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 * A number of functions handling low level keys or engines were deprecated
27 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
28 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
29 EVP_PKEY_get0_siphash(). Applications using engines should instead use
30 providers. Applications getting or setting low level keys in an EVP_PKEY
31 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
32 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
36 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
37 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
38 and require returning octet ptr parameters from providers that
39 would like to support them which complicates provider implementations.
43 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
44 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
45 RAND_set_rand_method(). Provider based random number generators should
46 be used instead via EVP_RAND(3).
50 * The SRP APIs have been deprecated. The old APIs do not work via providers,
51 and there is no EVP interface to them. Unfortunately there is no replacement
52 for these APIs at this time.
56 * Add a compile time option to prevent the caching of provider fetched
57 algorithms. This is enabled by including the no-cached-fetch option
58 at configuration time.
62 * The openssl speed command does not use low-level API calls anymore. This
63 implies some of the performance numbers might not be fully comparable
64 with the previous releases due to higher overhead. This applies
65 particularly to measuring performance on smaller data chunks.
69 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
70 Typically if OpenSSL has no EC or DH algorithms then it cannot support
71 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
72 through providers. Therefore third party providers may supply group
73 implementations even where there are no built-in ones. Attempting to create
74 TLS connections in such a build without also disabling TLSv1.3 at run time or
75 using third party provider groups may result in handshake failures. TLSv1.3
76 can be disabled at compile time using the "no-tls1_3" Configure option.
80 * The undocumented function X509_certificate_type() has been deprecated;
81 applications can use X509_get0_pubkey() and X509_get0_signature() to
82 get the same information.
86 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
87 functions. They are identical to BN_rand() and BN_rand_range()
92 * Removed RSA padding mode for SSLv23 (which was only used for
93 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
94 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
99 * Deprecated the obsolete X9.31 RSA key generation related functions
100 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
101 BN_X931_generate_prime_ex().
105 * The default key generation method for the regular 2-prime RSA keys was
106 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
107 Conditions Based on Auxiliary Probable Primes). This method is slower
108 than the original method.
112 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
113 They are replaced with the BN_check_prime() function that avoids possible
114 misuse and always uses at least 64 rounds of the Miller-Rabin
115 primality test. At least 64 rounds of the Miller-Rabin test are now also
116 used for all prime generation, including RSA key generation.
117 This increases key generation time, especially for larger keys.
121 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
122 as they are not useful with non-deprecated functions.
126 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
127 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
128 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
129 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
130 were used to collect all necessary data to form a HTTP request, and to
131 perform the HTTP transfer with that request. With OpenSSL 3.0, the
132 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
133 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
134 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
135 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
136 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
137 OSSL_HTTP_REQ_CTX_set_max_response_length().
139 *Rich Salz and Richard Levitte*
141 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
145 * Validation of SM2 keys has been separated from the validation of regular EC
146 keys, allowing to improve the SM2 validation process to reject loaded private
147 keys that are not conforming to the SM2 ISO standard.
148 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
153 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
154 switches: a validation failure triggers an early exit, returning a failure
155 exit status to the parent process.
159 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
160 to ignore unknown ciphers.
164 * The `-cipher-commands` and `-digest-commands` options
165 of the command line utility `list` have been deprecated.
166 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
170 * All of the low level EC_KEY functions have been deprecated including:
172 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
173 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
174 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
175 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
176 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
177 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
178 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
179 EC_KEY_METHOD_get_verify,
180 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
181 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
182 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
183 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
184 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
185 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
186 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
187 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
188 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
189 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
190 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
191 Applications that need to implement an EC_KEY_METHOD need to consider
192 implementation of the functionality in a special provider.
193 For replacement of the functions manipulating the EC_KEY objects
194 see the EVP_PKEY-EC(7) manual page.
196 Additionally functions that read and write EC_KEY objects such as
197 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
198 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
199 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
200 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
201 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
202 have also been deprecated. Applications should instead use the
203 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
205 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
206 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
207 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
208 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
209 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
211 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
213 * Deprecated all the libcrypto and libssl error string loading
214 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
215 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
216 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
217 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
218 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
219 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
220 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
221 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
222 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
223 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
225 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
226 now loads error strings automatically.
230 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
231 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
232 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
233 are to be used by servers requiring ephemeral DH keys. Instead applications
234 should consider using the built-in DH parameters that are available by
235 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
236 necessary then applications can use the alternative functions
237 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
238 replacement for the "callback" functions. The callback was originally useful
239 in order to have different parameters for export and non-export ciphersuites.
240 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
241 functions should be replaced by one of the other methods described above.
245 * The `-crypt` option to the `passwd` command line tool has been removed.
249 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
254 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
256 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
257 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
258 The inverse ciphers use AES decryption for wrapping, and
259 AES encryption for unwrapping.
263 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
264 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
265 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
266 OpenSSL 3.0 these are replaced by the more generic functions
267 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
268 The old versions have been converted to deprecated macros that just call the
273 * The security callback, which can be customised by application code, supports
274 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
275 in the "other" parameter. In most places this is what is passed. All these
276 places occur server side. However there was one client side call of this
277 security operation and it passed a DH object instead. This is incorrect
278 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
279 of the other locations. Therefore this client side call has been changed to
280 pass an EVP_PKEY instead.
284 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
285 interface. Their functionality remains unchanged.
289 * Added new option for 'openssl list', '-providers', which will display the
290 list of loaded providers, their names, version and status. It optionally
291 displays their gettable parameters.
295 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
296 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
297 type is internally recognised so the workaround is no longer needed.
299 Functionality is still retained as it is, but will only work with
300 EVP_PKEYs with a legacy internal key.
304 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
305 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
309 * Changed all "STACK" functions to be macros instead of inline functions. Macro
310 parameters are still checked for type safety at compile time via helper
315 * Remove the RAND_DRBG API
317 The RAND_DRBG API did not fit well into the new provider concept as
318 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
319 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
320 and some of its API calls are rather low-level. This holds in particular
321 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
323 Adding a compatibility layer to continue supporting the RAND_DRBG API as
324 a legacy API for a regular deprecation period turned out to come at the
325 price of complicating the new provider API unnecessarily. Since the
326 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
329 *Paul Dale and Matthias St. Pierre*
331 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
332 as well as actual hostnames.
336 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
337 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
338 conversely, silently ignore DTLS protocol version bounds when configuring
339 TLS-based contexts. The commands can be repeated to set bounds of both
340 types. The same applies with the corresponding "min_protocol" and
341 "max_protocol" command-line switches, in case some application uses both TLS
344 SSL_CTX instances that are created for a fixed protocol version (e.g.
345 `TLSv1_server_method()`) also silently ignore version bounds. Previously
346 attempts to apply bounds to these protocol versions would result in an
347 error. Now only the "version-flexible" SSL_CTX instances are subject to
348 limits in configuration files in command-line options.
352 * Deprecated the `ENGINE` API. Engines should be replaced with providers
357 * Reworked the recorded ERR codes to make better space for system errors.
358 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
359 given code is a system error (true) or an OpenSSL error (false).
363 * Reworked the test perl framework to better allow parallel testing.
365 *Nicola Tuveri and David von Oheimb*
367 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
368 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
372 * 'Configure' has been changed to figure out the configuration target if
373 none is given on the command line. Consequently, the 'config' script is
374 now only a mere wrapper. All documentation is changed to only mention
377 *Rich Salz and Richard Levitte*
379 * Added a library context `OSSL_LIB_CTX` that applications as well as
380 other libraries can use to form a separate context within which
381 libcrypto operations are performed.
383 There are two ways this can be used:
385 - Directly, by passing a library context to functions that take
386 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
388 - Indirectly, by creating a new library context and then assigning
389 it as the new default, with `OSSL_LIB_CTX_set0_default`.
391 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
392 apart from the functions directly related to `OSSL_LIB_CTX`, accept
393 NULL to indicate that the default library context should be used.
395 Library code that changes the default library context using
396 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
397 second call before returning to the caller.
399 _(Note: the library context was initially called `OPENSSL_CTX` and
400 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
404 * Handshake now fails if Extended Master Secret extension is dropped
409 * Dropped interactive mode from the `openssl` program. From now on,
410 running it without arguments is equivalent to `openssl help`.
414 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
415 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
416 While the old function names have been retained for backward compatibility
417 they should not be used in new developments
418 because their return values are confusing: Unlike other `_cmp()` functions
419 they do not return 0 in case their arguments are equal.
423 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
424 `EC_GROUP_get_field_type()`.
428 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
429 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
430 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
431 Applications should rely on the library automatically assigning a suitable
432 EC_METHOD internally upon EC_GROUP construction.
436 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
437 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
438 assigned internally without application intervention.
439 Users of EC_GROUP_new() should switch to a different suitable constructor.
443 * Add CAdES-BES signature verification support, mostly derived
444 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
446 *Filipe Raimundo da Silva*
448 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
452 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
453 functions are not widely used and now OpenSSL automatically perform this
454 conversion when needed.
458 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
459 EC_KEY_precompute_mult(). These functions are not widely used and
460 applications should instead switch to named curves which OpenSSL has
461 hardcoded lookup tables for.
465 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
466 should instead use the L<EC_POINT_mul(3)> function.
470 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
471 that are not applicable to the new provider model. Applications should
472 instead use EVP_default_properties_is_fips_enabled() and
473 EVP_default_properties_enable_fips().
477 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
478 is set, an unexpected EOF is ignored, it pretends a close notify was received
479 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
483 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
484 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
485 used and applications should instead use the
486 L<EC_POINT_set_affine_coordinates(3)> and
487 L<EC_POINT_get_affine_coordinates(3)> functions.
491 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
492 arrays to be more easily constructed via a series of utility functions.
493 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
494 the various push functions and finally convert to a passable OSSL_PARAM
495 array using OSSL_PARAM_BLD_to_param().
499 * The security strength of SHA1 and MD5 based signatures in TLS has been
500 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
501 working at the default security level of 1 and instead requires security
502 level 0. The security level can be changed either using the cipher string
503 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
507 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
508 contain a provider side internal key.
512 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
513 They are old functions that we don't use, and that you could disable with
514 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
518 * Project text documents not yet having a proper file name extension
519 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
520 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
524 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
525 have been converted to Markdown with the goal to produce documents
526 which not only look pretty when viewed online in the browser, but
527 remain well readable inside a plain text editor.
529 To achieve this goal, a 'minimalistic' Markdown style has been applied
530 which avoids formatting elements that interfere too much with the
531 reading flow in the text file. For example, it
533 * avoids [ATX headings][] and uses [setext headings][] instead
534 (which works for `<h1>` and `<h2>` headings only).
535 * avoids [inline links][] and uses [reference links][] instead.
536 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
538 [ATX headings]: https://github.github.com/gfm/#atx-headings
539 [setext headings]: https://github.github.com/gfm/#setext-headings
540 [inline links]: https://github.github.com/gfm/#inline-link
541 [reference links]: https://github.github.com/gfm/#reference-link
542 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
543 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
545 *Matthias St. Pierre*
547 * The test suite is changed to preserve results of each test recipe.
548 A new directory test-runs/ with subdirectories named like the
549 test recipes are created in the build tree for this purpose.
553 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
554 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
555 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
557 *David von Oheimb, Martin Peylo*
559 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
560 The legacy OCSP-focused and only partly documented API is retained for
561 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
565 * Added `util/check-format.pl`, a tool for checking adherence to the
566 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
567 The checks performed are incomplete and yield some false positives.
568 Still the tool should be useful for detecting most typical glitches.
572 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
573 If domain name resolution yields multiple IP addresses all of them are tried
574 after `connect()` failures.
578 * All of the low level RSA functions have been deprecated including:
580 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
581 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
582 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
583 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
584 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
585 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
586 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
587 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
588 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
589 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
590 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
591 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
592 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
593 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
594 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
595 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
596 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
597 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
598 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
599 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
600 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
601 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
602 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
603 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
604 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
605 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
606 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
607 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
609 Use of these low level functions has been informally discouraged for a long
610 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
611 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
612 L<EVP_PKEY_decrypt(3)>.
616 * X509 certificates signed using SHA1 are no longer allowed at security
618 In TLS/SSL the default security level is 1. It can be set either
619 using the cipher string with `@SECLEVEL`, or calling
620 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
621 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
623 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
624 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
625 options of the commands.
629 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
630 modified to use PKEY APIs. These commands are now in maintenance mode
631 and no new features will be added to them.
635 * The command line utility rsautl has been deprecated.
636 Instead use the pkeyutl program.
640 * The command line utilities genrsa and rsa have been modified to use PKEY
641 APIs. They now write PKCS#8 keys by default. These commands are now in
642 maintenance mode and no new features will be added to them.
646 * All of the low level DH functions have been deprecated including:
648 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
649 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
650 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
651 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
652 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
653 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
654 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
655 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
656 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
657 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
658 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
659 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
660 DH_meth_get_generate_params and DH_meth_set_generate_params.
662 Use of these low level functions has been informally discouraged for a long
663 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
664 and L<EVP_PKEY_derive(3)>.
666 Additionally functions that read and write DH objects such as d2i_DHparams,
667 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
668 functions have also been deprecated. Applications should instead use the
669 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
671 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
672 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
673 `EVP_PKEY_set1_DH()` are also deprecated.
674 Applications should instead either read or write an
675 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
676 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
678 *Paul Dale and Matt Caswell*
680 * All of the low level DSA functions have been deprecated including:
682 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
683 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
684 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
685 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
686 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
687 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
688 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
689 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
690 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
691 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
692 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
693 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
694 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
695 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
696 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
697 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
698 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
700 Use of these low level functions has been informally discouraged for a long
701 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
702 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
704 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
705 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
706 `EVP_PKEY_set1_DSA()` are also deprecated.
707 Applications should instead either read or write an
708 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
709 Or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
713 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
714 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
715 This means that applications don't have to look at the curve NID and
716 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
717 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
718 a no-op when the EVP_PKEY is already of the given type.
720 Parameter and key generation is also reworked to make it possible
721 to generate EVP_PKEY_SM2 parameters and keys without having to go
722 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
723 However, code that does the latter will still work as before.
727 * Deprecated low level ECDH and ECDSA functions. These include:
729 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
730 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
733 Use of these low level functions has been informally discouraged for a long
734 time. Instead applications should use the EVP_PKEY_derive(3),
735 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
739 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
740 and EVP_PKEY_decrypt() instead.
741 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
742 and EVP_PKEY_encrypt() instead.
746 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
747 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
748 a new formulation to include all the things it can be used for,
749 as well as words of caution.
753 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
754 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
758 * All of the low level HMAC functions have been deprecated including:
760 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
761 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
764 Use of these low level functions has been informally discouraged for a long
765 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
766 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
767 and L<EVP_MAC_final(3)>.
771 * Over two thousand fixes were made to the documentation, including:
772 - Common options (such as -rand/-writerand, TLS version control, etc)
773 were refactored and point to newly-enhanced descriptions in openssl.pod.
774 - Added style conformance for all options (with help from Richard Levitte),
775 documented all reported missing options, added a CI build to check
776 that all options are documented and that no unimplemented options
778 - Documented some internals, such as all use of environment variables.
779 - Addressed all internal broken L<> references.
783 * All of the low level CMAC functions have been deprecated including:
785 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
786 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
788 Use of these low level functions has been informally discouraged for a long
789 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
790 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
791 and L<EVP_MAC_final(3)>.
795 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
796 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
799 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
800 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
801 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
802 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
803 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
804 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
805 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
806 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
807 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
808 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
810 Use of these low level functions has been informally discouraged
811 for a long time. Applications should use the EVP_DigestInit_ex(3),
812 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
816 * Corrected the documentation of the return values from the `EVP_DigestSign*`
817 set of functions. The documentation mentioned negative values for some
818 errors, but this was never the case, so the mention of negative values
821 Code that followed the documentation and thereby check with something
822 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
826 * All of the low level cipher functions have been deprecated including:
828 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
829 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
830 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
831 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
832 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
833 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
834 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
835 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
836 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
837 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
838 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
839 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
840 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
841 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
842 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
843 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
844 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
845 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
846 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
847 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
848 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
849 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
850 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
851 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
852 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
853 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
854 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
855 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
856 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
858 Use of these low level functions has been informally discouraged for
859 a long time. Applications should use the high level EVP APIs, e.g.
860 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
861 equivalently named decrypt functions instead.
863 *Matt Caswell and Paul Dale*
865 * Removed include/openssl/opensslconf.h.in and replaced it with
866 include/openssl/configuration.h.in, which differs in not including
867 <openssl/macros.h>. A short header include/openssl/opensslconf.h
868 was added to include both.
