performing a lookup first
14895 (which will fill the cache with candidate certificates)
14896 and then examining the cache for matches. This is probably
14897 the best we can do without throwing out X509_LOOKUP
14898 entirely (maybe later...).
14900 The X509_VERIFY_CTX structure has been enhanced considerably.
14902 All certificate lookup operations now go via a get_issuer()
14903 callback. Although this currently uses an X509_STORE it
14904 can be replaced by custom lookups. This is a simple way
14905 to bypass the X509_STORE hackery necessary to make this
14906 work and makes it possible to use more efficient techniques
14907 in future. A very simple version which uses a simple
14908 STACK for its trusted certificate store is also provided
14909 using X509_STORE_CTX_trusted_stack().
14911 The verify_cb() and verify() callbacks now have equivalents
14912 in the X509_STORE_CTX structure.
14914 X509_STORE_CTX also has a 'flags' field which can be used
14915 to customise the verify behaviour.
14919 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14920 excludes S/MIME capabilities.
14924 * When a certificate request is read in keep a copy of the
14925 original encoding of the signed data and use it when outputting
14926 again. Signatures then use the original encoding rather than
14927 a decoded, encoded version which may cause problems if the
14928 request is improperly encoded.
14932 * For consistency with other BIO_puts implementations, call
14933 buffer_write(b, ...) directly in buffer_puts instead of calling
14936 In BIO_puts, increment b->num_write as in BIO_write.
14938 *Peter.Sylvester@EdelWeb.fr*
14940 * Fix BN_mul_word for the case where the word is 0. (We have to use
14941 BN_zero, we may not return a BIGNUM with an array consisting of
14942 words set to zero.)
14946 * Avoid calling abort() from within the library when problems are
14947 detected, except if preprocessor symbols have been defined
14948 (such as REF_CHECK, BN_DEBUG etc.).
14952 * New openssl application 'rsautl'. This utility can be
14953 used for low level RSA operations. DER public key
14954 BIO/fp routines also added.
14958 * New Configure entry and patches for compiling on QNX 4.
14960 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14962 * A demo state-machine implementation was sponsored by
14963 Nuron (<http://www.nuron.com/>) and is now available in
14964 demos/state_machine.
14968 * New options added to the 'dgst' utility for signature
14969 generation and verification.
14973 * Unrecognized PKCS#7 content types are now handled via a
14974 catch all ASN1_TYPE structure. This allows unsupported
14975 types to be stored as a "blob" and an application can
14976 encode and decode it manually.
14980 * Fix various signed/unsigned issues to make a_strex.c
14981 compile under VC++.
14983 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14985 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14986 length if passed a buffer. ASN1_INTEGER_to_BN failed
14987 if passed a NULL BN and its argument was negative.
14989 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14991 * Modification to PKCS#7 encoding routines to output definite
14992 length encoding. Since currently the whole structures are in
14993 memory there's not real point in using indefinite length
14994 constructed encoding. However if OpenSSL is compiled with
14995 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14999 * Added BIO_vprintf() and BIO_vsnprintf().
15003 * Added more prefixes to parse for in the strings written
15004 through a logging bio, to cover all the levels that are available
15005 through syslog. The prefixes are now:
15007 PANIC, EMERG, EMR => LOG_EMERG
15008 ALERT, ALR => LOG_ALERT
15009 CRIT, CRI => LOG_CRIT
15010 ERROR, ERR => LOG_ERR
15011 WARNING, WARN, WAR => LOG_WARNING
15012 NOTICE, NOTE, NOT => LOG_NOTICE
15013 INFO, INF => LOG_INFO
15014 DEBUG, DBG => LOG_DEBUG
15016 and as before, if none of those prefixes are present at the
15017 beginning of the string, LOG_ERR is chosen.
15019 On Win32, the `LOG_*` levels are mapped according to this:
15021 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15022 LOG_WARNING => EVENTLOG_WARNING_TYPE
15023 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15027 * Made it possible to reconfigure with just the configuration
15028 argument "reconf" or "reconfigure". The command line arguments
15029 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15030 and are retrieved from there when reconfiguring.
15036 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15038 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15042 * The obj_dat.pl script was messing up the sorting of object
15043 names. The reason was that it compared the quoted version
15044 of strings as a result "OCSP" > "OCSP Signing" because
15045 " > SPACE. Changed script to store unquoted versions of
15046 names and add quotes on output. It was also omitting some
15047 names from the lookup table if they were given a default
15048 value (that is if SN is missing it is given the same
15049 value as LN and vice versa), these are now added on the
15050 grounds that if an object has a name we should be able to
15051 look it up. Finally added warning output when duplicate
15052 short or long names are found.
15056 * Changes needed for Tandem NSK.
15058 *Scott Uroff <scott@xypro.com>*
15060 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15061 RSA_padding_check_SSLv23(), special padding was never detected
15062 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15063 version rollback attacks was not effective.
15065 In s23_clnt.c, don't use special rollback-attack detection padding
15066 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15067 client; similarly, in s23_srvr.c, don't do the rollback check if
15068 SSL 2.0 is the only protocol enabled in the server.
15072 * Make it possible to get hexdumps of unprintable data with 'openssl
15073 asn1parse'. By implication, the functions ASN1_parse_dump() and
15074 BIO_dump_indent() are added.
15078 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15079 these print out strings and name structures based on various
15080 flags including RFC2253 support and proper handling of
15081 multibyte characters. Added options to the 'x509' utility
15082 to allow the various flags to be set.
15086 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15087 Also change the functions X509_cmp_current_time() and
15088 X509_gmtime_adj() work with an ASN1_TIME structure,
15089 this will enable certificates using GeneralizedTime in validity
15090 dates to be checked.
15094 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15095 negative public key encodings) on by default,
15096 NO_NEG_PUBKEY_BUG can be set to disable it.
15100 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15101 content octets. An i2c_ASN1_OBJECT is unnecessary because
15102 the encoding can be trivially obtained from the structure.
15106 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15107 not read locks (`CRYPTO_r_[un]lock`).
15111 * A first attempt at creating official support for shared
15112 libraries through configuration. I've kept it so the
15113 default is static libraries only, and the OpenSSL programs
15114 are always statically linked for now, but there are
15115 preparations for dynamic linking in place.
15116 This has been tested on Linux and Tru64.
15120 * Randomness polling function for Win9x, as described in:
15121 Peter Gutmann, Software Generation of Practically Strong
15126 * Fix so PRNG is seeded in req if using an already existing
15131 * New options to smime application. -inform and -outform
15132 allow alternative formats for the S/MIME message including
15133 PEM and DER. The -content option allows the content to be
15134 specified separately. This should allow things like Netscape
15135 form signing output easier to verify.
15139 * Fix the ASN1 encoding of tags using the 'long form'.
15143 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15144 STRING types. These convert content octets to and from the
15145 underlying type. The actual tag and length octets are
15146 already assumed to have been read in and checked. These
15147 are needed because all other string types have virtually
15148 identical handling apart from the tag. By having versions
15149 of the ASN1 functions that just operate on content octets
15150 IMPLICIT tagging can be handled properly. It also allows
15151 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15152 and ASN1_INTEGER are identical apart from the tag.
15156 * Change the handling of OID objects as follows:
15158 - New object identifiers are inserted in objects.txt, following
15159 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15160 - objects.pl is used to process obj_mac.num and create a new
15162 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15165 This is currently kind of a hack, and the perl code in objects.pl
15166 isn't very elegant, but it works as I intended. The simplest way
15167 to check that it worked correctly is to look in obj_dat.h and
15168 check the array nid_objs and make sure the objects haven't moved
15169 around (this is important!). Additions are OK, as well as
15170 consistent name changes.
15174 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15178 * Addition of the command line parameter '-rand file' to 'openssl req'.
15179 The given file adds to whatever has already been seeded into the
15180 random pool through the RANDFILE configuration file option or
15181 environment variable, or the default random state file.
15185 * mkstack.pl now sorts each macro group into lexical order.
15186 Previously the output order depended on the order the files
15187 appeared in the directory, resulting in needless rewriting
15192 * Patches to make OpenSSL compile under Win32 again. Mostly
15193 work arounds for the VC++ problem that it treats func() as
15194 func(void). Also stripped out the parts of mkdef.pl that
15195 added extra typesafe functions: these no longer exist.
15199 * Reorganisation of the stack code. The macros are now all
15200 collected in safestack.h . Each macro is defined in terms of
15201 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15202 DEBUG_SAFESTACK is now handled in terms of function casts,
15203 this has the advantage of retaining type safety without the
15204 use of additional functions. If DEBUG_SAFESTACK is not defined
15205 then the non typesafe macros are used instead. Also modified the
15206 mkstack.pl script to handle the new form. Needs testing to see
15207 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15208 the default if no major problems. Similar behaviour for ASN1_SET_OF
15209 and PKCS12_STACK_OF.
15213 * When some versions of IIS use the 'NET' form of private key the
15214 key derivation algorithm is different. Normally MD5(password) is
15215 used as a 128 bit RC4 key. In the modified case
15216 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15217 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15218 as the old Netscape_RSA functions except they have an additional
15219 'sgckey' parameter which uses the modified algorithm. Also added
15220 an -sgckey command line option to the rsa utility. Thanks to
15221 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15222 algorithm to openssl-dev.
15226 * The evp_local.h macros were using 'c.##kname' which resulted in
15227 invalid expansion on some systems (SCO 5.0.5 for example).
15228 Corrected to 'c.kname'.
15230 *Phillip Porch <root@theporch.com>*
15232 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15233 a STACK of email addresses from a certificate or request, these look
15234 in the subject name and the subject alternative name extensions and
15235 omit any duplicate addresses.
15239 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15240 This makes DSA verification about 2 % faster.
15244 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15245 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15246 plus overhead for 1024 bit moduli).
15247 This makes exponentiations about 0.5 % faster for 1024 bit
15248 exponents (as measured by "openssl speed rsa2048").
15252 * Rename memory handling macros to avoid conflicts with other
15254 Malloc => OPENSSL_malloc
15255 Malloc_locked => OPENSSL_malloc_locked
15256 Realloc => OPENSSL_realloc
15257 Free => OPENSSL_free
15261 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15262 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15266 * CygWin32 support.
15268 *John Jarvie <jjarvie@newsguy.com>*
15270 * The type-safe stack code has been rejigged. It is now only compiled
15271 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15272 by default all type-specific stack functions are "#define"d back to
15273 standard stack functions. This results in more streamlined output
15274 but retains the type-safety checking possibilities of the original
15279 * The STACK code has been cleaned up, and certain type declarations
15280 that didn't make a lot of sense have been brought in line. This has
15281 also involved a cleanup of sorts in safestack.h to more correctly
15282 map type-safe stack functions onto their plain stack counterparts.
15283 This work has also resulted in a variety of "const"ifications of
15284 lots of the code, especially `_cmp` operations which should normally
15285 be prototyped with "const" parameters anyway.
15289 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15290 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15291 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15292 where all of 'md' is used each time the PRNG is used, but 'state'
15293 is used only indexed by a cyclic counter. As entropy may not be
15294 well distributed from the beginning, 'md' is important as a
15295 chaining variable. However, the output function chains only half
15296 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15297 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15298 in all of 'state' being rewritten, with the new values depending
15299 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15303 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15304 the handshake is continued after ssl_verify_cert_chain();
15305 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15306 can lead to 'unexplainable' connection aborts later.
15308 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15310 * Major EVP API cipher revision.
15311 Add hooks for extra EVP features. This allows various cipher
15312 parameters to be set in the EVP interface. Support added for variable
15313 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15314 setting of RC2 and RC5 parameters.
15316 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15319 Remove lots of duplicated code from the EVP library. For example *every*
15320 cipher init() function handles the 'iv' in the same way according to the
15321 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15322 for CFB and OFB modes they zero ctx->num.
15324 New functionality allows removal of S/MIME code RC2 hack.
15326 Most of the routines have the same form and so can be declared in terms
15329 By shifting this to the top level EVP_CipherInit() it can be removed from
15330 all individual ciphers. If the cipher wants to handle IVs or keys
15331 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15334 Change lots of functions like EVP_EncryptUpdate() to now return a
15335 value: although software versions of the algorithms cannot fail
15336 any installed hardware versions can.