870 This allows internal hacks where one might need to modify the set
871 of configured macros, for example this if deprecated symbols are
872 still supposed to be available internally:
874 #include <openssl/configuration.h>
876 #undef OPENSSL_NO_DEPRECATED
877 #define OPENSSL_SUPPRESS_DEPRECATED
879 #include <openssl/macros.h>
881 This should not be used by applications that use the exported
882 symbols, as that will lead to linking errors.
886 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
887 used in exponentiation with 512-bit moduli. No EC algorithms are
888 affected. Analysis suggests that attacks against 2-prime RSA1024,
889 3-prime RSA1536, and DSA1024 as a result of this defect would be very
890 difficult to perform and are not believed likely. Attacks against DH512
891 are considered just feasible. However, for an attack the target would
892 have to re-use the DH512 private key, which is not recommended anyway.
893 Also applications directly using the low level API BN_mod_exp may be
894 affected if they use BN_FLG_CONSTTIME.
899 * Most memory-debug features have been deprecated, and the functionality
900 replaced with no-ops.
904 * Added documentation for the STACK API.
908 * Introduced a new method type and API, OSSL_ENCODER, to
909 represent generic encoders. An implementation is expected to
910 be able to encode an object associated with a given name (such
911 as an algorithm name for an asymmetric key) into forms given by
912 implementation properties.
914 Encoders are primarily used from inside libcrypto, through
915 calls to functions like EVP_PKEY_print_private(),
916 PEM_write_bio_PrivateKey() and similar.
918 Encoders are specified in such a way that they can be made to
919 directly handle the provider side portion of an object, if this
920 provider side part comes from the same provider as the encoder
921 itself, but can also be made to handle objects in parametrized
922 form (as an OSSL_PARAM array of data). This allows a provider to
923 offer generic encoders as a service for any other provider.
927 * Added a .pragma directive to the syntax of configuration files, to
928 allow varying behavior in a supported and predictable manner.
929 Currently added pragma:
933 This allows dollar signs to be a keyword character unless it's
934 followed by a opening brace or parenthesis. This is useful for
935 platforms where dollar signs are commonly used in names, such as
936 volume names and system directory names on VMS.
940 * Added functionality to create an EVP_PKEY from user data. This
941 is effectively the same as creating a RSA, DH or DSA object and
942 then assigning them to an EVP_PKEY, but directly using algorithm
943 agnostic EVP functions. A benefit is that this should be future
944 proof for public key algorithms to come.
948 * Change the interpretation of the '--api' configuration option to
949 mean that this is a desired API compatibility level with no
950 further meaning. The previous interpretation, that this would
951 also mean to remove all deprecated symbols up to and including
952 the given version, no requires that 'no-deprecated' is also used
953 in the configuration.
955 When building applications, the desired API compatibility level
956 can be set with the OPENSSL_API_COMPAT macro like before. For
957 API compatibility version below 3.0, the old style numerical
958 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
959 For version 3.0 and on, the value is expected to be the decimal
960 value calculated from the major and minor version like this:
962 MAJOR * 10000 + MINOR * 100
966 -DOPENSSL_API_COMPAT=30000 For 3.0
967 -DOPENSSL_API_COMPAT=30200 For 3.2
969 To hide declarations that are deprecated up to and including the
970 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
971 given when building the application as well.
975 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
976 access to certificate and CRL stores via URIs and OSSL_STORE
979 This adds the following functions:
981 - X509_LOOKUP_store()
982 - X509_STORE_load_file()
983 - X509_STORE_load_path()
984 - X509_STORE_load_store()
985 - SSL_add_store_cert_subjects_to_stack()
986 - SSL_CTX_set_default_verify_store()
987 - SSL_CTX_load_verify_file()
988 - SSL_CTX_load_verify_dir()
989 - SSL_CTX_load_verify_store()
993 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
994 The presence of this system service is determined at run-time.
998 * Added functionality to create an EVP_PKEY context based on data
999 for methods from providers. This takes an algorithm name and a
1000 property query string and simply stores them, with the intent
1001 that any operation that uses this context will use those strings
1002 to fetch the needed methods implicitly, thereby making the port
1003 of application written for pre-3.0 OpenSSL easier.
1007 * The undocumented function NCONF_WIN32() has been deprecated; for
1008 conversion details see the HISTORY section of doc/man5/config.pod
1012 * Introduced the new functions EVP_DigestSignInit_ex() and
1013 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1014 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1015 pages for further details.
1019 * Over two thousand fixes were made to the documentation, including:
1020 adding missing command flags, better style conformance, documentation
1023 *Rich Salz, Richard Levitte*
1025 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1026 X25519, X448, Ed25519 and Ed448.
1030 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1035 * Deprecated the public definition of `ERR_STATE` as well as the function
1036 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1041 * Added ERR functionality to give callers access to the stored function
1042 names that have replaced the older function code based functions.
1044 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1045 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1046 ERR_peek_error_all() and ERR_peek_last_error_all().
1048 These functions have become deprecated: ERR_get_error_line(),
1049 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1050 ERR_peek_last_error_line_data() and ERR_func_error_string().
1052 Users are recommended to use ERR_get_error_all(), or to pick information
1053 with ERR_peek functions and finish off with getting the error code by using
1058 * Extended testing to be verbose for failing tests only. The make variables
1059 VERBOSE_FAILURE or VF can be used to enable this:
1061 $ make VF=1 test # Unix
1062 $ mms /macro=(VF=1) test ! OpenVMS
1063 $ nmake VF=1 test # Windows
1067 * Added the `-copy_extensions` option to the `x509` command for use with
1068 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1069 all extensions in the request are copied to the certificate or vice versa.
1071 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1073 * Added the `-copy_extensions` option to the `req` command for use with
1074 `-x509`. When given with the `copy` or `copyall` argument,
1075 all extensions in the certification request are copied to the certificate.
1079 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1080 they generate are by default RFC 5280 compliant in the following sense:
1081 There is a subjectKeyIdentifier extension with a hash value of the public key
1082 and for not self-signed certs there is an authorityKeyIdentifier extension
1083 with a keyIdentifier field or issuer information identifying the signing key.
1084 This is done unless some configuration overrides the new default behavior,
1085 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1089 * Added several checks to `X509_verify_cert()` according to requirements in
1090 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1091 (which may be done by using the CLI option `-x509_strict`):
1092 * The basicConstraints of CA certificates must be marked critical.
1093 * CA certificates must explicitly include the keyUsage extension.
1094 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1095 * The issuer name of any certificate must not be empty.
1096 * The subject name of CA certs, certs with keyUsage crlSign,
1097 and certs without subjectAlternativeName must not be empty.
1098 * If a subjectAlternativeName extension is given it must not be empty.
1099 * The signatureAlgorithm field and the cert signature must be consistent.
1100 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1101 must not be marked critical.
1102 * The authorityKeyIdentifier must be given for X.509v3 certs
1103 unless they are self-signed.
1104 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1108 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1109 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1113 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1114 used even when parsing explicit parameters, when loading a encoded key
1115 or calling `EC_GROUP_new_from_ecpkparameters()`/
1116 `EC_GROUP_new_from_ecparameters()`.
1117 This prevents bypass of security hardening and performance gains,
1118 especially for curves with specialized EC_METHODs.
1119 By default, if a key encoded with explicit parameters is loaded and later
1120 encoded, the output is still encoded with explicit parameters, even if
1121 internally a "named" EC_GROUP is used for computation.
1125 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1126 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1127 NULL. After this change, only the cofactor parameter can be NULL. It also
1128 does some minimal sanity checks on the passed order.
1133 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1134 An attack is simple, if the first CMS_recipientInfo is valid but the
1135 second CMS_recipientInfo is chosen ciphertext. If the second
1136 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1137 encryption key will be replaced by garbage, and the message cannot be
1138 decoded, but if the RSA decryption fails, the correct encryption key is
1139 used and the recipient will not notice the attack.
1140 As a work around for this potential attack the length of the decrypted
1141 key must be equal to the cipher default key length, in case the
1142 certifiate is not given and all recipientInfo are tried out.
1143 The old behaviour can be re-enabled in the CMS code by setting the
1144 CMS_DEBUG_DECRYPT flag.
1148 * Early start up entropy quality from the DEVRANDOM seed source has been
1149 improved for older Linux systems. The RAND subsystem will wait for
1150 /dev/random to be producing output before seeding from /dev/urandom.
1151 The seeded state is stored for future library initialisations using
1152 a system global shared memory segment. The shared memory identifier
1153 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1154 the desired value. The default identifier is 114.
1158 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1159 when primes for RSA keys are computed.
1160 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1161 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1162 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1163 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1164 This avoids possible fingerprinting of newly generated RSA modules.
1168 * Correct the extended master secret constant on EBCDIC systems. Without this
1169 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1170 negotiate EMS will fail. Unfortunately this also means that TLS connections
1171 between EBCDIC systems with this fix, and EBCDIC systems without this
1172 fix will fail if they negotiate EMS.
1176 * Changed the library initialisation so that the config file is now loaded
1177 by default. This was already the case for libssl. It now occurs for both
1178 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1179 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1183 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1184 where the former acts as a replacement for `ERR_put_error()`, and the
1185 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1186 `ERR_raise_data()` adds more flexibility by taking a format string and
1187 an arbitrary number of arguments following it, to be processed with
1192 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1193 to check if a named provider is loaded and available. When called, it
1194 will also activate all fallback providers if such are still present.
1198 * Enforce a minimum DH modulus size of 512 bits.
1202 * Changed DH parameters to generate the order q subgroup instead of 2q.
1203 Previously generated DH parameters are still accepted by DH_check
1204 but DH_generate_key works around that by clearing bit 0 of the
1205 private key for those. This avoids leaking bit 0 of the private key.
1209 * Significantly reduce secure memory usage by the randomness pools.
1213 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1218 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1219 algorithms. An implementation of a key exchange algorithm can be obtained
1220 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1221 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1222 the older EVP_PKEY_derive_init() function. See the man pages for the new
1223 functions for further details.
1227 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1231 * Removed the function names from error messages and deprecated the
1234 * Removed NextStep support and the macro OPENSSL_UNISTD
1238 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1239 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1240 Also removed "export var as function" capability; we do not export
1241 variables, only functions.
1245 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1246 an error and 1 indicating success. In previous versions of OpenSSL this
1247 was a void type. If a key was set longer than the maximum possible this
1252 * Support SM2 signing and verification schemes with X509 certificate.
1256 * Use SHA256 as the default digest for TS query in the `ts` app.
1260 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1261 This checks that the salt length is at least 128 bits, the derived key
1262 length is at least 112 bits, and that the iteration count is at least 1000.
1263 For backwards compatibility these checks are disabled by default in the
1264 default provider, but are enabled by default in the fips provider.
1265 To enable or disable these checks use the control
1266 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1270 * Default cipher lists/suites are now available via a function, the
1271 #defines are deprecated.
1275 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1276 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1277 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1281 * Join the directories crypto/x509 and crypto/x509v3
1285 * Added command 'openssl kdf' that uses the EVP_KDF API.
1289 * Added command 'openssl mac' that uses the EVP_MAC API.
1293 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1294 as default directories. Also added the command 'openssl info'
1295 for scripting purposes.
1299 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1300 deprecated. These undocumented functions were never integrated into the EVP
1301 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1302 Bi-directional IGE mode. These modes were never formally standardised and
1303 usage of these functions is believed to be very small. In particular
1304 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1305 is ever used. The security implications are believed to be minimal, but
1306 this issue was never fixed for backwards compatibility reasons. New code
1307 should not use these modes.
1311 * Add prediction resistance to the DRBG reseeding process.
1315 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1316 mandated by IEEE Std 1619-2018.
1320 * Added newline escaping functionality to a filename when using openssl dgst.
1321 This output format is to replicate the output format found in the `*sum`
1322 checksum programs. This aims to preserve backward compatibility.
1324 *Matt Eaton, Richard Levitte, and Paul Dale*
1326 * Removed the heartbeat message in DTLS feature, as it has very
1327 little usage and doesn't seem to fulfill a valuable purpose.
1328 The configuration option is now deprecated.
1332 * Changed the output of 'openssl {digestname} < file' to display the
1333 digest name in its output.
1337 * Added a new generic trace API which provides support for enabling
1338 instrumentation through trace output. This feature is mainly intended
1339 as an aid for developers and is disabled by default. To utilize it,
1340 OpenSSL needs to be configured with the `enable-trace` option.
1342 If the tracing API is enabled, the application can activate trace output
1343 by registering BIOs as trace channels for a number of tracing and debugging
1346 The `openssl` program has been expanded to enable any of the types
1347 available via environment variables defined by the user, and serves as
1348 one possible example on how to use this functionality.
1350 *Richard Levitte & Matthias St. Pierre*
1352 * Added build tests for C++. These are generated files that only do one
1353 thing, to include one public OpenSSL head file each. This tests that
1354 the public header files can be usefully included in a C++ application.
1356 This test isn't enabled by default. It can be enabled with the option
1357 'enable-buildtest-c++'.
1361 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1365 * Add KMAC to EVP_MAC.
1369 * Added property based algorithm implementation selection framework to
1374 * Added SCA hardening for modular field inversion in EC_GROUP through
1375 a new dedicated field_inv() pointer in EC_METHOD.
1376 This also addresses a leakage affecting conversions from projective
1377 to affine coordinates.
1379 *Billy Bob Brumley, Nicola Tuveri*
1381 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1382 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1383 those algorithms that were already supported through the EVP_PKEY API
1384 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1385 and scrypt are now wrappers that call EVP_KDF.
1389 * Build devcrypto engine as a dynamic engine.
1391 *Eneas U de Queiroz*
1393 * Add keyed BLAKE2 to EVP_MAC.
1397 * Fix a bug in the computation of the endpoint-pair shared secret used
1398 by DTLS over SCTP. This breaks interoperability with older versions
1399 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1400 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1401 interoperability with such broken implementations. However, enabling
1402 this switch breaks interoperability with correct implementations.
1404 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1405 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1409 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1413 * Change the license to the Apache License v2.0.
1417 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1419 - Major releases (indicated by incrementing the MAJOR release number)
1420 may introduce incompatible API/ABI changes.
1421 - Minor releases (indicated by incrementing the MINOR release number)
1422 may introduce new features but retain API/ABI compatibility.
1423 - Patch releases (indicated by incrementing the PATCH number)
1424 are intended for bug fixes and other improvements of existing
1425 features only (like improving performance or adding documentation)
1426 and retain API/ABI compatibility.
1430 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1434 * Remove the 'dist' target and add a tarball building script. The
1435 'dist' target has fallen out of use, and it shouldn't be
1436 necessary to configure just to create a source distribution.
1440 * Recreate the OS390-Unix config target. It no longer relies on a
1441 special script like it did for OpenSSL pre-1.1.0.
1445 * Instead of having the source directories listed in Configure, add
1446 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1451 * Add GMAC to EVP_MAC.
1455 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1459 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1460 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1461 to facilitate the continued use of MACs through raw private keys in
1462 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1466 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1467 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1471 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1472 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1473 are retained for backwards compatibility.
1477 * AES-XTS mode now enforces that its two keys are different to mitigate
1478 the attacked described in "Efficient Instantiations of Tweakable
1479 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1480 Details of this attack can be obtained from:
1481 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1485 * Rename the object files, i.e. give them other names than in previous
1486 versions. Their names now include the name of the final product, as
1487 well as its type mnemonic (bin, lib, shlib).
1491 * Added new option for 'openssl list', '-objects', which will display the
1492 list of built in objects, i.e. OIDs with names.
1496 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1497 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1502 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1503 improves application performance by removing data copies and providing
1504 applications with zero-copy system calls such as sendfile and splice.
1508 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1509 option is set, openssl cleanses (zeroize) plaintext bytes from
1510 internal buffers after delivering them to the application. Note,
1511 the application is still responsible for cleansing other copies
1512 (e.g.: data received by SSL_read(3)).
1516 * `PKCS12_parse` now maintains the order of the parsed certificates
1517 when outputting them via `*ca` (rather than reversing it).
1521 * Deprecated pthread fork support methods. These were unused so no
1522 replacement is required.
1524 - OPENSSL_fork_prepare()
1525 - OPENSSL_fork_parent()
1526 - OPENSSL_fork_child()
1533 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1535 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1537 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1538 create a unique hash value based on the issuer and serial number data
1539 contained within an X509 certificate. However it was failing to correctly
1540 handle any errors that may occur while parsing the issuer field (which might
1541 occur if the issuer field is maliciously constructed). This may subsequently
1542 result in a NULL pointer deref and a crash leading to a potential denial of
1548 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1549 padding mode to correctly check for rollback attacks. This is considered a
1550 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1555 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1556 functions. Previously they could overflow the output length argument in some
1557 cases where the input length is close to the maximum permissable length for
1558 an integer on the platform. In such cases the return value from the function
1559 call would be 1 (indicating success), but the output length value would be
1560 negative. This could cause applications to behave incorrectly or crash.