15340 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15341 this option is set, tolerate broken clients that send the negotiated
15342 protocol version number instead of the requested protocol version
15347 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15348 i.e. non-zero for export ciphersuites, zero otherwise.
15349 Previous versions had this flag inverted, inconsistent with
15350 rsa_tmp_cb (..._TMP_RSA_CB).
15352 *Bodo Moeller; problem reported by Amit Chopra*
15354 * Add missing DSA library text string. Work around for some IIS
15355 key files with invalid SEQUENCE encoding.
15359 * Add a document (doc/standards.txt) that list all kinds of standards
15360 and so on that are implemented in OpenSSL.
15364 * Enhance c_rehash script. Old version would mishandle certificates
15365 with the same subject name hash and wouldn't handle CRLs at all.
15366 Added -fingerprint option to crl utility, to support new c_rehash
15371 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15375 * Fix for SSL server purpose checking. Server checking was
15376 rejecting certificates which had extended key usage present
15377 but no ssl client purpose.
15379 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15381 * Make PKCS#12 code work with no password. The PKCS#12 spec
15382 is a little unclear about how a blank password is handled.
15383 Since the password in encoded as a BMPString with terminating
15384 double NULL a zero length password would end up as just the
15385 double NULL. However no password at all is different and is
15386 handled differently in the PKCS#12 key generation code. NS
15387 treats a blank password as zero length. MSIE treats it as no
15388 password on export: but it will try both on import. We now do
15389 the same: PKCS12_parse() tries zero length and no password if
15390 the password is set to "" or NULL (NULL is now a valid password:
15391 it wasn't before) as does the pkcs12 application.
15395 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15396 perror when PEM_read_bio_X509_REQ fails, the error message must
15397 be obtained from the error queue.
15401 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15402 it in ERR_remove_state if appropriate, and change ERR_get_state
15403 accordingly to avoid race conditions (this is necessary because
15404 thread_hash is no longer constant once set).
15408 * Bugfix for linux-elf makefile.one.
15412 * RSA_get_default_method() will now cause a default
15413 RSA_METHOD to be chosen if one doesn't exist already.
15414 Previously this was only set during a call to RSA_new()
15415 or RSA_new_method(NULL) meaning it was possible for
15416 RSA_get_default_method() to return NULL.
15420 * Added native name translation to the existing DSO code
15421 that will convert (if the flag to do so is set) filenames
15422 that are sufficiently small and have no path information
15423 into a canonical native form. Eg. "blah" converted to
15424 "libblah.so" or "blah.dll" etc.
15428 * New function ERR_error_string_n(e, buf, len) which is like
15429 ERR_error_string(e, buf), but writes at most 'len' bytes
15430 including the 0 terminator. For ERR_error_string_n, 'buf'
15433 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15435 * CONF library reworked to become more general. A new CONF
15436 configuration file reader "class" is implemented as well as a
15437 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15438 old `CONF_*` functions are still there, but are reimplemented to
15439 work in terms of the new functions. Also, a set of functions
15440 to handle the internal storage of the configuration data is
15441 provided to make it easier to write new configuration file
15442 reader "classes" (I can definitely see something reading a
15443 configuration file in XML format, for example), called `_CONF_*`,
15444 or "the configuration storage API"...
15446 The new configuration file reading functions are:
15448 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15449 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15451 NCONF_default, NCONF_WIN32
15453 NCONF_dump_fp, NCONF_dump_bio
15455 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15456 NCONF_new creates a new CONF object. This works in the same way
15457 as other interfaces in OpenSSL, like the BIO interface.
15458 `NCONF_dump_*` dump the internal storage of the configuration file,
15459 which is useful for debugging. All other functions take the same
15460 arguments as the old `CONF_*` functions with the exception of the
15461 first that must be a `CONF *` instead of a `LHASH *`.
15463 To make it easier to use the new classes with the old `CONF_*` functions,
15464 the function CONF_set_default_method is provided.
15468 * Add '-tls1' option to 'openssl ciphers', which was already
15469 mentioned in the documentation but had not been implemented.
15470 (This option is not yet really useful because even the additional
15471 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15475 * Initial DSO code added into libcrypto for letting OpenSSL (and
15476 OpenSSL-based applications) load shared libraries and bind to
15477 them in a portable way.
15479 *Geoff Thorpe, with contributions from Richard Levitte*
15481 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15483 * Make sure _lrotl and _lrotr are only used with MSVC.
15485 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15486 (the default implementation of RAND_status).
15488 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15489 to '-clrext' (= clear extensions), as intended and documented.
15490 *Bodo Moeller; inconsistency pointed out by Michael Attili
15491 <attili@amaxo.com>*
15493 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15494 was larger than the MD block size.
15496 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15498 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15499 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15500 using the passed key: if the passed key was a private key the result
15501 of X509_print(), for example, would be to print out all the private key
15506 * des_quad_cksum() byte order bug fix.
15507 *Ulf Möller, using the problem description in krb4-0.9.7, where
15508 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15510 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15513 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15515 * For easily testing in shell scripts whether some command
15516 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15517 returns with exit code 0 iff no command of the given name is available.
15518 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15519 the output goes to stdout and nothing is printed to stderr.
15520 Additional arguments are always ignored.
15522 Since for each cipher there is a command of the same name,
15523 the 'no-cipher' compilation switches can be tested this way.
15525 ('openssl no-XXX' is not able to detect pseudo-commands such
15526 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15530 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15534 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15535 is set; it will be thrown away anyway because each handshake creates
15537 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15538 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15539 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15540 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15544 * New s_client option -ign_eof: EOF at stdin is ignored, and
15545 'Q' and 'R' lose their special meanings (quit/renegotiate).
15546 This is part of what -quiet does; unlike -quiet, -ign_eof
15547 does not suppress any output.
15551 * Add compatibility options to the purpose and trust code. The
15552 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15553 accepts a certificate or CA, this was the previous behaviour,
15554 with all the associated security issues.
15556 X509_TRUST_COMPAT is the old trust behaviour: only and
15557 automatically trust self signed roots in certificate store. A
15558 new trust setting X509_TRUST_DEFAULT is used to specify that
15559 a purpose has no associated trust setting and it should instead
15560 use the value in the default purpose.
15564 * Fix the PKCS#8 DSA private key code so it decodes keys again
15565 and fix a memory leak.
15569 * In util/mkerr.pl (which implements 'make errors'), preserve
15570 reason strings from the previous version of the .c file, as
15571 the default to have only downcase letters (and digits) in
15572 automatically generated reasons codes is not always appropriate.
15576 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15577 using strerror. Previously, ERR_reason_error_string() returned
15578 library names as reason strings for SYSerr; but SYSerr is a special
15579 case where small numbers are errno values, not library numbers.
15583 * Add '-dsaparam' option to 'openssl dhparam' application. This
15584 converts DSA parameters into DH parameters. (When creating parameters,
15585 DSA_generate_parameters is used.)
15589 * Include 'length' (recommended exponent length) in C code generated
15590 by 'openssl dhparam -C'.
15594 * The second argument to set_label in perlasm was already being used
15595 so couldn't be used as a "file scope" flag. Moved to third argument
15600 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15601 instead of RAND_bytes for encryption IVs and salts.
15605 * Include RAND_status() into RAND_METHOD instead of implementing
15606 it only for md_rand.c Otherwise replacing the PRNG by calling
15607 RAND_set_rand_method would be impossible.
15611 * Don't let DSA_generate_key() enter an infinite loop if the random
15612 number generation fails.
15616 * New 'rand' application for creating pseudo-random output.
15620 * Added configuration support for Linux/IA64
15622 *Rolf Haberrecker <rolf@suse.de>*
15624 * Assembler module support for Mingw32.
15628 * Shared library support for HPUX (in shlib/).
15630 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15632 * Shared library support for Solaris gcc.
15634 *Lutz Behnke <behnke@trustcenter.de>*
15636 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15638 * PKCS7_encrypt() was adding text MIME headers twice because they
15639 were added manually and by SMIME_crlf_copy().
15643 * In bntest.c don't call BN_rand with zero bits argument.
15645 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15647 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15648 case was implemented. This caused BN_div_recp() to fail occasionally.
15652 * Add an optional second argument to the set_label() in the perl
15653 assembly language builder. If this argument exists and is set
15654 to 1 it signals that the assembler should use a symbol whose
15655 scope is the entire file, not just the current function. This
15656 is needed with MASM which uses the format label:: for this scope.
15658 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15660 * Change the ASN1 types so they are typedefs by default. Before
15661 almost all types were #define'd to ASN1_STRING which was causing
15662 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15667 * Change names of new functions to the new get1/get0 naming
15668 convention: After 'get1', the caller owns a reference count
15669 and has to call `..._free`; 'get0' returns a pointer to some
15670 data structure without incrementing reference counters.
15671 (Some of the existing 'get' functions increment a reference
15672 counter, some don't.)
15673 Similarly, 'set1' and 'add1' functions increase reference
15674 counters or duplicate objects.
15678 * Allow for the possibility of temp RSA key generation failure:
15679 the code used to assume it always worked and crashed on failure.
15683 * Fix potential buffer overrun problem in BIO_printf().
15684 *Ulf Möller, using public domain code by Patrick Powell; problem
15685 pointed out by David Sacerdote <das33@cornell.edu>*
15687 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15688 RAND_egd() and RAND_status(). In the command line application,
15689 the EGD socket can be specified like a seed file using RANDFILE
15694 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15695 Some CAs (e.g. Verisign) distribute certificates in this form.
15699 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15700 list to exclude them. This means that no special compilation option
15701 is needed to use anonymous DH: it just needs to be included in the
15706 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15707 EVP_MD_type. The old functionality is available in a new macro called
15708 EVP_MD_md(). Change code that uses it and update docs.
15712 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15713 where the `void *` argument is replaced by a function pointer argument.
15714 Previously `void *` was abused to point to functions, which works on
15715 many platforms, but is not correct. As these functions are usually
15716 called by macros defined in OpenSSL header files, most source code
15717 should work without changes.
15721 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15722 sections with information on -D... compiler switches used for
15723 compiling the library so that applications can see them. To enable
15724 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15725 must be defined. E.g.,
15726 #define OPENSSL_ALGORITHM_DEFINES
15727 #include <openssl/opensslconf.h>
15728 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15730 *Richard Levitte, Ulf and Bodo Möller*
15732 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15737 * Change the 'other' type in certificate aux info to a STACK_OF
15738 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15739 the required ASN1 format: arbitrary types determined by an OID.
15743 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15744 argument to 'req'. This is not because the function is newer or
15745 better than others it just uses the work 'NEW' in the certificate
15746 request header lines. Some software needs this.
15750 * Reorganise password command line arguments: now passwords can be
15751 obtained from various sources. Delete the PEM_cb function and make
15752 it the default behaviour: i.e. if the callback is NULL and the
15753 usrdata argument is not NULL interpret it as a null terminated pass
15754 phrase. If usrdata and the callback are NULL then the pass phrase
15755 is prompted for as usual.
15759 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15760 the support is automatically enabled. The resulting binaries will
15761 autodetect the card and use it if present.
15763 *Ben Laurie and Compaq Inc.*
15765 * Work around for Netscape hang bug. This sends certificate request
15766 and server done in one record. Since this is perfectly legal in the
15767 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15768 the bugs/SSLv3 entry for more info.
15772 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15776 * Add -rand argument to smime and pkcs12 applications and read/write
15781 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15785 * Add command line password options to the remaining applications.
15789 * Bug fix for BN_div_recp() for numerators with an even number of
15794 * More tests in bntest.c, and changed test_bn output.
15798 * ./config recognizes MacOS X now.
15802 * Bug fix for BN_div() when the first words of num and divisor are
15803 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15807 * Add support for various broken PKCS#8 formats, and command line
15808 options to produce them.
15812 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15813 get temporary BIGNUMs from a BN_CTX.
15817 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15822 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15823 include a #define from the old name to the new. The original intent
15824 was that statically linked binaries could for example just call
15825 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15826 link with digests. This never worked because SSLeay_add_all_digests()
15827 and SSLeay_add_all_ciphers() were in the same source file so calling
15828 one would link with the other. They are now in separate source files.