1565 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1566 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1567 could be exploited in a side channel attack to recover the password. Since
1568 the attack is local host only this is outside of the current OpenSSL
1569 threat model and therefore no CVE is assigned.
1571 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1576 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1578 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1579 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1580 If an attacker can control both items being compared then this could lead
1581 to a possible denial of service attack. OpenSSL itself uses the
1582 GENERAL_NAME_cmp function for two purposes:
1583 1) Comparing CRL distribution point names between an available CRL and a
1584 CRL distribution point embedded in an X509 certificate
1585 2) When verifying that a timestamp response token signer matches the
1586 timestamp authority name (exposed via the API functions
1587 TS_RESP_verify_response and TS_RESP_verify_token)
1592 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1594 * Certificates with explicit curve parameters are now disallowed in
1595 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1599 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1600 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1601 conversely, silently ignore DTLS protocol version bounds when configuring
1602 TLS-based contexts. The commands can be repeated to set bounds of both
1603 types. The same applies with the corresponding "min_protocol" and
1604 "max_protocol" command-line switches, in case some application uses both TLS
1607 SSL_CTX instances that are created for a fixed protocol version (e.g.
1608 TLSv1_server_method()) also silently ignore version bounds. Previously
1609 attempts to apply bounds to these protocol versions would result in an
1610 error. Now only the "version-flexible" SSL_CTX instances are subject to
1611 limits in configuration files in command-line options.
1615 * Handshake now fails if Extended Master Secret extension is dropped
1620 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1622 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1624 * Fixed segmentation fault in SSL_check_chain()
1625 Server or client applications that call the SSL_check_chain() function
1626 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1627 dereference as a result of incorrect handling of the
1628 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1629 or unrecognised signature algorithm is received from the peer. This could
1630 be exploited by a malicious peer in a Denial of Service attack.
1635 * Added AES consttime code for no-asm configurations
1636 an optional constant time support for AES was added
1637 when building openssl for no-asm.
1638 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1639 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1640 At this time this feature is by default disabled.
1641 It will be enabled by default in 3.0.
1645 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1647 * Revert the change of EOF detection while reading in libssl to avoid
1648 regressions in applications depending on the current way of reporting
1649 the EOF. As the existing method is not fully accurate the change to
1650 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1651 branch and will be present in the 3.0 release.
1655 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1656 when primes for RSA keys are computed.
1657 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1658 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1659 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1660 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1661 This avoids possible fingerprinting of newly generated RSA modules.
1665 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1667 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1668 while reading in libssl then we would report an error back to the
1669 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1670 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1671 therefore give a hint as to what went wrong.
1675 * Check that ed25519 and ed448 are allowed by the security level. Previously
1676 signature algorithms not using an MD were not being checked that they were
1677 allowed by the security level.
1681 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1682 was not quite right. The behaviour was not consistent between resumption
1683 and normal handshakes, and also not quite consistent with historical
1684 behaviour. The behaviour in various scenarios has been clarified and
1685 it has been updated to make it match historical behaviour as closely as
1690 * *[VMS only]* The header files that the VMS compilers include automatically,
1691 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1692 that the C++ compiler doesn't understand. This is a shortcoming in the
1693 compiler, but can be worked around with `__cplusplus` guards.
1695 C++ applications that use OpenSSL libraries must be compiled using the
1696 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1697 functions. Otherwise, only functions with symbols of less than 31
1698 characters can be used, as the linker will not be able to successfully
1699 resolve symbols with longer names.
1703 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1704 The presence of this system service is determined at run-time.
1708 * Added newline escaping functionality to a filename when using openssl dgst.
1709 This output format is to replicate the output format found in the `*sum`
1710 checksum programs. This aims to preserve backward compatibility.
1712 *Matt Eaton, Richard Levitte, and Paul Dale*
1714 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1719 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1721 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1722 number generator (RNG). This was intended to include protection in the
1723 event of a fork() system call in order to ensure that the parent and child
1724 processes did not share the same RNG state. However this protection was not
1725 being used in the default case.
1727 A partial mitigation for this issue is that the output from a high
1728 precision timer is mixed into the RNG state so the likelihood of a parent
1729 and child process sharing state is significantly reduced.
1731 If an application already calls OPENSSL_init_crypto() explicitly using
1732 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1735 *Matthias St. Pierre*
1737 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1738 used even when parsing explicit parameters, when loading a encoded key
1739 or calling `EC_GROUP_new_from_ecpkparameters()`/
1740 `EC_GROUP_new_from_ecparameters()`.
1741 This prevents bypass of security hardening and performance gains,
1742 especially for curves with specialized EC_METHODs.
1743 By default, if a key encoded with explicit parameters is loaded and later
1744 encoded, the output is still encoded with explicit parameters, even if
1745 internally a "named" EC_GROUP is used for computation.
1749 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1750 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1751 NULL. After this change, only the cofactor parameter can be NULL. It also
1752 does some minimal sanity checks on the passed order.
1757 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1758 An attack is simple, if the first CMS_recipientInfo is valid but the
1759 second CMS_recipientInfo is chosen ciphertext. If the second
1760 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1761 encryption key will be replaced by garbage, and the message cannot be
1762 decoded, but if the RSA decryption fails, the correct encryption key is
1763 used and the recipient will not notice the attack.
1764 As a work around for this potential attack the length of the decrypted
1765 key must be equal to the cipher default key length, in case the
1766 certifiate is not given and all recipientInfo are tried out.
1767 The old behaviour can be re-enabled in the CMS code by setting the
1768 CMS_DEBUG_DECRYPT flag.
1773 * Early start up entropy quality from the DEVRANDOM seed source has been
1774 improved for older Linux systems. The RAND subsystem will wait for
1775 /dev/random to be producing output before seeding from /dev/urandom.
1776 The seeded state is stored for future library initialisations using
1777 a system global shared memory segment. The shared memory identifier
1778 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1779 the desired value. The default identifier is 114.
1783 * Correct the extended master secret constant on EBCDIC systems. Without this
1784 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1785 negotiate EMS will fail. Unfortunately this also means that TLS connections
1786 between EBCDIC systems with this fix, and EBCDIC systems without this
1787 fix will fail if they negotiate EMS.
1791 * Use Windows installation paths in the mingw builds
1793 Mingw isn't a POSIX environment per se, which means that Windows
1794 paths should be used for installation.
1799 * Changed DH_check to accept parameters with order q and 2q subgroups.
1800 With order 2q subgroups the bit 0 of the private key is not secret
1801 but DH_generate_key works around that by clearing bit 0 of the
1802 private key for those. This avoids leaking bit 0 of the private key.
1806 * Significantly reduce secure memory usage by the randomness pools.
1810 * Revert the DEVRANDOM_WAIT feature for Linux systems
1812 The DEVRANDOM_WAIT feature added a select() call to wait for the
1813 /dev/random device to become readable before reading from the
1814 /dev/urandom device.
1816 It turned out that this change had negative side effects on
1817 performance which were not acceptable. After some discussion it
1818 was decided to revert this feature and leave it up to the OS
1819 resp. the platform maintainer to ensure a proper initialization
1820 during early boot time.
1822 *Matthias St. Pierre*
1824 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1826 * Add build tests for C++. These are generated files that only do one
1827 thing, to include one public OpenSSL head file each. This tests that
1828 the public header files can be usefully included in a C++ application.
1830 This test isn't enabled by default. It can be enabled with the option
1831 'enable-buildtest-c++'.
1835 * Enable SHA3 pre-hashing for ECDSA and DSA.
1839 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1840 This changes the size when using the `genpkey` command when no size is given.
1841 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1842 generation commands to use 2048 bits by default.
1846 * Reorganize the manual pages to consistently have RETURN VALUES,
1847 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1848 util/fix-doc-nits accordingly.
1850 *Paul Yang, Joshua Lock*
1852 * Add the missing accessor EVP_PKEY_get0_engine()
1856 * Have commands like `s_client` and `s_server` output the signature scheme
1857 along with other cipher suite parameters when debugging.
1861 * Make OPENSSL_config() error agnostic again.
1865 * Do the error handling in RSA decryption constant time.
1869 * Prevent over long nonces in ChaCha20-Poly1305.
1871 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1872 for every encryption operation. RFC 7539 specifies that the nonce value
1873 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1874 and front pads the nonce with 0 bytes if it is less than 12
1875 bytes. However it also incorrectly allows a nonce to be set of up to 16
1876 bytes. In this case only the last 12 bytes are significant and any
1877 additional leading bytes are ignored.
1879 It is a requirement of using this cipher that nonce values are
1880 unique. Messages encrypted using a reused nonce value are susceptible to
1881 serious confidentiality and integrity attacks. If an application changes
1882 the default nonce length to be longer than 12 bytes and then makes a
1883 change to the leading bytes of the nonce expecting the new value to be a
1884 new unique nonce then such an application could inadvertently encrypt
1885 messages with a reused nonce.
1887 Additionally the ignored bytes in a long nonce are not covered by the
1888 integrity guarantee of this cipher. Any application that relies on the
1889 integrity of these ignored leading bytes of a long nonce may be further
1890 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1891 is safe because no such use sets such a long nonce value. However user
1892 applications that use this cipher directly and set a non-default nonce
1893 length to be longer than 12 bytes may be vulnerable.
1895 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1901 * Add DEVRANDOM_WAIT feature for Linux systems
1903 On older Linux systems where the getrandom() system call is not available,
1904 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1905 Contrary to getrandom(), the /dev/urandom device will not block during
1906 early boot when the kernel CSPRNG has not been seeded yet.
1908 To mitigate this known weakness, use select() to wait for /dev/random to
1909 become readable before reading from /dev/urandom.
1911 * Ensure that SM2 only uses SM3 as digest algorithm
1915 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1917 * Change the info callback signals for the start and end of a post-handshake
1918 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1919 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1920 confused by this and assume that a TLSv1.2 renegotiation has started. This
1921 can break KeyUpdate handling. Instead we no longer signal the start and end
1922 of a post handshake message exchange (although the messages themselves are
1923 still signalled). This could break some applications that were expecting
1924 the old signals. However without this KeyUpdate is not usable for many
1929 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1931 * Timing vulnerability in DSA signature generation
1933 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1934 timing side channel attack. An attacker could use variations in the signing
1935 algorithm to recover the private key.
1937 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1942 * Timing vulnerability in ECDSA signature generation
1944 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1945 timing side channel attack. An attacker could use variations in the signing
1946 algorithm to recover the private key.
1948 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1953 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1954 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1955 of two gigabytes and the error handling improved.
1957 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1958 categorized as a normal bug, not a security issue, because the DRBG reseeds
1959 automatically and is fully functional even without additional randomness
1960 provided by the application.
1962 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1964 * Add a new ClientHello callback. Provides a callback interface that gives
1965 the application the ability to adjust the nascent SSL object at the
1966 earliest stage of ClientHello processing, immediately after extensions have
1967 been collected but before they have been processed. In particular, this
1968 callback can adjust the supported TLS versions in response to the contents
1973 * Add SM2 base algorithm support.
1977 * s390x assembly pack: add (improved) hardware-support for the following
1978 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1979 aes-cfb/cfb8, aes-ecb.
1983 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1984 parameter is no longer accepted, as it leads to a corrupt table. NULL
1985 pem_str is reserved for alias entries only.
1989 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1990 step for prime curves. The new implementation is based on formulae from
1991 differential addition-and-doubling in homogeneous projective coordinates
1992 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1993 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1994 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1995 to work in projective coordinates.
1997 *Billy Bob Brumley, Nicola Tuveri*
1999 * Change generating and checking of primes so that the error rate of not
2000 being prime depends on the intended use based on the size of the input.
2001 For larger primes this will result in more rounds of Miller-Rabin.
2002 The maximal error rate for primes with more than 1080 bits is lowered
2005 *Kurt Roeckx, Annie Yousar*
2007 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2011 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2012 moving between systems, and to avoid confusion when a Windows build is
2013 done with mingw vs with MSVC. For POSIX installs, there's still a
2014 symlink or copy named 'tsget' to avoid that confusion as well.
2018 * Revert blinding in ECDSA sign and instead make problematic addition
2019 length-invariant. Switch even to fixed-length Montgomery multiplication.
2023 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2024 step for binary curves. The new implementation is based on formulae from
2025 differential addition-and-doubling in mixed Lopez-Dahab projective
2026 coordinates, modified to independently blind the operands.
2028 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2030 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2031 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2032 EC_METHODs to implement their own specialized "ladder step", to take
2033 advantage of more favorable coordinate systems or more efficient
2034 differential addition-and-doubling algorithms.
2036 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2038 * Modified the random device based seed sources to keep the relevant
2039 file descriptors open rather than reopening them on each access.
2040 This allows such sources to operate in a chroot() jail without
2041 the associated device nodes being available. This behaviour can be
2042 controlled using RAND_keep_random_devices_open().
2046 * Numerous side-channel attack mitigations have been applied. This may have
2047 performance impacts for some algorithms for the benefit of improved
2048 security. Specific changes are noted in this change log by their respective
2053 * AIX shared library support overhaul. Switch to AIX "natural" way of
2054 handling shared libraries, which means collecting shared objects of
2055 different versions and bitnesses in one common archive. This allows to
2056 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2057 doesn't affect the way 3rd party applications are linked, only how
2058 multi-version installation is managed.
2062 * Make ec_group_do_inverse_ord() more robust and available to other
2063 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2064 mitigations are applied to the fallback BN_mod_inverse().
2065 When using this function rather than BN_mod_inverse() directly, new
2066 EC cryptosystem implementations are then safer-by-default.
2070 * Add coordinate blinding for EC_POINT and implement projective
2071 coordinate blinding for generic prime curves as a countermeasure to
2072 chosen point SCA attacks.
2074 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2076 * Add blinding to ECDSA and DSA signatures to protect against side channel
2077 attacks discovered by Keegan Ryan (NCC Group).
2081 * Enforce checking in the `pkeyutl` command to ensure that the input
2082 length does not exceed the maximum supported digest length when performing
2083 a sign, verify or verifyrecover operation.
2087 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2088 I/O in combination with something like select() or poll() will hang. This
2089 can be turned off again using SSL_CTX_clear_mode().
2090 Many applications do not properly handle non-application data records, and
2091 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2092 around the problems in those applications, but can also break some.
2093 It's recommended to read the manpages about SSL_read(), SSL_write(),
2094 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2095 SSL_CTX_set_read_ahead() again.
2099 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2100 now allow empty (zero character) pass phrases.
2104 * Apply blinding to binary field modular inversion and remove patent
2105 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2109 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2110 binary and prime elliptic curves.
2114 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2115 constant time fixed point multiplication.
2119 * Revise elliptic curve scalar multiplication with timing attack
2120 defenses: ec_wNAF_mul redirects to a constant time implementation
2121 when computing fixed point and variable point multiplication (which
2122 in OpenSSL are mostly used with secret scalars in keygen, sign,
2123 ECDH derive operations).
2124 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2127 * Updated CONTRIBUTING
2131 * Updated DRBG / RAND to request nonce and additional low entropy
2132 randomness from the system.
2134 *Matthias St. Pierre*
2136 * Updated 'openssl rehash' to use OpenSSL consistent default.
2140 * Moved the load of the ssl_conf module to libcrypto, which helps
2141 loading engines that libssl uses before libssl is initialised.
2145 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2149 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2151 *Ingo Schwarze, Rich Salz*
2153 * Added output of accepting IP address and port for 'openssl s_server'
2157 * Added a new API for TLSv1.3 ciphersuites:
2158 SSL_CTX_set_ciphersuites()
2159 SSL_set_ciphersuites()
2163 * Memory allocation failures consistently add an error to the error
2168 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2169 in libcrypto when run as setuid/setgid.
2173 * Load any config file by default when libssl is used.
2177 * Added new public header file <openssl/rand_drbg.h> and documentation
2178 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2180 *Matthias St. Pierre*
2182 * QNX support removed (cannot find contributors to get their approval
2183 for the license change).
2187 * TLSv1.3 replay protection for early data has been implemented. See the
2188 SSL_read_early_data() man page for further details.
2192 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2193 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2194 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2195 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2196 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2197 configuration has been separated out. See the ciphers man page or the
2198 SSL_CTX_set_ciphersuites() man page for more information.
2202 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2203 in responder mode now supports the new "-multi" option, which
2204 spawns the specified number of child processes to handle OCSP
2205 requests. The "-timeout" option now also limits the OCSP
2206 responder's patience to wait to receive the full client request
2207 on a newly accepted connection. Child processes are respawned
2208 as needed, and the CA index file is automatically reloaded
2209 when changed. This makes it possible to run the "ocsp" responder
2210 as a long-running service, making the OpenSSL CA somewhat more
2211 feature-complete. In this mode, most diagnostic messages logged
2212 after entering the event loop are logged via syslog(3) rather than
2217 * Added support for X448 and Ed448. Heavily based on original work by
2222 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2223 objects loaded. This adds the functions OSSL_STORE_expect() and
2224 OSSL_STORE_find() as well as needed tools to construct searches and
2225 get the search data out of them.
2229 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2230 version of OpenSSL should review their configuration settings to ensure
2231 that they are still appropriate for TLSv1.3. For further information see:
2232 <https://wiki.openssl.org/index.php/TLS1.3>
2236 * Grand redesign of the OpenSSL random generator
2238 The default RAND method now utilizes an AES-CTR DRBG according to
2239 NIST standard SP 800-90Ar1. The new random generator is essentially
2240 a port of the default random generator from the OpenSSL FIPS 2.0
2241 object module. It is a hybrid deterministic random bit generator
2242 using an AES-CTR bit stream and which seeds and reseeds itself
2243 automatically using trusted system entropy sources.
2245 Some of its new features are:
2246 - Support for multiple DRBG instances with seed chaining.
2247 - The default RAND method makes use of a DRBG.
2248 - There is a public and private DRBG instance.
2249 - The DRBG instances are fork-safe.
2250 - Keep all global DRBG instances on the secure heap if it is enabled.
2251 - The public and private DRBG instance are per thread for lock free
2254 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2256 * Changed Configure so it only says what it does and doesn't dump
2257 so much data. Instead, ./configdata.pm should be used as a script
2258 to display all sorts of configuration data.
2262 * Added processing of "make variables" to Configure.
2266 * Added SHA512/224 and SHA512/256 algorithm support.
2270 * The last traces of Netware support, first removed in 1.1.0, have
2275 * Get rid of Makefile.shared, and in the process, make the processing
2276 of certain files (rc.obj, or the .def/.map/.opt files produced from
2277 the ordinal files) more visible and hopefully easier to trace and
2278 debug (or make silent).
2282 * Make it possible to have environment variable assignments as
2283 arguments to config / Configure.
2287 * Add multi-prime RSA (RFC 8017) support.
2291 * Add SM3 implemented according to GB/T 32905-2016
2292 *Jack Lloyd <jack.lloyd@ribose.com>,*
2293 *Ronald Tse <ronald.tse@ribose.com>,*
2294 *Erick Borsboom <erick.borsboom@ribose.com>*
2296 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2297 as documented in RFC6066.
2298 Based on a patch from Tomasz Moń
2300 *Filipe Raimundo da Silva*
2302 * Add SM4 implemented according to GB/T 32907-2016.
2303 *Jack Lloyd <jack.lloyd@ribose.com>,*
2304 *Ronald Tse <ronald.tse@ribose.com>,*
2305 *Erick Borsboom <erick.borsboom@ribose.com>*
2307 * Reimplement -newreq-nodes and ERR_error_string_n; the
2308 original author does not agree with the license change.
2312 * Add ARIA AEAD TLS support.
2316 * Some macro definitions to support VS6 have been removed. Visual
2317 Studio 6 has not worked since 1.1.0
2321 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2322 without clearing the errors.
2326 * Add "atfork" functions. If building on a system that without
2327 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2328 requirements. The RAND facility now uses/requires this.
2336 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2337 not possible to disable entirely. However, it's still possible to
2338 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2341 To disable, configure with 'no-ui-console'. 'no-ui' is still
2342 possible to use as an alias. Check at compile time with the
2343 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2344 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2348 * Add a STORE module, which implements a uniform and URI based reader of
2349 stores that can contain keys, certificates, CRLs and numerous other
2350 objects. The main API is loosely based on a few stdio functions,
2351 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2352 OSSL_STORE_error and OSSL_STORE_close.
2353 The implementation uses backends called "loaders" to implement arbitrary
2354 URI schemes. There is one built in "loader" for the 'file' scheme.
2358 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2359 then adjusted to work on FreeBSD 8.4 as well.
2360 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2361 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2365 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2366 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2367 error code calls like this:
2369 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2371 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2372 that can be encoded in C. For the foreseeable future, this will only
2375 *Richard Levitte and Tim Hudson*
2377 * Removed BSD cryptodev engine.
2381 * Add a build target 'build_all_generated', to build all generated files
2382 and only that. This can be used to prepare everything that requires
2383 things like perl for a system that lacks perl and then move everything
2384 to that system and do the rest of the build there.
2388 * In the UI interface, make it possible to duplicate the user data. This
2389 can be used by engines that need to retain the data for a longer time
2390 than just the call where this user data is passed.
2394 * Ignore the '-named_curve auto' value for compatibility of applications
2397 *Tomáš Mráz <tmraz@fedoraproject.org>*
2399 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2400 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2401 alerts across multiple records (some of which could be empty). In practice
2402 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2403 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2404 support this at all. Supporting it adds significant complexity to the
2405 record layer, and its removal is unlikely to cause interoperability
2410 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2411 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2412 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2417 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2418 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2420 *Richard Levitte, Andy Polyakov*
2422 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2427 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2428 platform rather than 'mingw'.
2432 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2433 success if they are asked to add an object which already exists
2434 in the store. This change cascades to other functions which load
2435 certificates and CRLs.
2439 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2440 facilitate stack unwinding even from assembly subroutines.
2444 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2445 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2449 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2450 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2451 which is the minimum version we support.
2455 * Certificate time validation (X509_cmp_time) enforces stricter
2456 compliance with RFC 5280. Fractional seconds and timezone offsets
2457 are no longer allowed.
2461 * Add support for ARIA
2465 * s_client will now send the Server Name Indication (SNI) extension by
2466 default unless the new "-noservername" option is used. The server name is
2467 based on the host provided to the "-connect" option unless overridden by
2468 using "-servername".
2472 * Add support for SipHash
2476 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2477 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2478 prevent issues where no progress is being made and the peer continually
2479 sends unrecognised record types, using up resources processing them.
2483 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2484 using the algorithm defined in
2485 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2489 * Heartbeat support has been removed; the ABI is changed for now.
2491 *Richard Levitte, Rich Salz*
2493 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2497 * The RSA "null" method, which was partially supported to avoid patent
2498 issues, has been replaced to always returns NULL.
2505 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2507 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2508 used even when parsing explicit parameters, when loading a encoded key
2509 or calling `EC_GROUP_new_from_ecpkparameters()`/
2510 `EC_GROUP_new_from_ecparameters()`.
2511 This prevents bypass of security hardening and performance gains,
2512 especially for curves with specialized EC_METHODs.
2513 By default, if a key encoded with explicit parameters is loaded and later
2514 encoded, the output is still encoded with explicit parameters, even if
2515 internally a "named" EC_GROUP is used for computation.
2519 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2520 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2521 NULL. After this change, only the cofactor parameter can be NULL. It also
2522 does some minimal sanity checks on the passed order.
2527 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2528 An attack is simple, if the first CMS_recipientInfo is valid but the
2529 second CMS_recipientInfo is chosen ciphertext. If the second
2530 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2531 encryption key will be replaced by garbage, and the message cannot be
2532 decoded, but if the RSA decryption fails, the correct encryption key is
2533 used and the recipient will not notice the attack.
2534 As a work around for this potential attack the length of the decrypted
2535 key must be equal to the cipher default key length, in case the
2536 certifiate is not given and all recipientInfo are tried out.
2537 The old behaviour can be re-enabled in the CMS code by setting the
2538 CMS_DEBUG_DECRYPT flag.
2543 * Use Windows installation paths in the mingw builds
2545 Mingw isn't a POSIX environment per se, which means that Windows
2546 paths should be used for installation.
2551 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2553 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2554 This changes the size when using the `genpkey` command when no size is given.
2555 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2556 generation commands to use 2048 bits by default.
2560 * Prevent over long nonces in ChaCha20-Poly1305.
2562 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2563 for every encryption operation. RFC 7539 specifies that the nonce value
2564 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2565 and front pads the nonce with 0 bytes if it is less than 12
2566 bytes. However it also incorrectly allows a nonce to be set of up to 16
2567 bytes. In this case only the last 12 bytes are significant and any
2568 additional leading bytes are ignored.
2570 It is a requirement of using this cipher that nonce values are
2571 unique. Messages encrypted using a reused nonce value are susceptible to
2572 serious confidentiality and integrity attacks. If an application changes
2573 the default nonce length to be longer than 12 bytes and then makes a
2574 change to the leading bytes of the nonce expecting the new value to be a
2575 new unique nonce then such an application could inadvertently encrypt
2576 messages with a reused nonce.
2578 Additionally the ignored bytes in a long nonce are not covered by the
2579 integrity guarantee of this cipher. Any application that relies on the
2580 integrity of these ignored leading bytes of a long nonce may be further
2581 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2582 is safe because no such use sets such a long nonce value. However user
2583 applications that use this cipher directly and set a non-default nonce
2584 length to be longer than 12 bytes may be vulnerable.
2586 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2592 * Added SCA hardening for modular field inversion in EC_GROUP through
2593 a new dedicated field_inv() pointer in EC_METHOD.
2594 This also addresses a leakage affecting conversions from projective
2595 to affine coordinates.
2597 *Billy Bob Brumley, Nicola Tuveri*
2599 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2600 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2604 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2608 * Remove the 'dist' target and add a tarball building script. The
2609 'dist' target has fallen out of use, and it shouldn't be
2610 necessary to configure just to create a source distribution.
2614 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2616 * Timing vulnerability in DSA signature generation
2618 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2619 timing side channel attack. An attacker could use variations in the signing
2620 algorithm to recover the private key.
2622 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2627 * Timing vulnerability in ECDSA signature generation
2629 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2630 timing side channel attack. An attacker could use variations in the signing
2631 algorithm to recover the private key.
2633 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2638 * Add coordinate blinding for EC_POINT and implement projective
2639 coordinate blinding for generic prime curves as a countermeasure to
2640 chosen point SCA attacks.
2642 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2644 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2646 * Client DoS due to large DH parameter
2648 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2649 malicious server can send a very large prime value to the client. This will
2650 cause the client to spend an unreasonably long period of time generating a
2651 key for this prime resulting in a hang until the client has finished. This
2652 could be exploited in a Denial Of Service attack.
2654 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2659 * Cache timing vulnerability in RSA Key Generation
2661 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2662 a cache timing side channel attack. An attacker with sufficient access to
2663 mount cache timing attacks during the RSA key generation process could
2664 recover the private key.
2666 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2667 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2672 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2673 parameter is no longer accepted, as it leads to a corrupt table. NULL
2674 pem_str is reserved for alias entries only.
2678 * Revert blinding in ECDSA sign and instead make problematic addition
2679 length-invariant. Switch even to fixed-length Montgomery multiplication.
2683 * Change generating and checking of primes so that the error rate of not
2684 being prime depends on the intended use based on the size of the input.
2685 For larger primes this will result in more rounds of Miller-Rabin.
2686 The maximal error rate for primes with more than 1080 bits is lowered
2689 *Kurt Roeckx, Annie Yousar*
2691 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2695 * Add blinding to ECDSA and DSA signatures to protect against side channel
2696 attacks discovered by Keegan Ryan (NCC Group).
2700 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2701 now allow empty (zero character) pass phrases.
2705 * Certificate time validation (X509_cmp_time) enforces stricter
2706 compliance with RFC 5280. Fractional seconds and timezone offsets
2707 are no longer allowed.
2711 * Fixed a text canonicalisation bug in CMS
2713 Where a CMS detached signature is used with text content the text goes
2714 through a canonicalisation process first prior to signing or verifying a
2715 signature. This process strips trailing space at the end of lines, converts
2716 line terminators to CRLF and removes additional trailing line terminators
2717 at the end of a file. A bug in the canonicalisation process meant that
2718 some characters, such as form-feed, were incorrectly treated as whitespace
2719 and removed. This is contrary to the specification (RFC5485). This fix
2720 could mean that detached text data signed with an earlier version of
2721 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2722 signed with a fixed OpenSSL may fail to verify with an earlier version of
2723 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2724 and use the "-binary" flag (for the "cms" command line application) or set
2725 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2729 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2731 * Constructed ASN.1 types with a recursive definition could exceed the stack
2733 Constructed ASN.1 types with a recursive definition (such as can be found
2734 in PKCS7) could eventually exceed the stack given malicious input with
2735 excessive recursion. This could result in a Denial Of Service attack. There
2736 are no such structures used within SSL/TLS that come from untrusted sources
2737 so this is considered safe.
2739 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2745 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2747 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2748 effectively reduced to only comparing the least significant bit of each
2749 byte. This allows an attacker to forge messages that would be considered as
2750 authenticated in an amount of tries lower than that guaranteed by the
2751 security claims of the scheme. The module can only be compiled by the
2752 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2754 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2760 * Add a build target 'build_all_generated', to build all generated files
2761 and only that. This can be used to prepare everything that requires
2762 things like perl for a system that lacks perl and then move everything
2763 to that system and do the rest of the build there.
2767 * Backport SSL_OP_NO_RENGOTIATION
2769 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2770 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2771 changes this is no longer possible in 1.1.0. Therefore the new
2772 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2773 1.1.0 to provide equivalent functionality.
2775 Note that if an application built against 1.1.0h headers (or above) is run
2776 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2777 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2781 * Removed the OS390-Unix config target. It relied on a script that doesn't
2786 * rsaz_1024_mul_avx2 overflow bug on x86_64
2788 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2789 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2790 Analysis suggests that attacks against RSA and DSA as a result of this
2791 defect would be very difficult to perform and are not believed likely.
2792 Attacks against DH1024 are considered just feasible, because most of the
2793 work necessary to deduce information about a private key may be performed
2794 offline. The amount of resources required for such an attack would be
2795 significant. However, for an attack on TLS to be meaningful, the server
2796 would have to share the DH1024 private key among multiple clients, which is
2797 no longer an option since CVE-2016-0701.
2799 This only affects processors that support the AVX2 but not ADX extensions
2800 like Intel Haswell (4th generation).
2802 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2803 was originally found via the OSS-Fuzz project.
2808 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2810 * bn_sqrx8x_internal carry bug 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.
2824 This only affects processors that support the BMI1, BMI2 and ADX extensions
2825 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2827 This issue was reported to OpenSSL by the OSS-Fuzz project.
2832 * Malformed X.509 IPAddressFamily could cause OOB read
2834 If an X.509 certificate has a malformed IPAddressFamily extension,
2835 OpenSSL could do a one-byte buffer overread. The most likely result
2836 would be an erroneous display of the certificate in text format.
2838 This issue was reported to OpenSSL by the OSS-Fuzz project.
2843 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2845 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2846 platform rather than 'mingw'.
2850 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2851 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2852 which is the minimum version we support.
2856 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2858 * Encrypt-Then-Mac renegotiation crash
2860 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2861 negotiated where it was not in the original handshake (or vice-versa) then
2862 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2863 and servers are affected.
2865 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2870 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2872 * Truncated packet could crash via OOB read
2874 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2875 cipher is being used, then a truncated packet can cause that host to
2876 perform an out-of-bounds read, usually resulting in a crash.
2878 This issue was reported to OpenSSL by Robert Święcki of Google.
2883 * Bad (EC)DHE parameters cause a client crash
2885 If a malicious server supplies bad parameters for a DHE or ECDHE key
2886 exchange then this can result in the client attempting to dereference a
2887 NULL pointer leading to a client crash. This could be exploited in a Denial
2890 This issue was reported to OpenSSL by Guido Vranken.
2895 * BN_mod_exp may produce incorrect results on x86_64
2897 There is a carry propagating bug in the x86_64 Montgomery squaring
2898 procedure. No EC algorithms are affected. Analysis suggests that attacks
2899 against RSA and DSA as a result of this defect would be very difficult to
2900 perform and are not believed likely. Attacks against DH are considered just
2901 feasible (although very difficult) because most of the work necessary to
2902 deduce information about a private key may be performed offline. The amount
2903 of resources required for such an attack would be very significant and
2904 likely only accessible to a limited number of attackers. An attacker would
2905 additionally need online access to an unpatched system using the target
2906 private key in a scenario with persistent DH parameters and a private
2907 key that is shared between multiple clients. For example this can occur by
2908 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2909 similar to CVE-2015-3193 but must be treated as a separate problem.
2911 This issue was reported to OpenSSL by the OSS-Fuzz project.
2916 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2918 * ChaCha20/Poly1305 heap-buffer-overflow
2920 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2921 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2922 crash. This issue is not considered to be exploitable beyond a DoS.
2924 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2929 * CMS Null dereference
2931 Applications parsing invalid CMS structures can crash with a NULL pointer
2932 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2933 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2934 structure callback if an attempt is made to free certain invalid encodings.