15832 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15836 * Use a less unusual form of the Miller-Rabin primality test (it used
15837 a binary algorithm for exponentiation integrated into the Miller-Rabin
15838 loop, our standard modexp algorithms are faster).
15842 * Support for the EBCDIC character set completed.
15844 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15846 * Source code cleanups: use const where appropriate, eliminate casts,
15847 use `void *` instead of `char *` in lhash.
15851 * Bugfix: ssl3_send_server_key_exchange was not restartable
15852 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15853 this the server could overwrite ephemeral keys that the client
15858 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15859 using 50 iterations of the Rabin-Miller test.
15861 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15862 iterations of the Rabin-Miller test as required by the appendix
15863 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15864 As BN_is_prime_fasttest includes trial division, DSA parameter
15865 generation becomes much faster.
15867 This implies a change for the callback functions in DSA_is_prime
15868 and DSA_generate_parameters: The callback function is called once
15869 for each positive witness in the Rabin-Miller test, not just
15870 occasionally in the inner loop; and the parameters to the
15871 callback function now provide an iteration count for the outer
15872 loop rather than for the current invocation of the inner loop.
15873 DSA_generate_parameters additionally can call the callback
15874 function with an 'iteration count' of -1, meaning that a
15875 candidate has passed the trial division test (when q is generated
15876 from an application-provided seed, trial division is skipped).
15880 * New function BN_is_prime_fasttest that optionally does trial
15881 division before starting the Rabin-Miller test and has
15882 an additional BN_CTX * argument (whereas BN_is_prime always
15883 has to allocate at least one BN_CTX).
15884 'callback(1, -1, cb_arg)' is called when a number has passed the
15885 trial division stage.
15889 * Fix for bug in CRL encoding. The validity dates weren't being handled
15894 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15898 * New function BN_pseudo_rand().
15902 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15903 bignum version of BN_from_montgomery() with the working code from
15904 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15909 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15910 made it impossible to use the same SSL_SESSION data structure in
15911 SSL2 clients in multiple threads.
15915 * The return value of RAND_load_file() no longer counts bytes obtained
15916 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15917 to seed the PRNG (previously an explicit byte count was required).
15919 *Ulf Möller, Bodo Möller*
15921 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15922 used `char *` instead of `void *` and had casts all over the place.
15926 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15930 * Retain source code compatibility for BN_prime_checks macro:
15931 BN_is_prime(..., BN_prime_checks, ...) now uses
15932 BN_prime_checks_for_size to determine the appropriate number of
15933 Rabin-Miller iterations.
15937 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15938 DH_CHECK_P_NOT_SAFE_PRIME.
15939 (Check if this is true? OpenPGP calls them "strong".)
15943 * Merge the functionality of "dh" and "gendh" programs into a new program
15944 "dhparam". The old programs are retained for now but will handle DH keys
15945 (instead of parameters) in future.
15949 * Make the ciphers, s_server and s_client programs check the return values
15950 when a new cipher list is set.
15954 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15955 ciphers. Before when the 56bit ciphers were enabled the sorting was
15958 The syntax for the cipher sorting has been extended to support sorting by
15959 cipher-strength (using the strength_bits hard coded in the tables).
15960 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15962 Fix a bug in the cipher-command parser: when supplying a cipher command
15963 string with an "undefined" symbol (neither command nor alphanumeric
15964 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15965 an error is flagged.
15967 Due to the strength-sorting extension, the code of the
15968 ssl_create_cipher_list() function was completely rearranged. I hope that
15969 the readability was also increased :-)
15971 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15973 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15974 for the first serial number and places 2 in the serial number file. This
15975 avoids problems when the root CA is created with serial number zero and
15976 the first user certificate has the same issuer name and serial number
15981 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15982 the new code. Add documentation for this stuff.
15986 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15987 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15988 structures and behave in an analogous way to the X509v3 functions:
15989 they shouldn't be called directly but wrapper functions should be used
15992 So we also now have some wrapper functions that call the X509at functions
15993 when passed certificate requests. (TO DO: similar things can be done with
15994 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15995 things. Some of these need some d2i or i2d and print functionality
15996 because they handle more complex structures.)
16000 * Add missing #ifndefs that caused missing symbols when building libssl
16001 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16002 NO_RSA in `ssl/s2*.c`.
16004 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16006 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16007 has a return value which indicates the quality of the random data
16008 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16009 error queue. New function RAND_pseudo_bytes() generates output that is
16010 guaranteed to be unique but not unpredictable. RAND_add is like
16011 RAND_seed, but takes an extra argument for an entropy estimate
16012 (RAND_seed always assumes full entropy).
16016 * Do more iterations of Rabin-Miller probable prime test (specifically,
16017 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16018 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16019 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16020 false-positive rate of at most 2^-80 for random input.
16024 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16028 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16029 in the 0.9.5 release), this returns the chain
16030 from an X509_CTX structure with a dup of the stack and all
16031 the X509 reference counts upped: so the stack will exist
16032 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16035 Also make SSL_SESSION_print() print out the verify return
16040 * Add manpage for the pkcs12 command. Also change the default
16041 behaviour so MAC iteration counts are used unless the new
16042 -nomaciter option is used. This improves file security and
16043 only older versions of MSIE (4.0 for example) need it.
16047 * Honor the no-xxx Configure options when creating .DEF files.
16051 * Add PKCS#10 attributes to field table: challengePassword,
16052 unstructuredName and unstructuredAddress. These are taken from
16053 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16054 international characters are used.
16056 More changes to X509_ATTRIBUTE code: allow the setting of types
16057 based on strings. Remove the 'loc' parameter when adding
16058 attributes because these will be a SET OF encoding which is sorted
16063 * Initial changes to the 'req' utility to allow request generation
16064 automation. This will allow an application to just generate a template
16065 file containing all the field values and have req construct the
16068 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16069 used all over the place including certificate requests and PKCS#7
16070 structures. They are currently handled manually where necessary with
16071 some primitive wrappers for PKCS#7. The new functions behave in a
16072 manner analogous to the X509 extension functions: they allow
16073 attributes to be looked up by NID and added.
16075 Later something similar to the X509V3 code would be desirable to
16076 automatically handle the encoding, decoding and printing of the
16077 more complex types. The string types like challengePassword can
16078 be handled by the string table functions.
16080 Also modified the multi byte string table handling. Now there is
16081 a 'global mask' which masks out certain types. The table itself
16082 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16083 is useful when for example there is only one permissible type
16084 (as in countryName) and using the mask might result in no valid
16089 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16090 SSL_get_peer_finished to allow applications to obtain the latest
16091 Finished messages sent to the peer or expected from the peer,
16092 respectively. (SSL_get_peer_finished is usually the Finished message
16093 actually received from the peer, otherwise the protocol will be aborted.)
16095 As the Finished message are message digests of the complete handshake
16096 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16097 be used for external authentication procedures when the authentication
16098 provided by SSL/TLS is not desired or is not enough.
16102 * Enhanced support for Alpha Linux is added. Now ./config checks if
16103 the host supports BWX extension and if Compaq C is present on the
16104 $PATH. Just exploiting of the BWX extension results in 20-30%
16105 performance kick for some algorithms, e.g. DES and RC4 to mention
16106 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16111 * Add support for MS "fast SGC". This is arguably a violation of the
16112 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16113 weak crypto and after checking the certificate is SGC a second one
16114 with strong crypto. MS SGC stops the first handshake after receiving
16115 the server certificate message and sends a second client hello. Since
16116 a server will typically do all the time consuming operations before
16117 expecting any further messages from the client (server key exchange
16118 is the most expensive) there is little difference between the two.
16120 To get OpenSSL to support MS SGC we have to permit a second client
16121 hello message after we have sent server done. In addition we have to
16122 reset the MAC if we do get this second client hello.
16126 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16127 if a DER encoded private key is RSA or DSA traditional format. Changed
16128 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16129 format DER encoded private key. Newer code should use PKCS#8 format which
16130 has the key type encoded in the ASN1 structure. Added DER private key
16131 support to pkcs8 application.
16135 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16136 ciphersuites has been selected (as required by the SSL 3/TLS 1
16137 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16138 is set, we interpret this as a request to violate the specification
16139 (the worst that can happen is a handshake failure, and 'correct'
16140 behaviour would result in a handshake failure anyway).
16144 * In SSL_CTX_add_session, take into account that there might be multiple
16145 SSL_SESSION structures with the same session ID (e.g. when two threads
16146 concurrently obtain them from an external cache).
16147 The internal cache can handle only one SSL_SESSION with a given ID,
16148 so if there's a conflict, we now throw out the old one to achieve
16153 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16154 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16155 some routines that use cipher OIDs: some ciphers do not have OIDs
16156 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16161 * Simplify the trust setting structure and code. Now we just have
16162 two sequences of OIDs for trusted and rejected settings. These will
16163 typically have values the same as the extended key usage extension
16164 and any application specific purposes.
16166 The trust checking code now has a default behaviour: it will just
16167 check for an object with the same NID as the passed id. Functions can
16168 be provided to override either the default behaviour or the behaviour
16169 for a given id. SSL client, server and email already have functions
16170 in place for compatibility: they check the NID and also return "trusted"
16171 if the certificate is self signed.
16175 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16176 traditional format into an EVP_PKEY structure.
16180 * Add a password callback function PEM_cb() which either prompts for
16181 a password if usr_data is NULL or otherwise assumes it is a null
16182 terminated password. Allow passwords to be passed on command line
16183 environment or config files in a few more utilities.
16187 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16188 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16189 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16190 Update documentation.
16194 * Support for ASN1 "NULL" type. This could be handled before by using
16195 ASN1_TYPE but there wasn't any function that would try to read a NULL
16196 and produce an error if it couldn't. For compatibility we also have
16197 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16198 don't allocate anything because they don't need to.
16202 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16205 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16207 * Rebuild of the memory allocation routines used by OpenSSL code and
16208 possibly others as well. The purpose is to make an interface that
16209 provide hooks so anyone can build a separate set of allocation and
16210 deallocation routines to be used by OpenSSL, for example memory
16211 pool implementations, or something else, which was previously hard
16212 since Malloc(), Realloc() and Free() were defined as macros having
16213 the values malloc, realloc and free, respectively (except for Win32
16214 compilations). The same is provided for memory debugging code.
16215 OpenSSL already comes with functionality to find memory leaks, but
16216 this gives people a chance to debug other memory problems.
16218 With these changes, a new set of functions and macros have appeared:
16220 CRYPTO_set_mem_debug_functions() [F]
16221 CRYPTO_get_mem_debug_functions() [F]
16222 CRYPTO_dbg_set_options() [F]
16223 CRYPTO_dbg_get_options() [F]
16224 CRYPTO_malloc_debug_init() [M]
16226 The memory debug functions are NULL by default, unless the library
16227 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16228 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16229 gives the standard debugging functions that come with OpenSSL) or
16230 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16231 provided by the library user) must be used. When the standard
16232 debugging functions are used, CRYPTO_dbg_set_options can be used to
16233 request additional information:
16234 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16235 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16237 Also, things like CRYPTO_set_mem_functions will always give the
16238 expected result (the new set of functions is used for allocation
16239 and deallocation) at all times, regardless of platform and compiler
16242 To finish it up, some functions that were never use in any other
16243 way than through macros have a new API and new semantic:
16245 CRYPTO_dbg_malloc()
16246 CRYPTO_dbg_realloc()
16249 All macros of value have retained their old syntax.
16251 *Richard Levitte and Bodo Moeller*
16253 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16254 ordering of SMIMECapabilities wasn't in "strength order" and there
16255 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16260 * Some ASN1 types with illegal zero length encoding (INTEGER,
16261 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16263 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16265 * Merge in my S/MIME library for OpenSSL. This provides a simple
16266 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16267 functionality to handle multipart/signed properly) and a utility
16268 called 'smime' to call all this stuff. This is based on code I
16269 originally wrote for Celo who have kindly allowed it to be
16270 included in OpenSSL.
16274 * Add variants des_set_key_checked and des_set_key_unchecked of
16275 des_set_key (aka des_key_sched). Global variable des_check_key
16276 decides which of these is called by des_set_key; this way
16277 des_check_key behaves as it always did, but applications and
16278 the library itself, which was buggy for des_check_key == 1,
16279 have a cleaner way to pick the version they need.
16283 * New function PKCS12_newpass() which changes the password of a
16288 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16289 dynamic mix. In both cases the ids can be used as an index into the
16290 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16291 functions so they accept a list of the field values and the
16292 application doesn't need to directly manipulate the X509_TRUST
16297 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16302 * Modify the way the V3 extension code looks up extensions. This now
16303 works in a similar way to the object code: we have some "standard"
16304 extensions in a static table which is searched with OBJ_bsearch()
16305 and the application can add dynamic ones if needed. The file
16306 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16307 updated whenever a new extension is added to the core code and kept
16308 in ext_nid order. There is a simple program 'tabtest.c' which checks
16309 this. New extensions are not added too often so this file can readily
16310 be maintained manually.
16312 There are two big advantages in doing things this way. The extensions
16313 can be looked up immediately and no longer need to be "added" using
16314 X509V3_add_standard_extensions(): this function now does nothing.
16315 Side note: I get *lots* of email saying the extension code doesn't
16316 work because people forget to call this function.
16317 Also no dynamic allocation is done unless new extensions are added:
16318 so if we don't add custom extensions there is no need to call
16319 X509V3_EXT_cleanup().
16323 * Modify enc utility's salting as follows: make salting the default. Add a
16324 magic header, so unsalted files fail gracefully instead of just decrypting
16325 to garbage. This is because not salting is a big security hole, so people
16326 should be discouraged from doing it.
16330 * Fixes and enhancements to the 'x509' utility. It allowed a message
16331 digest to be passed on the command line but it only used this
16332 parameter when signing a certificate. Modified so all relevant
16333 operations are affected by the digest parameter including the
16334 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16335 DSA key was used because it didn't fix the digest.
16339 * Initial certificate chain verify code. Currently tests the untrusted
16340 certificates for consistency with the verify purpose (which is set
16341 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16343 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16344 this is because it will reject chains with invalid extensions whereas
16345 every previous version of OpenSSL and SSLeay made no checks at all.
16347 Trust code: checks the root CA for the relevant trust settings. Trust
16348 settings have an initial value consistent with the verify purpose: e.g.
16349 if the verify purpose is for SSL client use it expects the CA to be
16350 trusted for SSL client use. However the default value can be changed to
16351 permit custom trust settings: one example of this would be to only trust
16352 certificates from a specific "secure" set of CAs.
16354 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16355 which should be used for version portability: especially since the
16356 verify structure is likely to change more often now.
16358 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16359 to set them. If not set then assume SSL clients will verify SSL servers
16362 Two new options to the verify program: -untrusted allows a set of
16363 untrusted certificates to be passed in and -purpose which sets the
16364 intended purpose of the certificate. If a purpose is set then the
16365 new chain verify code is used to check extension consistency.
16369 * Support for the authority information access extension.
16373 * Modify RSA and DSA PEM read routines to transparently handle
16374 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16375 public keys in a format compatible with certificate
16376 SubjectPublicKeyInfo structures. Unfortunately there were already
16377 functions called *_PublicKey_* which used various odd formats so
16378 these are retained for compatibility: however the DSA variants were
16379 never in a public release so they have been deleted. Changed dsa/rsa
16380 utilities to handle the new format: note no releases ever handled public
16381 keys so we should be OK.
16383 The primary motivation for this change is to avoid the same fiasco
16384 that dogs private keys: there are several incompatible private key
16385 formats some of which are standard and some OpenSSL specific and
16386 require various evil hacks to allow partial transparent handling and
16387 even then it doesn't work with DER formats. Given the option anything
16388 other than PKCS#8 should be dumped: but the other formats have to
16389 stay in the name of compatibility.
16391 With public keys and the benefit of hindsight one standard format
16392 is used which works with EVP_PKEY, RSA or DSA structures: though
16393 it clearly returns an error if you try to read the wrong kind of key.
16395 Added a -pubkey option to the 'x509' utility to output the public key.
16396 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16397 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16398 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16399 that do the same as the `EVP_PKEY_assign_*()` except they up the
16400 reference count of the added key (they don't "swallow" the
16405 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16406 CRLs would fail if the file contained no certificates or no CRLs:
16407 added a new function to read in both types and return the number
16408 read: this means that if none are read it will be an error. The
16409 DER versions of the certificate and CRL reader would always fail
16410 because it isn't possible to mix certificates and CRLs in DER format
16411 without choking one or the other routine. Changed this to just read
16412 a certificate: this is the best we can do. Also modified the code
16413 in `apps/verify.c` to take notice of return codes: it was previously
16414 attempting to read in certificates from NULL pointers and ignoring
16415 any errors: this is one reason why the cert and CRL reader seemed
16416 to work. It doesn't check return codes from the default certificate
16417 routines: these may well fail if the certificates aren't installed.
16421 * Code to support otherName option in GeneralName.
16425 * First update to verify code. Change the verify utility
16426 so it warns if it is passed a self signed certificate:
16427 for consistency with the normal behaviour. X509_verify
16428 has been modified to it will now verify a self signed
16429 certificate if *exactly* the same certificate appears
16430 in the store: it was previously impossible to trust a
16431 single self signed certificate. This means that:
16432 openssl verify ss.pem
16433 now gives a warning about a self signed certificate but
16434 openssl verify -CAfile ss.pem ss.pem
16439 * For servers, store verify_result in SSL_SESSION data structure
16440 (and add it to external session representation).
16441 This is needed when client certificate verifications fails,
16442 but an application-provided verification callback (set by
16443 SSL_CTX_set_cert_verify_callback) allows accepting the session
16444 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16445 but returns 1): When the session is reused, we have to set
16446 ssl->verify_result to the appropriate error code to avoid
16449 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16451 * Fix a bug in the new PKCS#7 code: it didn't consider the
16452 case in PKCS7_dataInit() where the signed PKCS7 structure
16453 didn't contain any existing data because it was being created.
16455 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16457 * Add a salt to the key derivation routines in enc.c. This
16458 forms the first 8 bytes of the encrypted file. Also add a
16459 -S option to allow a salt to be input on the command line.
16463 * New function X509_cmp(). Oddly enough there wasn't a function
16464 to compare two certificates. We do this by working out the SHA1
16465 hash and comparing that. X509_cmp() will be needed by the trust
16470 * SSL_get1_session() is like SSL_get_session(), but increments
16471 the reference count in the SSL_SESSION returned.
16473 *Geoff Thorpe <geoff@eu.c2.net>*
16475 * Fix for 'req': it was adding a null to request attributes.
16476 Also change the X509_LOOKUP and X509_INFO code to handle
16477 certificate auxiliary information.
16481 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16486 * Add the possibility to add extra information to the memory leak
16487 detecting output, to form tracebacks, showing from where each
16488 allocation was originated: CRYPTO_push_info("constant string") adds
16489 the string plus current file name and line number to a per-thread
16490 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16491 is like calling CYRPTO_pop_info() until the stack is empty.
16492 Also updated memory leak detection code to be multi-thread-safe.
16496 * Add options -text and -noout to pkcs7 utility and delete the
16497 encryption options which never did anything. Update docs.
16501 * Add options to some of the utilities to allow the pass phrase
16502 to be included on either the command line (not recommended on
16503 OSes like Unix) or read from the environment. Update the
16504 manpages and fix a few bugs.
16508 * Add a few manpages for some of the openssl commands.
16512 * Fix the -revoke option in ca. It was freeing up memory twice,
16513 leaking and not finding already revoked certificates.
16517 * Extensive changes to support certificate auxiliary information.
16518 This involves the use of X509_CERT_AUX structure and X509_AUX
16519 functions. An X509_AUX function such as PEM_read_X509_AUX()
16520 can still read in a certificate file in the usual way but it
16521 will also read in any additional "auxiliary information". By
16522 doing things this way a fair degree of compatibility can be
16523 retained: existing certificates can have this information added
16524 using the new 'x509' options.
16526 Current auxiliary information includes an "alias" and some trust
16527 settings. The trust settings will ultimately be used in enhanced
16528 certificate chain verification routines: currently a certificate
16529 can only be trusted if it is self signed and then it is trusted
16534 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16535 The problem was that one of the replacement routines had not been working
16536 since SSLeay releases. For now the offending routine has been replaced
16537 with non-optimised assembler. Even so, this now gives around 95%
16538 performance improvement for 1024 bit RSA signs.
16542 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16543 handling. Most clients have the effective key size in bits equal to
16544 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16545 A few however don't do this and instead use the size of the decrypted key
16546 to determine the RC2 key length and the AlgorithmIdentifier to determine
16547 the effective key length. In this case the effective key length can still
16548 be 40 bits but the key length can be 168 bits for example. This is fixed
16549 by manually forcing an RC2 key into the EVP_PKEY structure because the
16550 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16551 the key length and effective key length are equal.
16555 * Add a bunch of functions that should simplify the creation of
16556 X509_NAME structures. Now you should be able to do:
16557 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16558 and have it automatically work out the correct field type and fill in
16559 the structures. The more adventurous can try:
16560 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16561 and it will (hopefully) work out the correct multibyte encoding.
16565 * Change the 'req' utility to use the new field handling and multibyte
16566 copy routines. Before the DN field creation was handled in an ad hoc
16567 way in req, ca, and x509 which was rather broken and didn't support
16568 BMPStrings or UTF8Strings. Since some software doesn't implement
16569 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16570 using the dirstring_type option. See the new comment in the default
16571 openssl.cnf for more info.
16575 * Make crypto/rand/md_rand.c more robust:
16576 - Assure unique random numbers after fork().
16577 - Make sure that concurrent threads access the global counter and
16578 md serializably so that we never lose entropy in them
16579 or use exactly the same state in multiple threads.
16580 Access to the large state is not always serializable because
16581 the additional locking could be a performance killer, and
16582 md should be large enough anyway.
16586 * New file `apps/app_rand.c` with commonly needed functionality
16587 for handling the random seed file.
16589 Use the random seed file in some applications that previously did not:
16591 dsaparam -genkey (which also ignored its '-rand' option),
16594 x509 (when signing).
16595 Except on systems with /dev/urandom, it is crucial to have a random
16596 seed file at least for key creation, DSA signing, and for DH exchanges;
16597 for RSA signatures we could do without one.
16599 gendh and gendsa (unlike genrsa) used to read only the first byte
16600 of each file listed in the '-rand' option. The function as previously
16601 found in genrsa is now in app_rand.c and is used by all programs
16602 that support '-rand'.
16606 * In RAND_write_file, use mode 0600 for creating files;
16607 don't just chmod when it may be too late.
16611 * Report an error from X509_STORE_load_locations
16612 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16616 * New function ASN1_mbstring_copy() this copies a string in either
16617 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16618 into an ASN1_STRING type. A mask of permissible types is passed
16619 and it chooses the "minimal" type to use or an error if not type
16624 * Add function equivalents to the various macros in asn1.h. The old
16625 macros are retained with an `M_` prefix. Code inside the library can
16626 use the `M_` macros. External code (including the openssl utility)
16627 should *NOT* in order to be "shared library friendly".
16631 * Add various functions that can check a certificate's extensions
16632 to see if it usable for various purposes such as SSL client,
16633 server or S/MIME and CAs of these types. This is currently
16634 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16635 verification. Also added a -purpose flag to x509 utility to
16636 print out all the purposes.
16640 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16645 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16646 for, obtain and decode and extension and obtain its critical flag.
16647 This allows all the necessary extension code to be handled in a
16648 single function call.
16652 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16653 platforms. See crypto/rc4/rc4_enc.c for further details.
16657 * New -noout option to asn1parse. This causes no output to be produced
16658 its main use is when combined with -strparse and -out to extract data
16659 from a file (which may not be in ASN.1 format).
16663 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16664 when producing the local key id.
16666 *Richard Levitte <levitte@stacken.kth.se>*
16668 * New option -dhparam in s_server. This allows a DH parameter file to be
16669 stated explicitly. If it is not stated then it tries the first server
16670 certificate file. The previous behaviour hard coded the filename
16675 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16676 a public key to be input or output. For example:
16677 openssl rsa -in key.pem -pubout -out pubkey.pem
16678 Also added necessary DSA public key functions to handle this.
16682 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16683 in the message. This was handled by allowing
16684 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16686 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16688 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16689 to the end of the strings whereas this didn't. This would cause problems
16690 if strings read with d2i_ASN1_bytes() were later modified.