2935 Only CHOICE structures using a callback which do not handle NULL value are
2938 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2943 * Montgomery multiplication may produce incorrect results
2945 There is a carry propagating bug in the Broadwell-specific Montgomery
2946 multiplication procedure that handles input lengths divisible by, but
2947 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2948 and DH private keys are impossible. This is because the subroutine in
2949 question is not used in operations with the private key itself and an input
2950 of the attacker's direct choice. Otherwise the bug can manifest itself as
2951 transient authentication and key negotiation failures or reproducible
2952 erroneous outcome of public-key operations with specially crafted input.
2953 Among EC algorithms only Brainpool P-512 curves are affected and one
2954 presumably can attack ECDH key negotiation. Impact was not analyzed in
2955 detail, because pre-requisites for attack are considered unlikely. Namely
2956 multiple clients have to choose the curve in question and the server has to
2957 share the private key among them, neither of which is default behaviour.
2958 Even then only clients that chose the curve will be affected.
2960 This issue was publicly reported as transient failures and was not
2961 initially recognized as a security issue. Thanks to Richard Morgan for
2962 providing reproducible case.
2967 * Removed automatic addition of RPATH in shared libraries and executables,
2968 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2972 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2974 * Fix Use After Free for large message sizes
2976 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2977 message larger than approx 16k is received then the underlying buffer to
2978 store the incoming message is reallocated and moved. Unfortunately a
2979 dangling pointer to the old location is left which results in an attempt to
2980 write to the previously freed location. This is likely to result in a
2981 crash, however it could potentially lead to execution of arbitrary code.
2983 This issue only affects OpenSSL 1.1.0a.
2985 This issue was reported to OpenSSL by Robert Święcki.
2990 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2992 * OCSP Status Request extension unbounded memory growth
2994 A malicious client can send an excessively large OCSP Status Request
2995 extension. If that client continually requests renegotiation, sending a
2996 large OCSP Status Request extension each time, then there will be unbounded
2997 memory growth on the server. This will eventually lead to a Denial Of
2998 Service attack through memory exhaustion. Servers with a default
2999 configuration are vulnerable even if they do not support OCSP. Builds using
3000 the "no-ocsp" build time option are not affected.
3002 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3007 * SSL_peek() hang on empty record
3009 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3010 sends an empty record. This could be exploited by a malicious peer in a
3011 Denial Of Service attack.
3013 This issue was reported to OpenSSL by Alex Gaynor.
3018 * Excessive allocation of memory in tls_get_message_header() and
3019 dtls1_preprocess_fragment()
3021 A (D)TLS message includes 3 bytes for its length in the header for the
3022 message. This would allow for messages up to 16Mb in length. Messages of
3023 this length are excessive and OpenSSL includes a check to ensure that a
3024 peer is sending reasonably sized messages in order to avoid too much memory
3025 being consumed to service a connection. A flaw in the logic of version
3026 1.1.0 means that memory for the message is allocated too early, prior to
3027 the excessive message length check. Due to way memory is allocated in
3028 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3029 to service a connection. This could lead to a Denial of Service through
3030 memory exhaustion. However, the excessive message length check still takes
3031 place, and this would cause the connection to immediately fail. Assuming
3032 that the application calls SSL_free() on the failed connection in a timely
3033 manner then the 21Mb of allocated memory will then be immediately freed
3034 again. Therefore the excessive memory allocation will be transitory in
3035 nature. This then means that there is only a security impact if:
3037 1) The application does not call SSL_free() in a timely manner in the event
3038 that the connection fails
3040 2) The application is working in a constrained environment where there is
3041 very little free memory
3043 3) The attacker initiates multiple connection attempts such that there are
3044 multiple connections in a state where memory has been allocated for the
3045 connection; SSL_free() has not yet been called; and there is insufficient
3046 memory to service the multiple requests.
3048 Except in the instance of (1) above any Denial Of Service is likely to be
3049 transitory because as soon as the connection fails the memory is
3050 subsequently freed again in the SSL_free() call. However there is an
3051 increased risk during this period of application crashes due to the lack of
3052 memory - which would then mean a more serious Denial of Service.
3054 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3055 (CVE-2016-6307 and CVE-2016-6308)
3059 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3060 had to be removed. Primary reason is that vendor assembler can't
3061 assemble our modules with -KPIC flag. As result it, assembly
3062 support, was not even available as option. But its lack means
3063 lack of side-channel resistant code, which is incompatible with
3064 security by todays standards. Fortunately gcc is readily available
3065 prepackaged option, which we firmly point at...
3069 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3071 * Windows command-line tool supports UTF-8 opt-in option for arguments
3072 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3073 (to any value) allows Windows user to access PKCS#12 file generated
3074 with Windows CryptoAPI and protected with non-ASCII password, as well
3075 as files generated under UTF-8 locale on Linux also protected with
3080 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3081 have been disabled by default and removed from DEFAULT, just like RC4.
3082 See the RC4 item below to re-enable both.
3086 * The method for finding the storage location for the Windows RAND seed file
3087 has changed. First we check %RANDFILE%. If that is not set then we check
3088 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3089 all else fails we fall back to C:\.
3093 * The EVP_EncryptUpdate() function has had its return type changed from void
3094 to int. A return of 0 indicates and error while a return of 1 indicates
3099 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3100 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3101 off the constant time implementation for RSA, DSA and DH have been made
3102 no-ops and deprecated.
3106 * Windows RAND implementation was simplified to only get entropy by
3107 calling CryptGenRandom(). Various other RAND-related tickets
3110 *Joseph Wylie Yandle, Rich Salz*
3112 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3113 and `OPENSSL_LH_`, respectively. The old names are available
3114 with API compatibility. They new names are now completely documented.
3118 * Unify TYPE_up_ref(obj) methods signature.
3119 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3120 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3121 int (instead of void) like all others TYPE_up_ref() methods.
3122 So now these methods also check the return value of CRYPTO_atomic_add(),
3123 and the validity of object reference counter.
3125 *fdasilvayy@gmail.com*
3127 * With Windows Visual Studio builds, the .pdb files are installed
3128 alongside the installed libraries and executables. For a static
3129 library installation, ossl_static.pdb is the associate compiler
3130 generated .pdb file to be used when linking programs.
3134 * Remove openssl.spec. Packaging files belong with the packagers.
3138 * Automatic Darwin/OSX configuration has had a refresh, it will now
3139 recognise x86_64 architectures automatically. You can still decide
3140 to build for a different bitness with the environment variable
3141 KERNEL_BITS (can be 32 or 64), for example:
3143 KERNEL_BITS=32 ./config
3147 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3148 256 bit AES and HMAC with SHA256.
3152 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3156 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3160 * To enable users to have their own config files and build file templates,
3161 Configure looks in the directory indicated by the environment variable
3162 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3163 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3164 name and is used as is.
3168 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3169 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3170 X509_CERT_FILE_CTX was removed.
3174 * "shared" builds are now the default. To create only static libraries use
3175 the "no-shared" Configure option.
3179 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3180 All of these option have not worked for some while and are fundamental
3185 * Make various cleanup routines no-ops and mark them as deprecated. Most
3186 global cleanup functions are no longer required because they are handled
3187 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3188 Explicitly de-initing can cause problems (e.g. where a library that uses
3189 OpenSSL de-inits, but an application is still using it). The affected
3190 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3191 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3192 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3193 COMP_zlib_cleanup().
3197 * --strict-warnings no longer enables runtime debugging options
3198 such as REF_DEBUG. Instead, debug options are automatically
3199 enabled with '--debug' builds.
3201 *Andy Polyakov, Emilia Käsper*
3203 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3204 have been moved out of the public header files. New functions for managing
3205 these have been added.
3209 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3210 objects have been moved out of the public header files. New
3211 functions for managing these have been added.
3215 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3216 have been moved out of the public header files. New functions for managing
3217 these have been added.
3221 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3222 moved out of the public header files. New functions for managing these
3227 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3231 * Removed the mk1mf build scripts.
3235 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3236 it is always safe to #include a header now.
3240 * Removed the aged BC-32 config and all its supporting scripts
3244 * Removed support for Ultrix, Netware, and OS/2.
3248 * Add support for HKDF.
3250 *Alessandro Ghedini*
3252 * Add support for blake2b and blake2s
3256 * Added support for "pipelining". Ciphers that have the
3257 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3258 encryptions/decryptions simultaneously. There are currently no built-in
3259 ciphers with this property but the expectation is that engines will be able
3260 to offer it to significantly improve throughput. Support has been extended
3261 into libssl so that multiple records for a single connection can be
3262 processed in one go (for >=TLS 1.1).
3266 * Added the AFALG engine. This is an async capable engine which is able to
3267 offload work to the Linux kernel. In this initial version it only supports
3268 AES128-CBC. The kernel must be version 4.1.0 or greater.
3272 * OpenSSL now uses a new threading API. It is no longer necessary to
3273 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3274 are two supported threading models: pthreads and windows threads. It is
3275 also possible to configure OpenSSL at compile time for "no-threads". The
3276 old threading API should no longer be used. The functions have been
3277 replaced with "no-op" compatibility macros.
3279 *Alessandro Ghedini, Matt Caswell*
3281 * Modify behavior of ALPN to invoke callback after SNI/servername
3282 callback, such that updates to the SSL_CTX affect ALPN.
3286 * Add SSL_CIPHER queries for authentication and key-exchange.
3290 * Changes to the DEFAULT cipherlist:
3291 - Prefer (EC)DHE handshakes over plain RSA.
3292 - Prefer AEAD ciphers over legacy ciphers.
3293 - Prefer ECDSA over RSA when both certificates are available.
3294 - Prefer TLSv1.2 ciphers/PRF.
3295 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3300 * Change the ECC default curve list to be this, in order: x25519,
3301 secp256r1, secp521r1, secp384r1.
3305 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3306 disabled by default. They can be re-enabled using the
3307 enable-weak-ssl-ciphers option to Configure.
3311 * If the server has ALPN configured, but supports no protocols that the
3312 client advertises, send a fatal "no_application_protocol" alert.
3313 This behaviour is SHALL in RFC 7301, though it isn't universally
3314 implemented by other servers.
3318 * Add X25519 support.
3319 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3320 for public and private key encoding using the format documented in
3321 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3322 key generation and key derivation.
3324 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3329 * Deprecate SRP_VBASE_get_by_user.
3330 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3331 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3332 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3333 seed, even if the seed is configured.
3335 Users should use SRP_VBASE_get1_by_user instead. Note that in
3336 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3337 also that even though configuring the SRP seed attempts to hide
3338 invalid usernames by continuing the handshake with fake
3339 credentials, this behaviour is not constant time and no strong
3340 guarantees are made that the handshake is indistinguishable from
3341 that of a valid user.
3345 * Configuration change; it's now possible to build dynamic engines
3346 without having to build shared libraries and vice versa. This
3347 only applies to the engines in `engines/`, those in `crypto/engine/`
3348 will always be built into libcrypto (i.e. "static").
3350 Building dynamic engines is enabled by default; to disable, use
3351 the configuration option "disable-dynamic-engine".
3353 The only requirements for building dynamic engines are the
3354 presence of the DSO module and building with position independent
3355 code, so they will also automatically be disabled if configuring
3356 with "disable-dso" or "disable-pic".
3358 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3359 are also taken away from openssl/opensslconf.h, as they are
3364 * Configuration change; if there is a known flag to compile
3365 position independent code, it will always be applied on the
3366 libcrypto and libssl object files, and never on the application
3367 object files. This means other libraries that use routines from
3368 libcrypto / libssl can be made into shared libraries regardless
3369 of how OpenSSL was configured.
3371 If this isn't desirable, the configuration options "disable-pic"
3372 or "no-pic" can be used to disable the use of PIC. This will
3373 also disable building shared libraries and dynamic engines.
3377 * Removed JPAKE code. It was experimental and has no wide use.
3381 * The INSTALL_PREFIX Makefile variable has been renamed to
3382 DESTDIR. That makes for less confusion on what this variable
3383 is for. Also, the configuration option --install_prefix is
3388 * Heartbeat for TLS has been removed and is disabled by default
3389 for DTLS; configure with enable-heartbeats. Code that uses the
3390 old #define's might need to be updated.
3392 *Emilia Käsper, Rich Salz*
3394 * Rename REF_CHECK to REF_DEBUG.
3398 * New "unified" build system
3400 The "unified" build system is aimed to be a common system for all
3401 platforms we support. With it comes new support for VMS.
3403 This system builds supports building in a different directory tree
3404 than the source tree. It produces one Makefile (for unix family
3405 or lookalikes), or one descrip.mms (for VMS).
3407 The source of information to make the Makefile / descrip.mms is
3408 small files called 'build.info', holding the necessary
3409 information for each directory with source to compile, and a
3410 template in Configurations, like unix-Makefile.tmpl or
3413 With this change, the library names were also renamed on Windows
3414 and on VMS. They now have names that are closer to the standard
3415 on Unix, and include the major version number, and in certain
3416 cases, the architecture they are built for. See "Notes on shared
3417 libraries" in INSTALL.
3419 We rely heavily on the perl module Text::Template.
3423 * Added support for auto-initialisation and de-initialisation of the library.
3424 OpenSSL no longer requires explicit init or deinit routines to be called,
3425 except in certain circumstances. See the OPENSSL_init_crypto() and
3426 OPENSSL_init_ssl() man pages for further information.
3430 * The arguments to the DTLSv1_listen function have changed. Specifically the
3431 "peer" argument is now expected to be a BIO_ADDR object.
3433 * Rewrite of BIO networking library. The BIO library lacked consistent
3434 support of IPv6, and adding it required some more extensive
3435 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3436 which hold all types of addresses and chains of address information.
3437 It also introduces a new API, with functions like BIO_socket,
3438 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3439 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3440 have been adapted accordingly.
3444 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3449 * CRIME protection: disable compression by default, even if OpenSSL is
3450 compiled with zlib enabled. Applications can still enable compression
3451 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3452 using the SSL_CONF library to configure compression.
3456 * The signature of the session callback configured with
3457 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3458 was explicitly marked as `const unsigned char*` instead of
3463 * Always DPURIFY. Remove the use of uninitialized memory in the
3464 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3468 * Removed many obsolete configuration items, including
3469 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3470 MD2_CHAR, MD2_INT, MD2_LONG
3472 IDEA_SHORT, IDEA_LONG
3473 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3475 *Rich Salz, with advice from Andy Polyakov*
3477 * Many BN internals have been moved to an internal header file.
3479 *Rich Salz with help from Andy Polyakov*
3481 * Configuration and writing out the results from it has changed.
3482 Files such as Makefile include/openssl/opensslconf.h and are now
3483 produced through general templates, such as Makefile.in and
3484 crypto/opensslconf.h.in and some help from the perl module
3487 Also, the center of configuration information is no longer
3488 Makefile. Instead, Configure produces a perl module in
3489 configdata.pm which holds most of the config data (in the hash
3490 table %config), the target data that comes from the target
3491 configuration in one of the `Configurations/*.conf` files (in
3496 * To clarify their intended purposes, the Configure options
3497 --prefix and --openssldir change their semantics, and become more
3498 straightforward and less interdependent.
3500 --prefix shall be used exclusively to give the location INSTALLTOP
3501 where programs, scripts, libraries, include files and manuals are
3502 going to be installed. The default is now /usr/local.
3504 --openssldir shall be used exclusively to give the default
3505 location OPENSSLDIR where certificates, private keys, CRLs are
3506 managed. This is also where the default openssl.cnf gets
3508 If the directory given with this option is a relative path, the
3509 values of both the --prefix value and the --openssldir value will
3510 be combined to become OPENSSLDIR.
3511 The default for --openssldir is INSTALLTOP/ssl.
3513 Anyone who uses --openssldir to specify where OpenSSL is to be
3514 installed MUST change to use --prefix instead.
3518 * The GOST engine was out of date and therefore it has been removed. An up
3519 to date GOST engine is now being maintained in an external repository.
3520 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3521 support for GOST ciphersuites (these are only activated if a GOST engine
3526 * EGD is no longer supported by default; use enable-egd when
3529 *Ben Kaduk and Rich Salz*
3531 * The distribution now has Makefile.in files, which are used to
3532 create Makefile's when Configure is run. *Configure must be run
3533 before trying to build now.*
3537 * The return value for SSL_CIPHER_description() for error conditions
3542 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3544 Obtaining and performing DNSSEC validation of TLSA records is
3545 the application's responsibility. The application provides
3546 the TLSA records of its choice to OpenSSL, and these are then
3547 used to authenticate the peer.
3549 The TLSA records need not even come from DNS. They can, for
3550 example, be used to implement local end-entity certificate or
3551 trust-anchor "pinning", where the "pin" data takes the form
3552 of TLSA records, which can augment or replace verification
3553 based on the usual WebPKI public certification authorities.