16692 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16694 * Fix for base64 decode bug. When a base64 bio reads only one line of
16695 data and it contains EOF it will end up returning an error. This is
16696 caused by input 46 bytes long. The cause is due to the way base64
16697 BIOs find the start of base64 encoded data. They do this by trying a
16698 trial decode on each line until they find one that works. When they
16699 do a flag is set and it starts again knowing it can pass all the
16700 data directly through the decoder. Unfortunately it doesn't reset
16701 the context it uses. This means that if EOF is reached an attempt
16702 is made to pass two EOFs through the context and this causes the
16703 resulting error. This can also cause other problems as well. As is
16704 usual with these problems it takes *ages* to find and the fix is
16705 trivial: move one line.
16707 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16709 * Ugly workaround to get s_client and s_server working under Windows. The
16710 old code wouldn't work because it needed to select() on sockets and the
16711 tty (for keypresses and to see if data could be written). Win32 only
16712 supports select() on sockets so we select() with a 1s timeout on the
16713 sockets and then see if any characters are waiting to be read, if none
16714 are present then we retry, we also assume we can always write data to
16715 the tty. This isn't nice because the code then blocks until we've
16716 received a complete line of data and it is effectively polling the
16717 keyboard at 1s intervals: however it's quite a bit better than not
16718 working at all :-) A dedicated Windows application might handle this
16719 with an event loop for example.
16723 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16724 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16725 will be called when RSA_sign() and RSA_verify() are used. This is useful
16726 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16727 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16728 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16729 This necessitated the support of an extra signature type NID_md5_sha1
16730 for SSL signatures and modifications to the SSL library to use it instead
16731 of calling RSA_public_decrypt() and RSA_private_encrypt().
16735 * Add new -verify -CAfile and -CApath options to the crl program, these
16736 will lookup a CRL issuers certificate and verify the signature in a
16737 similar way to the verify program. Tidy up the crl program so it
16738 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16739 less strict. It will now permit CRL extensions even if it is not
16740 a V2 CRL: this will allow it to tolerate some broken CRLs.
16744 * Initialize all non-automatic variables each time one of the openssl
16745 sub-programs is started (this is necessary as they may be started
16746 multiple times from the "OpenSSL>" prompt).
16748 *Lennart Bang, Bodo Moeller*
16750 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16751 removing all other RSA functionality (this is what NO_RSA does). This
16752 is so (for example) those in the US can disable those operations covered
16753 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16758 * Non-copying interface to BIO pairs.
16759 (still largely untested)
16763 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16764 ASCII string. This was handled independently in various places before.
16768 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16769 UTF8 strings a character at a time.
16773 * Use client_version from client hello to select the protocol
16774 (s23_srvr.c) and for RSA client key exchange verification
16775 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16779 * Add various utility functions to handle SPKACs, these were previously
16780 handled by poking round in the structure internals. Added new function
16781 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16782 print, verify and generate SPKACs. Based on an original idea from
16783 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16787 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16791 * Allow the config file extension section to be overwritten on the
16792 command line. Based on an original idea from Massimiliano Pala
16793 <madwolf@comune.modena.it>. The new option is called -extensions
16794 and can be applied to ca, req and x509. Also -reqexts to override
16795 the request extensions in req and -crlexts to override the crl extensions
16800 * Add new feature to the SPKAC handling in ca. Now you can include
16801 the same field multiple times by preceding it by "XXXX." for example:
16804 this is the same syntax as used in the req config file.
16808 * Allow certificate extensions to be added to certificate requests. These
16809 are specified in a 'req_extensions' option of the req section of the
16810 config file. They can be printed out with the -text option to req but
16811 are otherwise ignored at present.
16815 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16816 data read consists of only the final block it would not decrypted because
16817 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16818 A misplaced 'break' also meant the decrypted final block might not be
16819 copied until the next read.
16823 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16824 a few extra parameters to the DH structure: these will be useful if
16825 for example we want the value of 'q' or implement X9.42 DH.
16829 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16830 provides hooks that allow the default DSA functions or functions on a
16831 "per key" basis to be replaced. This allows hardware acceleration and
16832 hardware key storage to be handled without major modification to the
16833 library. Also added low level modexp hooks and CRYPTO_EX structure and
16834 associated functions.
16838 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16839 as "read only": it can't be written to and the buffer it points to will
16840 not be freed. Reading from a read only BIO is much more efficient than
16841 a normal memory BIO. This was added because there are several times when
16842 an area of memory needs to be read from a BIO. The previous method was
16843 to create a memory BIO and write the data to it, this results in two
16844 copies of the data and an O(n^2) reading algorithm. There is a new
16845 function BIO_new_mem_buf() which creates a read only memory BIO from
16846 an area of memory. Also modified the PKCS#7 routines to use read only
16851 * Bugfix: ssl23_get_client_hello did not work properly when called in
16852 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16853 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16854 but a retry condition occurred while trying to read the rest.
16858 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16859 NID_pkcs7_encrypted by default: this was wrong since this should almost
16860 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16861 the encrypted data type: this is a more sensible place to put it and it
16862 allows the PKCS#12 code to be tidied up that duplicated this
16867 * Changed obj_dat.pl script so it takes its input and output files on
16868 the command line. This should avoid shell escape redirection problems
16873 * Initial support for certificate extension requests, these are included
16874 in things like Xenroll certificate requests. Included functions to allow
16875 extensions to be obtained and added.
16879 * -crlf option to s_client and s_server for sending newlines as
16880 CRLF (as required by many protocols).
16884 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16886 * Install libRSAglue.a when OpenSSL is built with RSAref.
16888 *Ralf S. Engelschall*
16890 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16892 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16894 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16899 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16900 DH parameters/keys (q is lost during that conversion, but the resulting
16901 DH parameters contain its length).
16903 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16904 much faster than DH_generate_parameters (which creates parameters
16905 where `p = 2*q + 1`), and also the smaller q makes DH computations
16906 much more efficient (160-bit exponentiation instead of 1024-bit
16907 exponentiation); so this provides a convenient way to support DHE
16908 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16909 utter importance to use
16910 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16912 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16913 when such DH parameters are used, because otherwise small subgroup
16914 attacks may become possible!
16918 * Avoid memory leak in i2d_DHparams.
16922 * Allow the -k option to be used more than once in the enc program:
16923 this allows the same encrypted message to be read by multiple recipients.
16927 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16928 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16929 it will always use the numerical form of the OID, even if it has a short
16934 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16935 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16936 otherwise bn_mod_exp was called. In the case of hardware keys for example
16937 no private key components need be present and it might store extra data
16938 in the RSA structure, which cannot be accessed from bn_mod_exp.
16939 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16940 private key operations.
16944 * Added support for SPARC Linux.
16948 * pem_password_cb function type incompatibly changed from
16949 typedef int pem_password_cb(char *buf, int size, int rwflag);
16951 ....(char *buf, int size, int rwflag, void *userdata);
16952 so that applications can pass data to their callbacks:
16953 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16954 additional void * argument, which is just handed through whenever
16955 the password callback is called.
16957 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16959 New function SSL_CTX_set_default_passwd_cb_userdata.
16961 Compatibility note: As many C implementations push function arguments
16962 onto the stack in reverse order, the new library version is likely to
16963 interoperate with programs that have been compiled with the old
16964 pem_password_cb definition (PEM_whatever takes some data that
16965 happens to be on the stack as its last argument, and the callback
16966 just ignores this garbage); but there is no guarantee whatsoever that
16969 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16970 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16971 problems not only on Windows, but also on some Unix platforms.
16972 To avoid problematic command lines, these definitions are now in an
16973 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16974 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16978 * MIPS III/IV assembler module is reimplemented.
16982 * More DES library cleanups: remove references to srand/rand and
16983 delete an unused file.
16987 * Add support for the free Netwide assembler (NASM) under Win32,
16988 since not many people have MASM (ml) and it can be hard to obtain.
16989 This is currently experimental but it seems to work OK and pass all
16990 the tests. Check out INSTALL.W32 for info.
16994 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16995 without temporary keys kept an extra copy of the server key,
16996 and connections with temporary keys did not free everything in case
17001 * New function RSA_check_key and new openssl rsa option -check
17002 for verifying the consistency of RSA keys.
17004 *Ulf Moeller, Bodo Moeller*
17006 * Various changes to make Win32 compile work:
17007 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17008 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17009 comparison" warnings.
17010 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17014 * Add a debugging option to PKCS#5 v2 key generation function: when
17015 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17016 derived keys are printed to stderr.
17020 * Copy the flags in ASN1_STRING_dup().
17022 *Roman E. Pavlov <pre@mo.msk.ru>*
17024 * The x509 application mishandled signing requests containing DSA
17025 keys when the signing key was also DSA and the parameters didn't match.
17027 It was supposed to omit the parameters when they matched the signing key:
17028 the verifying software was then supposed to automatically use the CA's
17029 parameters if they were absent from the end user certificate.
17031 Omitting parameters is no longer recommended. The test was also
17032 the wrong way round! This was probably due to unusual behaviour in
17033 EVP_cmp_parameters() which returns 1 if the parameters match.
17034 This meant that parameters were omitted when they *didn't* match and
17035 the certificate was useless. Certificates signed with 'ca' didn't have
17038 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17040 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17041 The interface is as follows:
17042 Applications can use
17043 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17044 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17045 "off" is now the default.
17046 The library internally uses
17047 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17048 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17049 to disable memory-checking temporarily.
17051 Some inconsistent states that previously were possible (and were
17052 even the default) are now avoided.
17054 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17055 with each memory chunk allocated; this is occasionally more helpful
17056 than just having a counter.
17058 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17060 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17065 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17066 which largely parallels "options", but is for changing API behaviour,
17067 whereas "options" are about protocol behaviour.
17068 Initial "mode" flags are:
17070 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17071 a single record has been written.
17072 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17073 retries use the same buffer location.
17074 (But all of the contents must be
17079 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17082 * Fix problems with no-hmac etc.
17084 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17086 * New functions RSA_get_default_method(), RSA_set_method() and
17087 RSA_get_method(). These allows replacement of RSA_METHODs without having
17088 to mess around with the internals of an RSA structure.
17092 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17093 Also really enable memory leak checks in openssl.c and in some
17096 *Chad C. Mulligan, Bodo Moeller*
17098 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17099 up the length of negative integers. This has now been simplified to just
17100 store the length when it is first determined and use it later, rather
17101 than trying to keep track of where data is copied and updating it to
17103 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17105 * Add a new function PKCS7_signatureVerify. This allows the verification
17106 of a PKCS#7 signature but with the signing certificate passed to the
17107 function itself. This contrasts with PKCS7_dataVerify which assumes the
17108 certificate is present in the PKCS#7 structure. This isn't always the
17109 case: certificates can be omitted from a PKCS#7 structure and be
17110 distributed by "out of band" means (such as a certificate database).
17114 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17115 function prototypes in pem.h, also change util/mkdef.pl to add the
17116 necessary function names.
17120 * mk1mf.pl (used by Windows builds) did not properly read the
17121 options set by Configure in the top level Makefile, and Configure
17122 was not even able to write more than one option correctly.
17123 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17127 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17128 file to be loaded from a BIO or FILE pointer. The BIO version will
17129 for example allow memory BIOs to contain config info.
17133 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17134 Whoever hopes to achieve shared-library compatibility across versions
17135 must use this, not the compile-time macro.
17136 (Exercise 0.9.4: Which is the minimum library version required by
17138 Note: All this applies only to multi-threaded programs, others don't
17143 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17144 through a BIO pair triggered the default case, i.e.
17145 SSLerr(...,SSL_R_UNKNOWN_STATE).
17149 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17150 can use the SSL library even if none of the specific BIOs is
17155 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17156 for the encoded length.
17158 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17160 * Add initial documentation of the X509V3 functions.
17164 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17165 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17166 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17167 secure PKCS#8 private key format with a high iteration count.
17171 * Fix determination of Perl interpreter: A perl or perl5
17172 *directory* in $PATH was also accepted as the interpreter.
17174 *Ralf S. Engelschall*
17176 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17177 wrong with it but it was very old and did things like calling
17178 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17179 unusual formatting.
17183 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17184 to use the new extension code.
17188 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17189 with macros. This should make it easier to change their form, add extra
17190 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17195 * Add to configuration table a new entry that can specify an alternative
17196 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17197 according to Mark Crispin <MRC@Panda.COM>.
17201 * DES CBC did not update the IV. Weird.
17205 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17206 Changing the behaviour of the former might break existing programs --
17207 where IV updating is needed, des_ncbc_encrypt can be used.
17210 * When bntest is run from "make test" it drives bc to check its
17211 calculations, as well as internally checking them. If an internal check
17212 fails, it needs to cause bc to give a non-zero result or make test carries
17213 on without noticing the failure. Fixed.
17217 * DES library cleanups.
17221 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17222 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17223 ciphers. NOTE: although the key derivation function has been verified
17224 against some published test vectors it has not been extensively tested
17225 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17230 * Instead of "mkdir -p", which is not fully portable, use new
17231 Perl script "util/mkdir-p.pl".
17235 * Rewrite the way password based encryption (PBE) is handled. It used to
17236 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17237 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17238 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17239 the 'parameter' field of the AlgorithmIdentifier is passed to the
17240 underlying key generation function so it must do its own ASN1 parsing.
17241 This has also changed the EVP_PBE_CipherInit() function which now has a
17242 'parameter' argument instead of literal salt and iteration count values
17243 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17247 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17248 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17249 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17250 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17251 value was just used as a "magic string" and not used directly its
17252 value doesn't matter.
17256 * Introduce some semblance of const correctness to BN. Shame C doesn't
17261 * "linux-sparc64" configuration (ultrapenguin).
17263 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17264 "linux-sparc" configuration.
17266 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17268 * config now generates no-xxx options for missing ciphers.
17272 * Support the EBCDIC character set (work in progress).
17273 File ebcdic.c not yet included because it has a different license.
17275 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17277 * Support BS2000/OSD-POSIX.
17279 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17281 * Make callbacks for key generation use `void *` instead of `char *`.
17285 * Make S/MIME samples compile (not yet tested).
17289 * Additional typesafe stacks.
17293 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17297 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17299 * New configuration variant "sco5-gcc".
17301 * Updated some demos.
17303 *Sean O Riordain, Wade Scholine*
17305 * Add missing BIO_free at exit of pkcs12 application.
17309 * Fix memory leak in conf.c.
17313 * Updates for Win32 to assembler version of MD5.
17317 * Set #! path to perl in `apps/der_chop` to where we found it
17318 instead of using a fixed path.
17322 * SHA library changes for irix64-mips4-cc.
17326 * Improvements for VMS support.
17330 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17332 * Bignum library bug fix. IRIX 6 passes "make test" now!
17333 This also avoids the problems with SC4.2 and unpatched SC5.
17335 *Andy Polyakov <appro@fy.chalmers.se>*
17337 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17338 These are required because of the typesafe stack would otherwise break
17339 existing code. If old code used a structure member which used to be STACK
17340 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17341 sk_num or sk_value it would produce an error because the num, data members
17342 are not present in STACK_OF. Now it just produces a warning. sk_set
17343 replaces the old method of assigning a value to sk_value
17344 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17345 that does this will no longer work (and should use sk_set instead) but
17346 this could be regarded as a "questionable" behaviour anyway.
17350 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17351 correctly handle encrypted S/MIME data.
17355 * Change type of various DES function arguments from des_cblock
17356 (which means, in function argument declarations, pointer to char)
17357 to des_cblock * (meaning pointer to array with 8 char elements),
17358 which allows the compiler to do more typechecking; it was like
17359 that back in SSLeay, but with lots of ugly casts.
17361 Introduce new type const_des_cblock.
17365 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17366 problems: find RecipientInfo structure that matches recipient certificate
17367 and initialise the ASN1 structures properly based on passed cipher.
17371 * Belatedly make the BN tests actually check the results.
17375 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17376 to and from BNs: it was completely broken. New compilation option
17377 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17378 key elements as negative integers.
17382 * Reorganize and speed up MD5.
17384 *Andy Polyakov <appro@fy.chalmers.se>*
17388 *Richard Levitte <richard@levitte.org>*
17390 * New option -out to asn1parse to allow the parsed structure to be
17391 output to a file. This is most useful when combined with the -strparse
17392 option to examine the output of things like OCTET STRINGS.
17396 * Make SSL library a little more fool-proof by not requiring any longer
17397 that `SSL_set_{accept,connect}_state` be called before
17398 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17399 in many applications because usually everything *appeared* to work as
17400 intended anyway -- now it really works as intended).
17404 * Move openssl.cnf out of lib/.
17408 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17409 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17410 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17412 *Ralf S. Engelschall*
17414 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17415 handle PKCS#7 enveloped data properly.
17417 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17419 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17420 copying pointers. The cert_st handling is changed by this in
17421 various ways (and thus what used to be known as ctx->default_cert
17422 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17423 any longer when s->cert does not give us what we need).
17424 ssl_cert_instantiate becomes obsolete by this change.
17425 As soon as we've got the new code right (possibly it already is?),
17426 we have solved a couple of bugs of the earlier code where s->cert
17427 was used as if it could not have been shared with other SSL structures.
17429 Note that using the SSL API in certain dirty ways now will result
17430 in different behaviour than observed with earlier library versions:
17431 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17432 does not influence s as it used to.
17434 In order to clean up things more thoroughly, inside SSL_SESSION
17435 we don't use CERT any longer, but a new structure SESS_CERT
17436 that holds per-session data (if available); currently, this is
17437 the peer's certificate chain and, for clients, the server's certificate
17438 and temporary key. CERT holds only those values that can have
17439 meaningful defaults in an SSL_CTX.
17443 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17444 from the internal representation. Various PKCS#7 fixes: remove some
17445 evil casts and set the enc_dig_alg field properly based on the signing
17450 * Allow PKCS#12 password to be set from the command line or the
17451 environment. Let 'ca' get its config file name from the environment
17452 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17457 * Allow certificate policies extension to use an IA5STRING for the
17458 organization field. This is contrary to the PKIX definition but
17459 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17464 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17465 without disallowing inline assembler and the like for non-pedantic builds.
17469 * Support Borland C++ builder.
17471 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17477 * SHA-1 cleanups and performance enhancements.
17479 *Andy Polyakov <appro@fy.chalmers.se>*
17481 * Sparc v8plus assembler for the bignum library.
17483 *Andy Polyakov <appro@fy.chalmers.se>*
17485 * Accept any -xxx and +xxx compiler options in Configure.
17489 * Update HPUX configuration.
17493 * Add missing `sk_<type>_unshift()` function to safestack.h
17495 *Ralf S. Engelschall*
17497 * New function SSL_CTX_use_certificate_chain_file that sets the
17498 "extra_cert"s in addition to the certificate. (This makes sense
17499 only for "PEM" format files, as chains as a whole are not
17504 * Support verify_depth from the SSL API.
17505 x509_vfy.c had what can be considered an off-by-one-error:
17506 Its depth (which was not part of the external interface)
17507 was actually counting the number of certificates in a chain;
17508 now it really counts the depth.
17512 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17513 instead of X509err, which often resulted in confusing error
17514 messages since the error codes are not globally unique
17515 (e.g. an alleged error in ssl3_accept when a certificate
17516 didn't match the private key).
17518 * New function SSL_CTX_set_session_id_context that allows to set a default
17519 value (so that you don't need SSL_set_session_id_context for each
17520 connection using the SSL_CTX).
17524 * OAEP decoding bug fix.
17528 * Support INSTALL_PREFIX for package builders, as proposed by
17533 * New Configure options "threads" and "no-threads". For systems
17534 where the proper compiler options are known (currently Solaris
17535 and Linux), "threads" is the default.
17539 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17543 * Install various scripts to $(OPENSSLDIR)/misc, not to
17544 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17545 such as /usr/local/bin.
17549 * "make linux-shared" to build shared libraries.
17551 *Niels Poppe <niels@netbox.org>*
17553 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17557 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17558 extension adding in x509 utility.
17562 * Remove NOPROTO sections and error code comments.
17566 * Partial rewrite of the DEF file generator to now parse the ANSI
17571 * New Configure options --prefix=DIR and --openssldir=DIR.
17575 * Complete rewrite of the error code script(s). It is all now handled
17576 by one script at the top level which handles error code gathering,
17577 header rewriting and C source file generation. It should be much better
17578 than the old method: it now uses a modified version of Ulf's parser to
17579 read the ANSI prototypes in all header files (thus the old K&R definitions
17580 aren't needed for error creation any more) and do a better job of
17581 translating function codes into names. The old 'ASN1 error code embedded
17582 in a comment' is no longer necessary and it doesn't use .err files which
17583 have now been deleted. Also the error code call doesn't have to appear all
17584 on one line (which resulted in some large lines...).
17588 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17592 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17593 0 (which usually indicates a closed connection), but continue reading.
17597 * Fix some race conditions.
17601 * Add support for CRL distribution points extension. Add Certificate
17602 Policies and CRL distribution points documentation.
17606 * Move the autogenerated header file parts to crypto/opensslconf.h.
17610 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17611 8 of keying material. Merlin has also confirmed interop with this fix
17612 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17614 *Merlin Hughes <merlin@baltimore.ie>*
17616 * Fix lots of warnings.
17618 *Richard Levitte <levitte@stacken.kth.se>*
17620 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17621 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17623 *Richard Levitte <levitte@stacken.kth.se>*
17625 * Fix problems with sizeof(long) == 8.
17627 *Andy Polyakov <appro@fy.chalmers.se>*
17629 * Change functions to ANSI C.
17633 * Fix typos in error codes.
17635 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17637 * Remove defunct assembler files from Configure.
17641 * SPARC v8 assembler BIGNUM implementation.
17643 *Andy Polyakov <appro@fy.chalmers.se>*
17645 * Support for Certificate Policies extension: both print and set.
17646 Various additions to support the r2i method this uses.
17650 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17651 return a const string when you are expecting an allocated buffer.
17655 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17656 types DirectoryString and DisplayText.
17660 * Add code to allow r2i extensions to access the configuration database,
17661 add an LHASH database driver and add several ctx helper functions.
17665 * Fix an evil bug in bn_expand2() which caused various BN functions to
17666 fail when they extended the size of a BIGNUM.
17670 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17671 support typesafe stack.
17675 * Fix typo in SSL_[gs]et_options().
17677 *Nils Frostberg <nils@medcom.se>*
17679 * Delete various functions and files that belonged to the (now obsolete)
17680 old X509V3 handling code.
17684 * New Configure option "rsaref".
17688 * Don't auto-generate pem.h.
17692 * Introduce type-safe ASN.1 SETs.
17696 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17698 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17700 * Introduce type-safe STACKs. This will almost certainly break lots of code
17701 that links with OpenSSL (well at least cause lots of warnings), but fear
17702 not: the conversion is trivial, and it eliminates loads of evil casts. A
17703 few STACKed things have been converted already. Feel free to convert more.
17704 In the fullness of time, I'll do away with the STACK type altogether.
17708 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17709 specified in `<certfile>` by updating the entry in the index.txt file.
17710 This way one no longer has to edit the index.txt file manually for
17711 revoking a certificate. The -revoke option does the gory details now.
17713 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17715 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17716 `-text` option at all and this way the `-noout -text` combination was
17717 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17719 *Ralf S. Engelschall*
17721 * Make sure a corresponding plain text error message exists for the
17722 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17723 verify callback function determined that a certificate was revoked.
17725 *Ralf S. Engelschall*
17727 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17728 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17729 all available ciphers including rc5, which was forgotten until now.
17730 In order to let the testing shell script know which algorithms
17731 are available, a new (up to now undocumented) command
17732 `openssl list-cipher-commands` is used.
17736 * Bugfix: s_client occasionally would sleep in select() when
17737 it should have checked SSL_pending() first.
17741 * New functions DSA_do_sign and DSA_do_verify to provide access to
17742 the raw DSA values prior to ASN.1 encoding.
17746 * Tweaks to Configure
17748 *Niels Poppe <niels@netbox.org>*
17750 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17755 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17759 * New config option to avoid instructions that are illegal on the 80386.