3557 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3558 continues to support deprecated interfaces in default builds.
3559 However, applications are strongly advised to compile their
3560 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3561 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3562 or the 1.1.0 releases.
3564 In environments in which all applications have been ported to
3565 not use any deprecated interfaces OpenSSL's Configure script
3566 should be used with the --api=1.1.0 option to entirely remove
3567 support for the deprecated features from the library and
3568 unconditionally disable them in the installed headers.
3569 Essentially the same effect can be achieved with the "no-deprecated"
3570 argument to Configure, except that this will always restrict
3571 the build to just the latest API, rather than a fixed API
3574 As applications are ported to future revisions of the API,
3575 they should update their compile-time OPENSSL_API_COMPAT define
3576 accordingly, but in most cases should be able to continue to
3577 compile with later releases.
3579 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3580 0x10000000L and 0x00908000L, respectively. However those
3581 versions did not support the OPENSSL_API_COMPAT feature, and
3582 so applications are not typically tested for explicit support
3583 of just the undeprecated features of either release.
3587 * Add support for setting the minimum and maximum supported protocol.
3588 It can bet set via the SSL_set_min_proto_version() and
3589 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3590 MaxProtocol. It's recommended to use the new APIs to disable
3591 protocols instead of disabling individual protocols using
3592 SSL_set_options() or SSL_CONF's Protocol. This change also
3593 removes support for disabling TLS 1.2 in the OpenSSL TLS
3594 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3598 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3602 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3603 and integrates ECDSA and ECDH functionality into EC. Implementations can
3604 now redirect key generation and no longer need to convert to or from
3607 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3608 include the ec.h header file instead.
3612 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3613 ciphers who are no longer supported and drops support the ephemeral RSA key
3614 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3618 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3619 opaque. For HMAC_CTX, the following constructors and destructors
3622 HMAC_CTX *HMAC_CTX_new(void);
3623 void HMAC_CTX_free(HMAC_CTX *ctx);
3625 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3626 destroy such methods has been added. See EVP_MD_meth_new(3) and
3627 EVP_CIPHER_meth_new(3) for documentation.
3630 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3631 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3632 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3633 an already created structure.
3634 2) For consistency with the majority of our object creators and
3635 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3636 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3637 for deprecated builds.
3641 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3642 cryptographic operations to be performed asynchronously as long as an
3643 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3644 further details. Libssl has also had this capability integrated with the
3645 introduction of the new mode SSL_MODE_ASYNC and associated error
3646 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3647 pages. This work was developed in partnership with Intel Corp.
3651 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3652 always enabled now. If you want to disable the support you should
3653 exclude it using the list of supported ciphers. This also means that the
3654 "-no_ecdhe" option has been removed from s_server.
3658 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3659 SSL_{CTX_}set1_curves() which can set a list.
3663 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3664 curve you want to support using SSL_{CTX_}set1_curves().
3668 * State machine rewrite. The state machine code has been significantly
3669 refactored in order to remove much duplication of code and solve issues
3670 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3671 further details). This change does have some associated API changes.
3672 Notably the SSL_state() function has been removed and replaced by
3673 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3674 SSL_set_state() has been removed altogether. The previous handshake states
3675 defined in ssl.h and ssl3.h have also been removed.
3679 * All instances of the string "ssleay" in the public API were replaced
3680 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3681 Some error codes related to internal RSA_eay API's were renamed.
3685 * The demo files in crypto/threads were moved to demo/threads.
3689 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3692 *Matt Caswell, Rich Salz*
3694 * New ASN.1 embed macro.
3696 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3697 structure is not allocated: it is part of the parent. That is instead of
3705 This reduces memory fragmentation and make it impossible to accidentally
3706 set a mandatory field to NULL.
3708 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3709 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3710 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3715 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3719 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3720 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3721 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3722 DES and RC4 ciphersuites.
3726 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3727 This changes the decoding behaviour for some invalid messages,
3728 though the change is mostly in the more lenient direction, and
3729 legacy behaviour is preserved as much as possible.
3733 * Fix no-stdio build.
3734 *David Woodhouse <David.Woodhouse@intel.com> and also*
3735 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3737 * New testing framework
3738 The testing framework has been largely rewritten and is now using
3739 perl and the perl modules Test::Harness and an extended variant of
3740 Test::More called OpenSSL::Test to do its work. All test scripts in
3741 test/ have been rewritten into test recipes, and all direct calls to
3742 executables in test/Makefile have become individual recipes using the
3743 simplified testing OpenSSL::Test::Simple.
3745 For documentation on our testing modules, do:
3747 perldoc test/testlib/OpenSSL/Test/Simple.pm
3748 perldoc test/testlib/OpenSSL/Test.pm
3752 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3753 are used; the latter aborts on memory leaks (usually checked on exit).
3754 Some undocumented "set malloc, etc., hooks" functions were removed
3755 and others were changed. All are now documented.
3759 * In DSA_generate_parameters_ex, if the provided seed is too short,
3762 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3764 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3765 from RFC4279, RFC4785, RFC5487, RFC5489.
3767 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3768 original RSA_PSK patch.
3772 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3773 era flag was never set throughout the codebase (only read). Also removed
3774 SSL3_FLAGS_POP_BUFFER which was only used if
3775 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3779 * Changed the default name options in the "ca", "crl", "req" and "x509"
3780 to be "oneline" instead of "compat".
3784 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3785 not aware of clients that still exhibit this bug, and the workaround
3786 hasn't been working properly for a while.
3790 * The return type of BIO_number_read() and BIO_number_written() as well as
3791 the corresponding num_read and num_write members in the BIO structure has
3792 changed from unsigned long to uint64_t. On platforms where an unsigned
3793 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3798 * Given the pervasive nature of TLS extensions it is inadvisable to run
3799 OpenSSL without support for them. It also means that maintaining
3800 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3801 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3805 * Removed support for the two export grade static DH ciphersuites
3806 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3807 were newly added (along with a number of other static DH ciphersuites) to
3808 1.0.2. However the two export ones have *never* worked since they were
3809 introduced. It seems strange in any case to be adding new export
3810 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3814 * Version negotiation has been rewritten. In particular SSLv23_method(),
3815 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3816 and turned into macros which simply call the new preferred function names
3817 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3818 should use the new names instead. Also as part of this change the ssl23.h
3819 header file has been removed.
3823 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3824 code and the associated standard is no longer considered fit-for-purpose.
3828 * RT2547 was closed. When generating a private key, try to make the
3829 output file readable only by the owner. This behavior change might
3830 be noticeable when interacting with other software.
3832 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3837 * Added HTTP GET support to the ocsp command.
3841 * Changed default digest for the dgst and enc commands from MD5 to
3846 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3850 * Added support for TLS extended master secret from
3851 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3852 initial patch which was a great help during development.
3856 * All libssl internal structures have been removed from the public header
3857 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3858 now redundant). Users should not attempt to access internal structures
3859 directly. Instead they should use the provided API functions.
3863 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3864 Access to deprecated functions can be re-enabled by running config with
3865 "enable-deprecated". In addition applications wishing to use deprecated
3866 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3867 will, by default, disable some transitive includes that previously existed
3868 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3872 * Added support for OCB mode. OpenSSL has been granted a patent license
3873 compatible with the OpenSSL license for use of OCB. Details are available
3874 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3875 for OCB can be removed by calling config with no-ocb.
3879 * SSLv2 support has been removed. It still supports receiving a SSLv2
3880 compatible client hello.
3884 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3885 done while fixing the error code for the key-too-small case.
3887 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3889 * CA.sh has been removed; use CA.pl instead.
3893 * Removed old DES API.
3897 * Remove various unsupported platforms:
3903 Sinix/ReliantUNIX RM400
3908 16-bit platforms such as WIN16
3912 * Clean up OPENSSL_NO_xxx #define's
3913 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3914 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3915 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3916 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3917 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3918 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3919 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3920 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3921 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3922 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3926 * Cleaned up dead code
3927 Remove all but one '#ifdef undef' which is to be looked at.
3931 * Clean up calling of xxx_free routines.
3932 Just like free(), fix most of the xxx_free routines to accept
3933 NULL. Remove the non-null checks from callers. Save much code.
3937 * Add secure heap for storage of private keys (when possible).
3938 Add BIO_s_secmem(), CBIGNUM, etc.
3939 Contributed by Akamai Technologies under our Corporate CLA.
3943 * Experimental support for a new, fast, unbiased prime candidate generator,
3944 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3946 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3948 * New output format NSS in the sess_id command line tool. This allows
3949 exporting the session id and the master key in NSS keylog format.
3951 *Martin Kaiser <martin@kaiser.cx>*
3953 * Harmonize version and its documentation. -f flag is used to display
3956 *mancha <mancha1@zoho.com>*
3958 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3959 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3961 *mancha <mancha1@zoho.com>*
3963 * Fix some double frees. These are not thought to be exploitable.
3965 *mancha <mancha1@zoho.com>*
3967 * A missing bounds check in the handling of the TLS heartbeat extension
3968 can be used to reveal up to 64k of memory to a connected client or
3971 Thanks for Neel Mehta of Google Security for discovering this bug and to
3972 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3973 preparing the fix ([CVE-2014-0160])
3975 *Adam Langley, Bodo Moeller*
3977 * Fix for the attack described in the paper "Recovering OpenSSL
3978 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3979 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3980 <http://eprint.iacr.org/2014/140>
3982 Thanks to Yuval Yarom and Naomi Benger for discovering this
3983 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3985 *Yuval Yarom and Naomi Benger*
3987 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3988 this fixes a limitation in previous versions of OpenSSL.
3992 * Experimental encrypt-then-mac support.
3994 Experimental support for encrypt then mac from
3995 draft-gutmann-tls-encrypt-then-mac-02.txt
3997 To enable it set the appropriate extension number (0x42 for the test
3998 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4000 For non-compliant peers (i.e. just about everything) this should have no
4003 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4007 * Add EVP support for key wrapping algorithms, to avoid problems with
4008 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4009 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4010 algorithms and include tests cases.
4014 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4019 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4020 MGF1 digest and OAEP label.
4024 * Make openssl verify return errors.
4026 *Chris Palmer <palmer@google.com> and Ben Laurie*
4028 * New function ASN1_TIME_diff to calculate the difference between two
4029 ASN1_TIME structures or one structure and the current time.
4033 * Update fips_test_suite to support multiple command line options. New
4034 test to induce all self test errors in sequence and check expected
4039 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4040 sign or verify all in one operation.
4044 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4045 test programs and fips_test_suite. Includes functionality to parse
4046 the minimal script output of fipsalgest.pl directly.
4050 * Add authorisation parameter to FIPS_module_mode_set().
4054 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4058 * Use separate DRBG fields for internal and external flags. New function
4059 FIPS_drbg_health_check() to perform on demand health checking. Add
4060 generation tests to fips_test_suite with reduced health check interval to
4061 demonstrate periodic health checking. Add "nodh" option to
4062 fips_test_suite to skip very slow DH test.
4066 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4071 * More extensive health check for DRBG checking many more failure modes.
4072 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4073 combination: call this in fips_test_suite.
4077 * Add support for canonical generation of DSA parameter 'g'. See
4080 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4081 POST to handle HMAC cases.
4085 * Add functions FIPS_module_version() and FIPS_module_version_text()
4086 to return numerical and string versions of the FIPS module number.
4090 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4091 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4092 outside the validated module in the FIPS capable OpenSSL.
4096 * Minor change to DRBG entropy callback semantics. In some cases
4097 there is no multiple of the block length between min_len and
4098 max_len. Allow the callback to return more than max_len bytes
4099 of entropy but discard any extra: it is the callback's responsibility
4100 to ensure that the extra data discarded does not impact the
4101 requested amount of entropy.
4105 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4106 information in FIPS186-3, SP800-57 and SP800-131A.
4110 * CCM support via EVP. Interface is very similar to GCM case except we
4111 must supply all data in one chunk (i.e. no update, final) and the
4112 message length must be supplied if AAD is used. Add algorithm test
4117 * Initial version of POST overhaul. Add POST callback to allow the status
4118 of POST to be monitored and/or failures induced. Modify fips_test_suite
4119 to use callback. Always run all selftests even if one fails.
4123 * XTS support including algorithm test driver in the fips_gcmtest program.
4124 Note: this does increase the maximum key length from 32 to 64 bytes but
4125 there should be no binary compatibility issues as existing applications
4126 will never use XTS mode.
4130 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4131 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4132 performs algorithm blocking for unapproved PRNG types. Also do not
4133 set PRNG type in FIPS_mode_set(): leave this to the application.
4134 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4135 the standard OpenSSL PRNG: set additional data to a date time vector.
4139 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4140 This shouldn't present any incompatibility problems because applications
4141 shouldn't be using these directly and any that are will need to rethink
4142 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4146 * Extensive self tests and health checking required by SP800-90 DRBG.
4147 Remove strength parameter from FIPS_drbg_instantiate and always
4148 instantiate at maximum supported strength.
4152 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4156 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4160 * New function DH_compute_key_padded() to compute a DH key and pad with
4161 leading zeroes if needed: this complies with SP800-56A et al.
4165 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4166 anything, incomplete, subject to change and largely untested at present.
4170 * Modify fipscanisteronly build option to only build the necessary object
4171 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4175 * Add experimental option FIPSSYMS to give all symbols in
4176 fipscanister.o and FIPS or fips prefix. This will avoid
4177 conflicts with future versions of OpenSSL. Add perl script
4178 util/fipsas.pl to preprocess assembly language source files
4179 and rename any affected symbols.
4183 * Add selftest checks and algorithm block of non-fips algorithms in
4184 FIPS mode. Remove DES2 from selftests.
4188 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4189 return internal method without any ENGINE dependencies. Add new
4190 tiny fips sign and verify functions.
4194 * New build option no-ec2m to disable characteristic 2 code.
4198 * New build option "fipscanisteronly". This only builds fipscanister.o
4199 and (currently) associated fips utilities. Uses the file Makefile.fips
4200 instead of Makefile.org as the prototype.
4204 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4205 Update fips_gcmtest to use IV generator.
4209 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4210 setting output buffer to NULL. The `*Final` function must be
4211 called although it will not retrieve any additional data. The tag
4212 can be set or retrieved with a ctrl. The IV length is by default 12
4213 bytes (96 bits) but can be set to an alternative value. If the IV
4214 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4219 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4220 underlying do_cipher function handles all cipher semantics itself
4221 including padding and finalisation. This is useful if (for example)
4222 an ENGINE cipher handles block padding itself. The behaviour of
4223 do_cipher is subtly changed if this flag is set: the return value
4224 is the number of characters written to the output buffer (zero is
4225 no longer an error code) or a negative error code. Also if the
4226 input buffer is NULL and length 0 finalisation should be performed.
4230 * If a candidate issuer certificate is already part of the constructed
4231 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4235 * Improve forward-security support: add functions
4237 void SSL_CTX_set_not_resumable_session_callback(
4238 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4239 void SSL_set_not_resumable_session_callback(
4240 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4242 for use by SSL/TLS servers; the callback function will be called whenever a
4243 new session is created, and gets to decide whether the session may be
4244 cached to make it resumable (return 0) or not (return 1). (As by the
4245 SSL/TLS protocol specifications, the session_id sent by the server will be
4246 empty to indicate that the session is not resumable; also, the server will
4247 not generate RFC 4507 (RFC 5077) session tickets.)
4249 A simple reasonable callback implementation is to return is_forward_secure.
4250 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4251 by the SSL/TLS server library, indicating whether it can provide forward
4254 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4256 * New -verify_name option in command line utilities to set verification
4261 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4262 Add CMAC pkey methods.
4266 * Experimental renegotiation in s_server -www mode. If the client
4267 browses /reneg connection is renegotiated. If /renegcert it is
4268 renegotiated requesting a certificate.
4272 * Add an "external" session cache for debugging purposes to s_server. This
4273 should help trace issues which normally are only apparent in deployed
4274 multi-process servers.
4278 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4279 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4280 BIO_set_cipher() and some obscure PEM functions were changed so they
4281 can now return an error. The RAND changes required a change to the
4282 RAND_METHOD structure.
4286 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4287 a gcc attribute to warn if the result of a function is ignored. This
4288 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4289 whose return value is often ignored.
4293 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4294 These allow SCTs (signed certificate timestamps) to be requested and
4295 validated when establishing a connection.
4297 *Rob Percival <robpercival@google.com>*
4302 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4304 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4305 used even when parsing explicit parameters, when loading a encoded key
4306 or calling `EC_GROUP_new_from_ecpkparameters()`/
4307 `EC_GROUP_new_from_ecparameters()`.
4308 This prevents bypass of security hardening and performance gains,
4309 especially for curves with specialized EC_METHODs.