17760 The default code is faster, but requires at least a 486.
17764 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17765 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17766 same as SSL2_VERSION anyway.
17770 * New "-showcerts" option for s_client.
17774 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17775 application. Various cleanups and fixes.
17779 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17780 modify error routines to work internally. Add error codes and PBE init
17781 to library startup routines.
17785 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17786 packing functions to asn1 and evp. Changed function names and error
17787 codes along the way.
17791 * PKCS12 integration: and so it begins... First of several patches to
17792 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17793 objects to objects.h
17797 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17798 and display support for Thawte strong extranet extension.
17802 * Add LinuxPPC support.
17804 *Jeff Dubrule <igor@pobox.org>*
17806 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17807 bn_div_words in alpha.s.
17809 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17811 * Make sure the RSA OAEP test is skipped under -DRSAref because
17812 OAEP isn't supported when OpenSSL is built with RSAref.
17814 *Ulf Moeller <ulf@fitug.de>*
17816 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17817 so they no longer are missing under -DNOPROTO.
17819 *Soren S. Jorvang <soren@t.dk>*
17821 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17823 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17824 doesn't work when the session is reused. Coming soon!
17828 * Fix a security hole, that allows sessions to be reused in the wrong
17829 context thus bypassing client cert protection! All software that uses
17830 client certs and session caches in multiple contexts NEEDS PATCHING to
17831 allow session reuse! A fuller solution is in the works.
17833 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17835 * Some more source tree cleanups (removed obsolete files
17836 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17837 permission on "config" script to be executable) and a fix for the INSTALL
17840 *Ulf Moeller <ulf@fitug.de>*
17842 * Remove some legacy and erroneous uses of malloc, free instead of
17845 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17847 * Make rsa_oaep_test return non-zero on error.
17849 *Ulf Moeller <ulf@fitug.de>*
17851 * Add support for native Solaris shared libraries. Configure
17852 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17853 if someone would make that last step automatic.
17855 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17857 * ctx_size was not built with the right compiler during "make links". Fixed.
17861 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17862 except NULL ciphers". This means the default cipher list will no longer
17863 enable NULL ciphers. They need to be specifically enabled e.g. with
17864 the string "DEFAULT:eNULL".
17868 * Fix to RSA private encryption routines: if p < q then it would
17869 occasionally produce an invalid result. This will only happen with
17870 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17874 * Be less restrictive and allow also `perl util/perlpath.pl
17875 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17876 because this way one can also use an interpreter named `perl5` (which is
17877 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17878 installed as `perl`).
17880 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17882 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17884 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17886 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17887 advapi32.lib to Win32 build and change the pem test comparison
17888 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17889 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17890 and crypto/des/ede_cbcm_enc.c.
17894 * DES quad checksum was broken on big-endian architectures. Fixed.
17898 * Comment out two functions in bio.h that aren't implemented. Fix up the
17899 Win32 test batch file so it (might) work again. The Win32 test batch file
17900 is horrible: I feel ill....
17904 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17905 in e_os.h. Audit of header files to check ANSI and non ANSI
17906 sections: 10 functions were absent from non ANSI section and not exported
17907 from Windows DLLs. Fixed up libeay.num for new functions.
17911 * Make `openssl version` output lines consistent.
17913 *Ralf S. Engelschall*
17915 * Fix Win32 symbol export lists for BIO functions: Added
17916 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17917 to ms/libeay{16,32}.def.
17919 *Ralf S. Engelschall*
17921 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17922 fine under Unix and passes some trivial tests I've now added. But the
17923 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17924 added to make sure no one expects that this stuff really works in the
17925 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17926 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17929 *Ralf S. Engelschall*
17931 * Fix the generation of two part addresses in perl.
17933 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17935 * Add config entry for Linux on MIPS.
17937 *John Tobey <jtobey@channel1.com>*
17939 * Make links whenever Configure is run, unless we are on Windoze.
17943 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17944 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17949 * Add a useful kludge to allow package maintainers to specify compiler and
17950 other platforms details on the command line without having to patch the
17951 Configure script every time: One now can use
17952 `perl Configure <id>:<details>`,
17953 i.e. platform ids are allowed to have details appended
17954 to them (separated by colons). This is treated as there would be a static
17955 pre-configured entry in Configure's %table under key `<id>` with value
17956 `<details>` and `perl Configure <id>` is called. So, when you want to
17957 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17958 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17959 now, which overrides the FreeBSD-elf entry on-the-fly.
17961 *Ralf S. Engelschall*
17963 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17967 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17968 on the `perl Configure ...` command line. This way one can compile
17969 OpenSSL libraries with Position Independent Code (PIC) which is needed
17970 for linking it into DSOs.
17972 *Ralf S. Engelschall*
17974 * Remarkably, export ciphers were totally broken and no-one had noticed!
17979 * Cleaned up the LICENSE document: The official contact for any license
17980 questions now is the OpenSSL core team under openssl-core@openssl.org.
17981 And add a paragraph about the dual-license situation to make sure people
17982 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17983 to the OpenSSL toolkit.
17985 *Ralf S. Engelschall*
17987 * General source tree makefile cleanups: Made `making xxx in yyy...`
17988 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17989 Additionally cleaned up the `make links` target: Remove unnecessary
17990 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17991 to speed processing and no longer clutter the display with confusing
17992 stuff. Instead only the actually done links are displayed.
17994 *Ralf S. Engelschall*
17996 * Permit null encryption ciphersuites, used for authentication only. It used
17997 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17998 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18003 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18004 signed attributes when verifying signatures (this would break them),
18005 the detached data encoding was wrong and public keys obtained using
18006 X509_get_pubkey() weren't freed.
18010 * Add text documentation for the BUFFER functions. Also added a work around
18011 to a Win95 console bug. This was triggered by the password read stuff: the
18012 last character typed gets carried over to the next fread(). If you were
18013 generating a new cert request using 'req' for example then the last
18014 character of the passphrase would be CR which would then enter the first
18019 * Added the new 'Includes OpenSSL Cryptography Software' button as
18020 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18021 button and can be used by applications based on OpenSSL to show the
18022 relationship to the OpenSSL project.
18024 *Ralf S. Engelschall*
18026 * Remove confusing variables in function signatures in files
18027 ssl/ssl_lib.c and ssl/ssl.h.
18029 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18031 * Don't install bss_file.c under PREFIX/include/
18033 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18035 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18036 functions that return function pointers and has support for NT specific
18037 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18038 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18039 unsigned to signed types: this was killing the Win32 compile.
18043 * Add new certificate file to stack functions,
18044 SSL_add_dir_cert_subjects_to_stack() and
18045 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18046 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18047 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18048 This means that Apache-SSL and similar packages don't have to mess around
18049 to add as many CAs as they want to the preferred list.
18053 * Experiment with doxygen documentation. Currently only partially applied to
18055 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18056 openssl.doxy as the configuration file.
18060 * Get rid of remaining C++-style comments which strict C compilers hate.
18062 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18064 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18065 compiled in by default: it has problems with large keys.
18069 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18070 DH private keys and/or callback functions which directly correspond to
18071 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18072 is needed for applications which have to configure certificates on a
18073 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18075 For the RSA certificate situation is makes no difference, but
18076 for the DSA certificate situation this fixes the "no shared cipher"
18077 problem where the OpenSSL cipher selection procedure failed because the
18078 temporary keys were not overtaken from the context and the API provided
18079 no way to reconfigure them.
18080 The new functions now let applications reconfigure the stuff and they
18081 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18082 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18083 non-public-API function ssl_cert_instantiate() is used as a helper
18084 function and also to reduce code redundancy inside ssl_rsa.c.
18086 *Ralf S. Engelschall*
18088 * Move s_server -dcert and -dkey options out of the undocumented feature
18089 area because they are useful for the DSA situation and should be
18090 recognized by the users.
18092 *Ralf S. Engelschall*
18094 * Fix the cipher decision scheme for export ciphers: the export bits are
18095 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18096 SSL_EXP_MASK. So, the original variable has to be used instead of the
18097 already masked variable.
18099 *Richard Levitte <levitte@stacken.kth.se>*
18101 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18103 *Richard Levitte <levitte@stacken.kth.se>*
18105 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18106 from `int` to `unsigned int` because it is a length and initialized by
18107 EVP_DigestFinal() which expects an `unsigned int *`.
18109 *Richard Levitte <levitte@stacken.kth.se>*
18111 * Don't hard-code path to Perl interpreter on shebang line of Configure
18112 script. Instead use the usual Shell->Perl transition trick.
18114 *Ralf S. Engelschall*
18116 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18117 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18118 -noout -modulus` as it's already the case for `openssl rsa -noout
18119 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18120 currently the public key is printed (a decision which was already done by
18121 `openssl dsa -modulus` in the past) which serves a similar purpose.
18122 Additionally the NO_RSA no longer completely removes the whole -modulus
18123 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18126 *Ralf S. Engelschall*
18128 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18129 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18131 *Arne Ansper <arne@ats.cyber.ee>*
18133 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18134 to be added. Now both 'req' and 'ca' can use new objects defined in the
18139 * Add cool BIO that does syslog (or event log on NT).
18141 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18143 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18144 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18145 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18146 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18150 * Add preliminary config info for new extension code.
18154 * Make RSA_NO_PADDING really use no padding.
18156 *Ulf Moeller <ulf@fitug.de>*
18158 * Generate errors when private/public key check is done.
18162 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18163 for some CRL extensions and new objects added.
18167 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18168 key usage extension and fuller support for authority key id.
18172 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18173 padding method for RSA, which is recommended for new applications in PKCS
18174 #1 v2.0 (RFC 2437, October 1998).
18175 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18176 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18177 against Bleichbacher's attack on RSA.
18178 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18181 * Updates to the new SSL compression code
18183 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18185 * Fix so that the version number in the master secret, when passed
18186 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18187 (because the server will not accept higher), that the version number
18188 is 0x03,0x01, not 0x03,0x00
18190 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18192 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18193 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18194 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18198 * Support for RAW extensions where an arbitrary extension can be
18199 created by including its DER encoding. See `apps/openssl.cnf` for
18204 * Make sure latest Perl versions don't interpret some generated C array
18205 code as Perl array code in the crypto/err/err_genc.pl script.
18207 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18209 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18210 not many people have the assembler. Various Win32 compilation fixes and
18211 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18212 build instructions.
18216 * Modify configure script 'Configure' to automatically create crypto/date.h
18217 file under Win32 and also build pem.h from pem.org. New script
18218 util/mkfiles.pl to create the MINFO file on environments that can't do a
18219 'make files': perl util/mkfiles.pl >MINFO should work.
18223 * Major rework of DES function declarations, in the pursuit of correctness
18224 and purity. As a result, many evil casts evaporated, and some weirdness,
18225 too. You may find this causes warnings in your code. Zapping your evil
18226 casts will probably fix them. Mostly.
18230 * Fix for a typo in asn1.h. Bug fix to object creation script
18231 obj_dat.pl. It considered a zero in an object definition to mean
18232 "end of object": none of the objects in objects.h have any zeros
18233 so it wasn't spotted.
18235 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18237 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18238 Masking (CBCM). In the absence of test vectors, the best I have been able
18239 to do is check that the decrypt undoes the encrypt, so far. Send me test
18240 vectors if you have them.
18244 * Correct calculation of key length for export ciphers (too much space was
18245 allocated for null ciphers). This has not been tested!
18249 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18250 message is now correct (it understands "crypto" and "ssl" on its
18251 command line). There is also now an "update" option. This will update
18252 the util/ssleay.num and util/libeay.num files with any new functions.
18254 perl util/mkdef.pl crypto ssl update
18255 it will update them.
18259 * Overhauled the Perl interface:
18260 - ported BN stuff to OpenSSL's different BN library
18261 - made the perl/ source tree CVS-aware
18262 - renamed the package from SSLeay to OpenSSL (the files still contain
18263 their history because I've copied them in the repository)
18264 - removed obsolete files (the test scripts will be replaced
18265 by better Test::Harness variants in the future)
18267 *Ralf S. Engelschall*
18269 * First cut for a very conservative source tree cleanup:
18270 1. merge various obsolete readme texts into doc/ssleay.txt
18271 where we collect the old documents and readme texts.