4310 By default, if a key encoded with explicit parameters is loaded and later
4311 encoded, the output is still encoded with explicit parameters, even if
4312 internally a "named" EC_GROUP is used for computation.
4316 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4317 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4318 NULL. After this change, only the cofactor parameter can be NULL. It also
4319 does some minimal sanity checks on the passed order.
4324 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4325 An attack is simple, if the first CMS_recipientInfo is valid but the
4326 second CMS_recipientInfo is chosen ciphertext. If the second
4327 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4328 encryption key will be replaced by garbage, and the message cannot be
4329 decoded, but if the RSA decryption fails, the correct encryption key is
4330 used and the recipient will not notice the attack.
4331 As a work around for this potential attack the length of the decrypted
4332 key must be equal to the cipher default key length, in case the
4333 certifiate is not given and all recipientInfo are tried out.
4334 The old behaviour can be re-enabled in the CMS code by setting the
4335 CMS_DEBUG_DECRYPT flag.
4340 * Document issue with installation paths in diverse Windows builds
4342 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4343 binaries and run-time config file.
4348 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4350 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4351 This changes the size when using the `genpkey` command when no size is given.
4352 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4353 generation commands to use 2048 bits by default.
4357 * Add FIPS support for Android Arm 64-bit
4359 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4360 Module in Version 2.0.10. For some reason, the corresponding target
4361 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4362 built with FIPS support on Android Arm 64-bit. This omission has been
4365 *Matthias St. Pierre*
4367 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4369 * 0-byte record padding oracle
4371 If an application encounters a fatal protocol error and then calls
4372 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4373 then OpenSSL can respond differently to the calling application if a 0 byte
4374 record is received with invalid padding compared to if a 0 byte record is
4375 received with an invalid MAC. If the application then behaves differently
4376 based on that in a way that is detectable to the remote peer, then this
4377 amounts to a padding oracle that could be used to decrypt data.
4379 In order for this to be exploitable "non-stitched" ciphersuites must be in
4380 use. Stitched ciphersuites are optimised implementations of certain
4381 commonly used ciphersuites. Also the application must call SSL_shutdown()
4382 twice even if a protocol error has occurred (applications should not do
4383 this but some do anyway).
4385 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4386 Aviram, with additional investigation by Steven Collison and Andrew
4387 Hourselt. It was reported to OpenSSL on 10th December 2018.
4392 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4396 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4398 * Microarchitecture timing vulnerability in ECC scalar multiplication
4400 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4401 shown to be vulnerable to a microarchitecture timing side channel attack.
4402 An attacker with sufficient access to mount local timing attacks during
4403 ECDSA signature generation could recover the private key.
4405 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4406 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4412 * Timing vulnerability in DSA signature generation
4414 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4415 timing side channel attack. An attacker could use variations in the signing
4416 algorithm to recover the private key.
4418 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4423 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4424 Module, accidentally introduced while backporting security fixes from the
4425 development branch and hindering the use of ECC in FIPS mode.
4429 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4431 * Client DoS due to large DH parameter
4433 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4434 malicious server can send a very large prime value to the client. This will
4435 cause the client to spend an unreasonably long period of time generating a
4436 key for this prime resulting in a hang until the client has finished. This
4437 could be exploited in a Denial Of Service attack.
4439 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4444 * Cache timing vulnerability in RSA Key Generation
4446 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4447 a cache timing side channel attack. An attacker with sufficient access to
4448 mount cache timing attacks during the RSA key generation process could
4449 recover the private key.
4451 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4452 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4457 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4458 parameter is no longer accepted, as it leads to a corrupt table. NULL
4459 pem_str is reserved for alias entries only.
4463 * Revert blinding in ECDSA sign and instead make problematic addition
4464 length-invariant. Switch even to fixed-length Montgomery multiplication.
4468 * Change generating and checking of primes so that the error rate of not
4469 being prime depends on the intended use based on the size of the input.
4470 For larger primes this will result in more rounds of Miller-Rabin.
4471 The maximal error rate for primes with more than 1080 bits is lowered
4474 *Kurt Roeckx, Annie Yousar*
4476 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4480 * Add blinding to ECDSA and DSA signatures to protect against side channel
4481 attacks discovered by Keegan Ryan (NCC Group).
4485 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4486 now allow empty (zero character) pass phrases.
4490 * Certificate time validation (X509_cmp_time) enforces stricter
4491 compliance with RFC 5280. Fractional seconds and timezone offsets
4492 are no longer allowed.
4496 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4498 * Constructed ASN.1 types with a recursive definition could exceed the stack
4500 Constructed ASN.1 types with a recursive definition (such as can be found
4501 in PKCS7) could eventually exceed the stack given malicious input with
4502 excessive recursion. This could result in a Denial Of Service attack. There
4503 are no such structures used within SSL/TLS that come from untrusted sources
4504 so this is considered safe.
4506 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4512 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4514 * Read/write after SSL object in error state
4516 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4517 mechanism. The intent was that if a fatal error occurred during a handshake
4518 then OpenSSL would move into the error state and would immediately fail if
4519 you attempted to continue the handshake. This works as designed for the
4520 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4521 SSL_connect()), however due to a bug it does not work correctly if
4522 SSL_read() or SSL_write() is called directly. In that scenario, if the
4523 handshake fails then a fatal error will be returned in the initial function
4524 call. If SSL_read()/SSL_write() is subsequently called by the application
4525 for the same SSL object then it will succeed and the data is passed without
4526 being decrypted/encrypted directly from the SSL/TLS record layer.
4528 In order to exploit this issue an application bug would have to be present
4529 that resulted in a call to SSL_read()/SSL_write() being issued after having
4530 already received a fatal error.
4532 This issue was reported to OpenSSL by David Benjamin (Google).
4537 * rsaz_1024_mul_avx2 overflow bug on x86_64
4539 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4540 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4541 Analysis suggests that attacks against RSA and DSA as a result of this
4542 defect would be very difficult to perform and are not believed likely.
4543 Attacks against DH1024 are considered just feasible, because most of the
4544 work necessary to deduce information about a private key may be performed
4545 offline. The amount of resources required for such an attack would be
4546 significant. However, for an attack on TLS to be meaningful, the server
4547 would have to share the DH1024 private key among multiple clients, which is
4548 no longer an option since CVE-2016-0701.
4550 This only affects processors that support the AVX2 but not ADX extensions
4551 like Intel Haswell (4th generation).
4553 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4554 was originally found via the OSS-Fuzz project.
4559 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4561 * bn_sqrx8x_internal carry bug on x86_64
4563 There is a carry propagating bug in the x86_64 Montgomery squaring
4564 procedure. No EC algorithms are affected. Analysis suggests that attacks
4565 against RSA and DSA as a result of this defect would be very difficult to
4566 perform and are not believed likely. Attacks against DH are considered just
4567 feasible (although very difficult) because most of the work necessary to
4568 deduce information about a private key may be performed offline. The amount
4569 of resources required for such an attack would be very significant and
4570 likely only accessible to a limited number of attackers. An attacker would
4571 additionally need online access to an unpatched system using the target
4572 private key in a scenario with persistent DH parameters and a private
4573 key that is shared between multiple clients.
4575 This only affects processors that support the BMI1, BMI2 and ADX extensions
4576 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4578 This issue was reported to OpenSSL by the OSS-Fuzz project.
4583 * Malformed X.509 IPAddressFamily could cause OOB read
4585 If an X.509 certificate has a malformed IPAddressFamily extension,
4586 OpenSSL could do a one-byte buffer overread. The most likely result
4587 would be an erroneous display of the certificate in text format.
4589 This issue was reported to OpenSSL by the OSS-Fuzz project.
4594 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4596 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4597 platform rather than 'mingw'.
4601 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4603 * Truncated packet could crash via OOB read
4605 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4606 cipher is being used, then a truncated packet can cause that host to
4607 perform an out-of-bounds read, usually resulting in a crash.
4609 This issue was reported to OpenSSL by Robert Święcki of Google.
4614 * BN_mod_exp may produce incorrect results on x86_64
4616 There is a carry propagating bug in the x86_64 Montgomery squaring
4617 procedure. No EC algorithms are affected. Analysis suggests that attacks
4618 against RSA and DSA as a result of this defect would be very difficult to
4619 perform and are not believed likely. Attacks against DH are considered just
4620 feasible (although very difficult) because most of the work necessary to
4621 deduce information about a private key may be performed offline. The amount
4622 of resources required for such an attack would be very significant and
4623 likely only accessible to a limited number of attackers. An attacker would
4624 additionally need online access to an unpatched system using the target
4625 private key in a scenario with persistent DH parameters and a private
4626 key that is shared between multiple clients. For example this can occur by
4627 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4628 similar to CVE-2015-3193 but must be treated as a separate problem.
4630 This issue was reported to OpenSSL by the OSS-Fuzz project.
4635 * Montgomery multiplication may produce incorrect results
4637 There is a carry propagating bug in the Broadwell-specific Montgomery
4638 multiplication procedure that handles input lengths divisible by, but
4639 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4640 and DH private keys are impossible. This is because the subroutine in
4641 question is not used in operations with the private key itself and an input
4642 of the attacker's direct choice. Otherwise the bug can manifest itself as
4643 transient authentication and key negotiation failures or reproducible
4644 erroneous outcome of public-key operations with specially crafted input.
4645 Among EC algorithms only Brainpool P-512 curves are affected and one
4646 presumably can attack ECDH key negotiation. Impact was not analyzed in
4647 detail, because pre-requisites for attack are considered unlikely. Namely
4648 multiple clients have to choose the curve in question and the server has to
4649 share the private key among them, neither of which is default behaviour.
4650 Even then only clients that chose the curve will be affected.
4652 This issue was publicly reported as transient failures and was not
4653 initially recognized as a security issue. Thanks to Richard Morgan for
4654 providing reproducible case.
4659 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4660 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4661 prevent issues where no progress is being made and the peer continually
4662 sends unrecognised record types, using up resources processing them.
4666 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4668 * Missing CRL sanity check
4670 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4671 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4672 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4674 This issue only affects the OpenSSL 1.0.2i
4679 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4681 * OCSP Status Request extension unbounded memory growth
4683 A malicious client can send an excessively large OCSP Status Request
4684 extension. If that client continually requests renegotiation, sending a
4685 large OCSP Status Request extension each time, then there will be unbounded
4686 memory growth on the server. This will eventually lead to a Denial Of
4687 Service attack through memory exhaustion. Servers with a default
4688 configuration are vulnerable even if they do not support OCSP. Builds using
4689 the "no-ocsp" build time option are not affected.
4691 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4696 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4699 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4705 * OOB write in MDC2_Update()
4707 An overflow can occur in MDC2_Update() either if called directly or
4708 through the EVP_DigestUpdate() function using MDC2. If an attacker
4709 is able to supply very large amounts of input data after a previous
4710 call to EVP_EncryptUpdate() with a partial block then a length check
4711 can overflow resulting in a heap corruption.
4713 The amount of data needed is comparable to SIZE_MAX which is impractical
4716 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4721 * Malformed SHA512 ticket DoS
4723 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4724 DoS attack where a malformed ticket will result in an OOB read which will
4727 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4728 a custom server callback and ticket lookup mechanism.
4730 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4735 * OOB write in BN_bn2dec()
4737 The function BN_bn2dec() does not check the return value of BN_div_word().
4738 This can cause an OOB write if an application uses this function with an
4739 overly large BIGNUM. This could be a problem if an overly large certificate
4740 or CRL is printed out from an untrusted source. TLS is not affected because
4741 record limits will reject an oversized certificate before it is parsed.
4743 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4748 * OOB read in TS_OBJ_print_bio()
4750 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4751 the total length the OID text representation would use and not the amount
4752 of data written. This will result in OOB reads when large OIDs are
4755 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4760 * Pointer arithmetic undefined behaviour
4762 Avoid some undefined pointer arithmetic
4764 A common idiom in the codebase is to check limits in the following manner:
4767 Where "p" points to some malloc'd data of SIZE bytes and
4770 "len" here could be from some externally supplied data (e.g. from a TLS
4773 The rules of C pointer arithmetic are such that "p + len" is only well
4774 defined where len <= SIZE. Therefore the above idiom is actually
4775 undefined behaviour.
4777 For example this could cause problems if some malloc implementation
4778 provides an address for "p" such that "p + len" actually overflows for
4779 values of len that are too big and therefore p + len < limit.
4781 This issue was reported to OpenSSL by Guido Vranken
4786 * Constant time flag not preserved in DSA signing
4788 Operations in the DSA signing algorithm should run in constant time in
4789 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4790 implementation means that a non-constant time codepath is followed for
4791 certain operations. This has been demonstrated through a cache-timing
4792 attack to be sufficient for an attacker to recover the private DSA key.
4794 This issue was reported by César Pereida (Aalto University), Billy Brumley
4795 (Tampere University of Technology), and Yuval Yarom (The University of
4796 Adelaide and NICTA).
4801 * DTLS buffered message DoS
4803 In a DTLS connection where handshake messages are delivered out-of-order
4804 those messages that OpenSSL is not yet ready to process will be buffered
4805 for later use. Under certain circumstances, a flaw in the logic means that
4806 those messages do not get removed from the buffer even though the handshake
4807 has been completed. An attacker could force up to approx. 15 messages to
4808 remain in the buffer when they are no longer required. These messages will
4809 be cleared when the DTLS connection is closed. The default maximum size for
4810 a message is 100k. Therefore the attacker could force an additional 1500k
4811 to be consumed per connection. By opening many simulataneous connections an
4812 attacker could cause a DoS attack through memory exhaustion.
4814 This issue was reported to OpenSSL by Quan Luo.
4819 * DTLS replay protection DoS
4821 A flaw in the DTLS replay attack protection mechanism means that records
4822 that arrive for future epochs update the replay protection "window" before
4823 the MAC for the record has been validated. This could be exploited by an
4824 attacker by sending a record for the next epoch (which does not have to
4825 decrypt or have a valid MAC), with a very large sequence number. This means
4826 that all subsequent legitimate packets are dropped causing a denial of
4827 service for a specific DTLS connection.
4829 This issue was reported to OpenSSL by the OCAP audit team.
4834 * Certificate message OOB reads
4836 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4837 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4838 theoretical DoS risk but this has not been observed in practice on common
4841 The messages affected are client certificate, client certificate request
4842 and server certificate. As a result the attack can only be performed
4843 against a client or a server which enables client authentication.
4845 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4850 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4852 * Prevent padding oracle in AES-NI CBC MAC check
4854 A MITM attacker can use a padding oracle attack to decrypt traffic
4855 when the connection uses an AES CBC cipher and the server support
4858 This issue was introduced as part of the fix for Lucky 13 padding
4859 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4860 constant time by making sure that always the same bytes are read and
4861 compared against either the MAC or padding bytes. But it no longer
4862 checked that there was enough data to have both the MAC and padding
4865 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4870 * Fix EVP_EncodeUpdate overflow
4872 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4873 Base64 encoding of binary data. If an attacker is able to supply very large
4874 amounts of input data then a length check can overflow resulting in a heap
4877 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4878 the `PEM_write_bio*` family of functions. These are mainly used within the
4879 OpenSSL command line applications, so any application which processes data
4880 from an untrusted source and outputs it as a PEM file should be considered
4881 vulnerable to this issue. User applications that call these APIs directly
4882 with large amounts of untrusted data may also be vulnerable.
4884 This issue was reported by Guido Vranken.
4889 * Fix EVP_EncryptUpdate overflow
4891 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4892 is able to supply very large amounts of input data after a previous call to
4893 EVP_EncryptUpdate() with a partial block then a length check can overflow
4894 resulting in a heap corruption. Following an analysis of all OpenSSL
4895 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4896 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4897 the first called function after an EVP_EncryptInit(), and therefore that
4898 specific call must be safe. The second form is where the length passed to
4899 EVP_EncryptUpdate() can be seen from the code to be some small value and
4900 therefore there is no possibility of an overflow. Since all instances are
4901 one of these two forms, it is believed that there can be no overflows in
4902 internal code due to this problem. It should be noted that
4903 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4904 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4905 of these calls have also been analysed too and it is believed there are no
4906 instances in internal usage where an overflow could occur.
4908 This issue was reported by Guido Vranken.
4913 * Prevent ASN.1 BIO excessive memory allocation
4915 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4916 a short invalid encoding can cause allocation of large amounts of memory
4917 potentially consuming excessive resources or exhausting memory.
4919 Any application parsing untrusted data through d2i BIO functions is
4920 affected. The memory based functions such as d2i_X509() are *not* affected.
4921 Since the memory based functions are used by the TLS library, TLS
4922 applications are not affected.
4924 This issue was reported by Brian Carpenter.
4931 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4932 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4933 in arbitrary stack data being returned in the buffer.