18272 2. remove the first part of files where I'm already sure that we no
18273 longer need them because of three reasons: either they are just temporary
18274 files which were left by Eric or they are preserved original files where
18275 I've verified that the diff is also available in the CVS via "cvs diff
18276 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18277 the crypto/md/ stuff).
18279 *Ralf S. Engelschall*
18281 * More extension code. Incomplete support for subject and issuer alt
18282 name, issuer and authority key id. Change the i2v function parameters
18283 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18284 what that's for :-) Fix to ASN1 macro which messed up
18285 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18289 * Preliminary support for ENUMERATED type. This is largely copied from the
18294 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18296 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18298 * Make sure `make rehash` target really finds the `openssl` program.
18300 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18302 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18303 like to hear about it if this slows down other processors.
18307 * Add CygWin32 platform information to Configure script.
18309 *Alan Batie <batie@aahz.jf.intel.com>*
18311 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18313 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18315 * New program nseq to manipulate netscape certificate sequences
18319 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18324 * Fixes to BN code. Previously the default was to define BN_RECURSION
18325 but the BN code had some problems that would cause failures when
18326 doing certificate verification and some other functions.
18328 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18330 * Add ASN1 and PEM code to support netscape certificate sequences.
18334 * Add ASN1 and PEM code to support netscape certificate sequences.
18338 * Add several PKIX and private extended key usage OIDs.
18342 * Modify the 'ca' program to handle the new extension code. Modify
18343 openssl.cnf for new extension format, add comments.
18347 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18348 and add a sample to openssl.cnf so req -x509 now adds appropriate
18353 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18354 error code, add initial support to X509_print() and x509 application.
18358 * Takes a deep breath and start adding X509 V3 extension support code. Add
18359 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18360 stuff is currently isolated and isn't even compiled yet.
18364 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18365 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18366 Removed the versions check from X509 routines when loading extensions:
18367 this allows certain broken certificates that don't set the version
18368 properly to be processed.
18372 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18373 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18374 can still be regenerated with "make depend".
18378 * Spelling mistake in C version of CAST-128.
18380 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18382 * Changes to the error generation code. The perl script err-code.pl
18383 now reads in the old error codes and retains the old numbers, only
18384 adding new ones if necessary. It also only changes the .err files if new
18385 codes are added. The makefiles have been modified to only insert errors
18386 when needed (to avoid needlessly modifying header files). This is done
18387 by only inserting errors if the .err file is newer than the auto generated
18388 C file. To rebuild all the error codes from scratch (the old behaviour)
18389 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18390 or delete all the .err files.
18394 * CAST-128 was incorrectly implemented for short keys. The C version has
18395 been fixed, but is untested. The assembler versions are also fixed, but
18396 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18397 to regenerate it if needed.
18398 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18399 Hagino <itojun@kame.net>*
18401 * File was opened incorrectly in randfile.c.
18403 *Ulf Möller <ulf@fitug.de>*
18405 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18406 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18407 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18408 al: it's just almost always a UTCTime. Note this patch adds new error
18409 codes so do a "make errors" if there are problems.
18413 * Correct Linux 1 recognition in config.
18415 *Ulf Möller <ulf@fitug.de>*
18417 * Remove pointless MD5 hash when using DSA keys in ca.
18419 *Anonymous <nobody@replay.com>*
18421 * Generate an error if given an empty string as a cert directory. Also
18422 generate an error if handed NULL (previously returned 0 to indicate an
18423 error, but didn't set one).
18425 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18427 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18431 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18432 parameters. This was causing a warning which killed off the Win32 compile.
18436 * Remove C++ style comments from crypto/bn/bn_local.h.
18438 *Neil Costigan <neil.costigan@celocom.com>*
18440 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18441 based on a text string, looking up short and long names and finally
18442 "dot" format. The "dot" format stuff didn't work. Added new function
18443 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18444 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18445 OID is not part of the table.
18449 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18450 X509_LOOKUP_by_alias().
18454 * Sort openssl functions by name.
18458 * Get the `gendsa` command working and add it to the `list` command. Remove
18459 encryption from sample DSA keys (in case anyone is interested the password
18464 * Make *all* `*_free` functions accept a NULL pointer.
18466 *Frans Heymans <fheymans@isaserver.be>*
18468 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18471 *Anonymous <nobody@replay.com>*
18473 * s_server should send the CAfile as acceptable CAs, not its own cert.
18475 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18477 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18479 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18481 * Temp key "for export" tests were wrong in s3_srvr.c.
18483 *Anonymous <nobody@replay.com>*
18485 * Add prototype for temp key callback functions
18486 SSL_CTX_set_tmp_{rsa,dh}_callback().
18490 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18491 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18495 * X509_name_add_entry() freed the wrong thing after an error.
18497 *Arne Ansper <arne@ats.cyber.ee>*
18499 * rsa_eay.c would attempt to free a NULL context.
18501 *Arne Ansper <arne@ats.cyber.ee>*
18503 * BIO_s_socket() had a broken should_retry() on Windoze.
18505 *Arne Ansper <arne@ats.cyber.ee>*
18507 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18509 *Arne Ansper <arne@ats.cyber.ee>*
18511 * Make sure the already existing X509_STORE->depth variable is initialized
18512 in X509_STORE_new(), but document the fact that this variable is still
18513 unused in the certificate verification process.
18515 *Ralf S. Engelschall*
18517 * Fix the various library and `apps/` files to free up pkeys obtained from
18518 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18522 * Fix reference counting in X509_PUBKEY_get(). This makes
18523 demos/maurice/example2.c work, amongst others, probably.
18525 *Steve Henson and Ben Laurie*
18527 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18528 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18529 are no longer created. This way we have a single and consistent command
18530 line interface `openssl <command>`, similar to `cvs <command>`.
18532 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18534 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18535 BIT STRING wrapper always have zero unused bits.
18539 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18543 * Make the top-level INSTALL documentation easier to understand.
18547 * Makefiles updated to exit if an error occurs in a sub-directory
18548 make (including if user presses ^C) [Paul Sutton]
18550 * Make Montgomery context stuff explicit in RSA data structure.
18554 * Fix build order of pem and err to allow for generated pem.h.
18558 * Fix renumbering bug in X509_NAME_delete_entry().
18562 * Enhanced the err-ins.pl script so it makes the error library number
18563 global and can add a library name. This is needed for external ASN1 and
18564 other error libraries.
18568 * Fixed sk_insert which never worked properly.
18572 * Fix ASN1 macros so they can handle indefinite length constructed
18573 EXPLICIT tags. Some non standard certificates use these: they can now
18578 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18579 into a single doc/ssleay.txt bundle. This way the information is still
18580 preserved but no longer messes up this directory. Now it's new room for
18581 the new set of documentation files.
18583 *Ralf S. Engelschall*
18585 * SETs were incorrectly DER encoded. This was a major pain, because they
18586 shared code with SEQUENCEs, which aren't coded the same. This means that
18587 almost everything to do with SETs or SEQUENCEs has either changed name or
18588 number of arguments.
18590 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18592 * Fix test data to work with the above.
18596 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18597 was already fixed by Eric for 0.9.1 it seems.
18599 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18601 * Autodetect FreeBSD3.
18605 * Fix various bugs in Configure. This affects the following platforms:
18609 unixware-2.0-pentium
18614 * Eliminate generated files from CVS. Reorder tests to regenerate files
18615 before they are needed.
18619 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18623 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18625 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18626 changed SSLeay to OpenSSL in version strings.
18628 *Ralf S. Engelschall*
18630 * Some fixups to the top-level documents.
18634 * Fixed the nasty bug where rsaref.h was not found under compile-time
18635 because the symlink to include/ was missing.
18637 *Ralf S. Engelschall*
18639 * Incorporated the popular no-RSA/DSA-only patches
18640 which allow to compile a RSA-free SSLeay.
18642 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18644 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18645 when "ssleay" is still not found.
18647 *Ralf S. Engelschall*
18649 * Added more platforms to Configure: Cray T3E, HPUX 11,
18651 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18653 * Updated the README file.
18655 *Ralf S. Engelschall*
18657 * Added various .cvsignore files in the CVS repository subdirs
18658 to make a "cvs update" really silent.
18660 *Ralf S. Engelschall*
18662 * Recompiled the error-definition header files and added
18663 missing symbols to the Win32 linker tables.
18665 *Ralf S. Engelschall*
18667 * Cleaned up the top-level documents;
18668 o new files: CHANGES and LICENSE
18669 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18670 o merged COPYRIGHT into LICENSE
18671 o removed obsolete TODO file
18672 o renamed MICROSOFT to INSTALL.W32
18674 *Ralf S. Engelschall*
18676 * Removed dummy files from the 0.9.1b source tree:
18678 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18679 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18680 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18681 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18682 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18685 *Ralf S. Engelschall*
18687 * Added various platform portability fixes.
18691 * The Genesis of the OpenSSL rpject:
18692 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18693 Young and Tim J. Hudson created while they were working for C2Net until
18696 *The OpenSSL Project*
18698 ### Changes between 0.9.0b and 0.9.1b [not released]
18700 * Updated a few CA certificates under certs/
18704 * Changed some BIGNUM api stuff.
18708 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18709 DGUX x86, Linux Alpha, etc.
18713 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18714 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18719 * Add -strparse option to asn1pars program which parses nested
18722 *Dr Stephen Henson <shenson@bigfoot.com>*
18724 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18728 * DSA fix for "ca" program.
18732 * Added "-genkey" option to "dsaparam" program.
18736 * Added RIPE MD160 (rmd160) message digest.
18740 * Added -a (all) option to "ssleay version" command.
18744 * Added PLATFORM define which is the id given to Configure.
18748 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18752 * Extended the ASN.1 parser routines.
18756 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18760 * Added a BN_CTX to the BN library.
18764 * Fixed the weak key values in DES library
18768 * Changed API in EVP library for cipher aliases.
18772 * Added support for RC2/64bit cipher.
18776 * Converted the lhash library to the crypto/mem.c functions.
18780 * Added more recognized ASN.1 object ids.
18784 * Added more RSA padding checks for SSL/TLS.
18788 * Added BIO proxy/filter functionality.
18792 * Added extra_certs to SSL_CTX which can be used
18793 send extra CA certificates to the client in the CA cert chain sending
18794 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18798 * Now Fortezza is denied in the authentication phase because
18799 this is key exchange mechanism is not supported by SSLeay at all.
18803 * Additional PKCS1 checks.
18807 * Support the string "TLSv1" for all TLS v1 ciphers.
18811 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18812 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18816 * Fixed a few memory leaks.
18820 * Fixed various code and comment typos.
18824 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18825 bytes sent in the client random.
18827 *Edward Bishop <ebishop@spyglass.com>*
18831 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18832 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18833 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18834 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18835 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18836 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18837 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18838 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18839 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18840 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18841 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18842 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18843 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18844 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18845 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18846 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18847 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18848 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18849 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18850 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18851 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18852 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18853 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18854 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18855 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18856 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18857 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18858 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18859 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18860 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18861 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18862 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18863 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18864 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18865 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18866 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18867 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18868 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18869 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18870 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18871 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18872 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18873 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18874 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18875 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18876 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18877 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18878 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18879 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18880 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18881 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18882 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18883 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18884 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18885 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18886 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18887 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18888 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18889 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18890 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18891 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18892 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18893 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18894 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18895 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18896 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18897 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18898 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18899 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18900 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18901 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18902 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18903 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18904 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18905 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18906 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18907 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18908 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18909 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18910 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18911 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18912 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18913 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18914 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18915 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18916 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18917 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18918 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18919 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18920 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18921 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18922 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18923 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18924 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18925 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18926 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18927 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18928 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18929 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18930 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18931 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18932 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18933 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18934 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18935 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18936 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18937 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18938 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18939 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18940 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18941 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18942 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18943 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18944 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18945 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18946 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18947 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18948 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18949 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18950 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18951 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18952 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18953 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18954 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18955 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18956 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18957 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18958 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18959 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18960 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18961 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18962 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18963 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18964 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18965 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18966 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18967 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18968 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18969 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18970 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18971 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18972 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18973 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18974 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18975 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18976 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18977 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18978 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18979 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18980 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18981 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18982 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18983 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18984 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18985 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18986 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18987 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18988 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18989 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18990 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18991 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18992 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655