4935 This issue was reported by Guido Vranken.
4940 * Modify behavior of ALPN to invoke callback after SNI/servername
4941 callback, such that updates to the SSL_CTX affect ALPN.
4945 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4950 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4951 methods are enabled and ssl2 is disabled the methods return NULL.
4955 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4957 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4958 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4959 provide any "EXPORT" or "LOW" strength ciphers.
4963 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4964 is by default disabled at build-time. Builds that are not configured with
4965 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4966 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4967 will need to explicitly call either of:
4969 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4971 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4973 as appropriate. Even if either of those is used, or the application
4974 explicitly uses the version-specific SSLv2_method() or its client and
4975 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4976 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4977 ciphers, and SSLv2 56-bit DES are no longer available.
4982 * Fix a double-free in DSA code
4984 A double free bug was discovered when OpenSSL parses malformed DSA private
4985 keys and could lead to a DoS attack or memory corruption for applications
4986 that receive DSA private keys from untrusted sources. This scenario is
4989 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4995 * Disable SRP fake user seed to address a server memory leak.
4997 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4999 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5000 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5001 was changed to ignore the "fake user" SRP seed, even if the seed
5004 Users should use SRP_VBASE_get1_by_user instead. Note that in
5005 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5006 also that even though configuring the SRP seed attempts to hide
5007 invalid usernames by continuing the handshake with fake
5008 credentials, this behaviour is not constant time and no strong
5009 guarantees are made that the handshake is indistinguishable from
5010 that of a valid user.
5015 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5017 In the BN_hex2bn function the number of hex digits is calculated using an
5018 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5019 large values of `i` this can result in `bn_expand` not allocating any
5020 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5021 field as NULL leading to a subsequent NULL ptr deref. For very large values
5022 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5023 In this case memory is allocated to the internal BIGNUM data field, but it
5024 is insufficiently sized leading to heap corruption. A similar issue exists
5025 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5026 is ever called by user applications with very large untrusted hex/dec data.
5027 This is anticipated to be a rare occurrence.
5029 All OpenSSL internal usage of these functions use data that is not expected
5030 to be untrusted, e.g. config file data or application command line
5031 arguments. If user developed applications generate config file data based
5032 on untrusted data then it is possible that this could also lead to security
5033 consequences. This is also anticipated to be rare.
5035 This issue was reported to OpenSSL by Guido Vranken.
5040 * Fix memory issues in `BIO_*printf` functions
5042 The internal `fmtstr` function used in processing a "%s" format string in
5043 the `BIO_*printf` functions could overflow while calculating the length of a
5044 string and cause an OOB read when printing very long strings.
5046 Additionally the internal `doapr_outch` function can attempt to write to an
5047 OOB memory location (at an offset from the NULL pointer) in the event of a
5048 memory allocation failure. In 1.0.2 and below this could be caused where
5049 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5050 could be in processing a very long "%s" format string. Memory leaks can
5053 The first issue may mask the second issue dependent on compiler behaviour.
5054 These problems could enable attacks where large amounts of untrusted data
5055 is passed to the `BIO_*printf` functions. If applications use these functions
5056 in this way then they could be vulnerable. OpenSSL itself uses these
5057 functions when printing out human-readable dumps of ASN.1 data. Therefore
5058 applications that print this data could be vulnerable if the data is from
5059 untrusted sources. OpenSSL command line applications could also be
5060 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5061 as command line arguments.
5063 Libssl is not considered directly vulnerable. Additionally certificates etc
5064 received via remote connections via libssl are also unlikely to be able to
5065 trigger these issues because of message size limits enforced within libssl.
5067 This issue was reported to OpenSSL Guido Vranken.
5072 * Side channel attack on modular exponentiation
5074 A side-channel attack was found which makes use of cache-bank conflicts on
5075 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5076 of RSA keys. The ability to exploit this issue is limited as it relies on
5077 an attacker who has control of code in a thread running on the same
5078 hyper-threaded core as the victim thread which is performing decryptions.
5080 This issue was reported to OpenSSL by Yuval Yarom, The University of
5081 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5082 Nadia Heninger, University of Pennsylvania with more information at
5083 <http://cachebleed.info>.
5088 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5089 if no keysize is specified with default_bits. This fixes an
5090 omission in an earlier change that changed all RSA/DSA key generation
5091 commands to use 2048 bits by default.
5095 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5097 * DH small subgroups
5099 Historically OpenSSL only ever generated DH parameters based on "safe"
5100 primes. More recently (in version 1.0.2) support was provided for
5101 generating X9.42 style parameter files such as those required for RFC 5114
5102 support. The primes used in such files may not be "safe". Where an
5103 application is using DH configured with parameters based on primes that are
5104 not "safe" then an attacker could use this fact to find a peer's private
5105 DH exponent. This attack requires that the attacker complete multiple
5106 handshakes in which the peer uses the same private DH exponent. For example
5107 this could be used to discover a TLS server's private DH exponent if it's
5108 reusing the private DH exponent or it's using a static DH ciphersuite.
5110 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5111 TLS. It is not on by default. If the option is not set then the server
5112 reuses the same private DH exponent for the life of the server process and
5113 would be vulnerable to this attack. It is believed that many popular
5114 applications do set this option and would therefore not be at risk.
5116 The fix for this issue adds an additional check where a "q" parameter is
5117 available (as is the case in X9.42 based parameters). This detects the
5118 only known attack, and is the only possible defense for static DH
5119 ciphersuites. This could have some performance impact.
5121 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5122 default and cannot be disabled. This could have some performance impact.
5124 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5129 * SSLv2 doesn't block disabled ciphers
5131 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5132 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5133 been disabled, provided that the SSLv2 protocol was not also disabled via
5136 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5137 and Sebastian Schinzel.
5142 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5144 * BN_mod_exp may produce incorrect results on x86_64
5146 There is a carry propagating bug in the x86_64 Montgomery squaring
5147 procedure. No EC algorithms are affected. Analysis suggests that attacks
5148 against RSA and DSA as a result of this defect would be very difficult to
5149 perform and are not believed likely. Attacks against DH are considered just
5150 feasible (although very difficult) because most of the work necessary to
5151 deduce information about a private key may be performed offline. The amount
5152 of resources required for such an attack would be very significant and
5153 likely only accessible to a limited number of attackers. An attacker would
5154 additionally need online access to an unpatched system using the target
5155 private key in a scenario with persistent DH parameters and a private
5156 key that is shared between multiple clients. For example this can occur by
5157 default in OpenSSL DHE based SSL/TLS ciphersuites.
5159 This issue was reported to OpenSSL by Hanno Böck.
5164 * Certificate verify crash with missing PSS parameter
5166 The signature verification routines will crash with a NULL pointer
5167 dereference if presented with an ASN.1 signature using the RSA PSS
5168 algorithm and absent mask generation function parameter. Since these
5169 routines are used to verify certificate signature algorithms this can be
5170 used to crash any certificate verification operation and exploited in a
5171 DoS attack. Any application which performs certificate verification is
5172 vulnerable including OpenSSL clients and servers which enable client
5175 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5180 * X509_ATTRIBUTE memory leak
5182 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5183 memory. This structure is used by the PKCS#7 and CMS routines so any
5184 application which reads PKCS#7 or CMS data from untrusted sources is
5185 affected. SSL/TLS is not affected.
5187 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5193 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5194 This changes the decoding behaviour for some invalid messages,
5195 though the change is mostly in the more lenient direction, and
5196 legacy behaviour is preserved as much as possible.
5200 * In DSA_generate_parameters_ex, if the provided seed is too short,
5203 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5205 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5207 * Alternate chains certificate forgery
5209 During certificate verification, OpenSSL will attempt to find an
5210 alternative certificate chain if the first attempt to build such a chain
5211 fails. An error in the implementation of this logic can mean that an
5212 attacker could cause certain checks on untrusted certificates to be
5213 bypassed, such as the CA flag, enabling them to use a valid leaf
5214 certificate to act as a CA and "issue" an invalid certificate.
5216 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5221 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5223 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5224 incompatibility in the handling of HMAC. The previous ABI has now been
5229 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5231 * Malformed ECParameters causes infinite loop
5233 When processing an ECParameters structure OpenSSL enters an infinite loop
5234 if the curve specified is over a specially malformed binary polynomial
5237 This can be used to perform denial of service against any
5238 system which processes public keys, certificate requests or
5239 certificates. This includes TLS clients and TLS servers with
5240 client authentication enabled.
5242 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5247 * Exploitable out-of-bounds read in X509_cmp_time
5249 X509_cmp_time does not properly check the length of the ASN1_TIME
5250 string and can read a few bytes out of bounds. In addition,
5251 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5254 An attacker can use this to craft malformed certificates and CRLs of
5255 various sizes and potentially cause a segmentation fault, resulting in
5256 a DoS on applications that verify certificates or CRLs. TLS clients
5257 that verify CRLs are affected. TLS clients and servers with client
5258 authentication enabled may be affected if they use custom verification
5261 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5262 independently by Hanno Böck.
5267 * PKCS7 crash with missing EnvelopedContent
5269 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5270 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5271 with missing content and trigger a NULL pointer dereference on parsing.
5273 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5274 structures from untrusted sources are affected. OpenSSL clients and
5275 servers are not affected.
5277 This issue was reported to OpenSSL by Michal Zalewski (Google).
5282 * CMS verify infinite loop with unknown hash function
5284 When verifying a signedData message the CMS code can enter an infinite loop
5285 if presented with an unknown hash function OID. This can be used to perform
5286 denial of service against any system which verifies signedData messages using
5288 This issue was reported to OpenSSL by Johannes Bauer.
5293 * Race condition handling NewSessionTicket
5295 If a NewSessionTicket is received by a multi-threaded client when attempting to
5296 reuse a previous ticket then a race condition can occur potentially leading to
5297 a double free of the ticket data.
5302 * Only support 256-bit or stronger elliptic curves with the
5303 'ecdh_auto' setting (server) or by default (client). Of supported
5304 curves, prefer P-256 (both).
5308 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5310 * ClientHello sigalgs DoS fix
5312 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5313 invalid signature algorithms extension a NULL pointer dereference will
5314 occur. This can be exploited in a DoS attack against the server.
5316 This issue was was reported to OpenSSL by David Ramos of Stanford
5320 *Stephen Henson and Matt Caswell*
5322 * Multiblock corrupted pointer fix
5324 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5325 feature only applies on 64 bit x86 architecture platforms that support AES
5326 NI instructions. A defect in the implementation of "multiblock" can cause
5327 OpenSSL's internal write buffer to become incorrectly set to NULL when
5328 using non-blocking IO. Typically, when the user application is using a
5329 socket BIO for writing, this will only result in a failed connection.
5330 However if some other BIO is used then it is likely that a segmentation
5331 fault will be triggered, thus enabling a potential DoS attack.
5333 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5338 * Segmentation fault in DTLSv1_listen fix
5340 The DTLSv1_listen function is intended to be stateless and processes the
5341 initial ClientHello from many peers. It is common for user code to loop
5342 over the call to DTLSv1_listen until a valid ClientHello is received with
5343 an associated cookie. A defect in the implementation of DTLSv1_listen means
5344 that state is preserved in the SSL object from one invocation to the next
5345 that can lead to a segmentation fault. Errors processing the initial
5346 ClientHello can trigger this scenario. An example of such an error could be
5347 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5350 This issue was reported to OpenSSL by Per Allansson.
5355 * Segmentation fault in ASN1_TYPE_cmp fix
5357 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5358 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5359 certificate signature algorithm consistency this can be used to crash any
5360 certificate verification operation and exploited in a DoS attack. Any
5361 application which performs certificate verification is vulnerable including
5362 OpenSSL clients and servers which enable client authentication.
5367 * Segmentation fault for invalid PSS parameters fix
5369 The signature verification routines will crash with a NULL pointer
5370 dereference if presented with an ASN.1 signature using the RSA PSS
5371 algorithm and invalid parameters. Since these routines are used to verify
5372 certificate signature algorithms this can be used to crash any
5373 certificate verification operation and exploited in a DoS attack. Any
5374 application which performs certificate verification is vulnerable including
5375 OpenSSL clients and servers which enable client authentication.
5377 This issue was was reported to OpenSSL by Brian Carpenter.
5382 * ASN.1 structure reuse memory corruption fix
5384 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5385 memory corruption via an invalid write. Such reuse is and has been
5386 strongly discouraged and is believed to be rare.
5388 Applications that parse structures containing CHOICE or ANY DEFINED BY
5389 components may be affected. Certificate parsing (d2i_X509 and related
5390 functions) are however not affected. OpenSSL clients and servers are
5396 * PKCS7 NULL pointer dereferences fix
5398 The PKCS#7 parsing code does not handle missing outer ContentInfo
5399 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5400 missing content and trigger a NULL pointer dereference on parsing.
5402 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5403 otherwise parse PKCS#7 structures from untrusted sources are
5404 affected. OpenSSL clients and servers are not affected.
5406 This issue was reported to OpenSSL by Michal Zalewski (Google).
5411 * DoS via reachable assert in SSLv2 servers fix
5413 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5414 servers that both support SSLv2 and enable export cipher suites by sending
5415 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5417 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5418 (OpenSSL development team).
5423 * Empty CKE with client auth and DHE fix
5425 If client auth is used then a server can seg fault in the event of a DHE
5426 ciphersuite being selected and a zero length ClientKeyExchange message
5427 being sent by the client. This could be exploited in a DoS attack.
5432 * Handshake with unseeded PRNG fix
5434 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5435 with an unseeded PRNG. The conditions are:
5436 - The client is on a platform where the PRNG has not been seeded
5437 automatically, and the user has not seeded manually
5438 - A protocol specific client method version has been used (i.e. not
5439 SSL_client_methodv23)
5440 - A ciphersuite is used that does not require additional random data from
5441 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5443 If the handshake succeeds then the client random that has been used will
5444 have been generated from a PRNG with insufficient entropy and therefore the
5445 output may be predictable.
5447 For example using the following command with an unseeded openssl will
5448 succeed on an unpatched platform:
5450 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5455 * Use After Free following d2i_ECPrivatekey error fix
5457 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5458 could cause a use after free condition. This, in turn, could cause a double
5459 free in several private key parsing functions (such as d2i_PrivateKey
5460 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5461 for applications that receive EC private keys from untrusted
5462 sources. This scenario is considered rare.
5464 This issue was discovered by the BoringSSL project and fixed in their
5470 * X509_to_X509_REQ NULL pointer deref fix
5472 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5473 the certificate key is invalid. This function is rarely used in practice.
5475 This issue was discovered by Brian Carpenter.
5480 * Removed the export ciphers from the DEFAULT ciphers
5484 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5486 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5487 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5488 So far those who have to target multiple platforms would compromise
5489 and argue that binary targeting say ARMv5 would still execute on
5490 ARMv8. "Universal" build resolves this compromise by providing
5491 near-optimal performance even on newer platforms.
5495 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5496 (other platforms pending).
5498 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5500 * Add support for the SignedCertificateTimestampList certificate and
5501 OCSP response extensions from RFC6962.
5505 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5506 for corner cases. (Certain input points at infinity could lead to
5507 bogus results, with non-infinity inputs mapped to infinity too.)
5511 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5512 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5513 common cases are optimized and there still is room for further
5514 improvements. Vector Permutation AES for Altivec is also added.
5518 * Add support for little-endian ppc64 Linux target.
5520 *Marcelo Cerri (IBM)*
5522 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5523 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5524 are optimized and there still is room for further improvements.
5525 Both 32- and 64-bit modes are supported.
5527 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5529 * Improved ARMv7 NEON support.
5533 * Support for SPARC Architecture 2011 crypto extensions, first
5534 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5535 SHA256/512, MD5, GHASH and modular exponentiation.
5537 *Andy Polyakov, David Miller*
5539 * Accelerated modular exponentiation for Intel processors, a.k.a.
5542 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5544 * Support for new and upcoming Intel processors, including AVX2,
5545 BMI and SHA ISA extensions. This includes additional "stitched"
5546 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5549 This work was sponsored by Intel Corp.
5553 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5554 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5555 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5559 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5560 this fixes a limitation in previous versions of OpenSSL.
5564 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5565 MGF1 digest and OAEP label.
5569 * Add EVP support for key wrapping algorithms, to avoid problems with
5570 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5571 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5572 algorithms and include tests cases.
5576 * Add functions to allocate and set the fields of an ECDSA_METHOD
5579 *Douglas E. Engert, Steve Henson*
5581 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5582 difference in days and seconds between two tm or ASN1_TIME structures.
5586 * Add -rev test option to s_server to just reverse order of characters
5587 received by client and send back to server. Also prints an abbreviated
5588 summary of the connection parameters.
5592 * New option -brief for s_client and s_server to print out a brief summary
5593 of connection parameters.
5597 * Add callbacks for arbitrary TLS extensions.
5599 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5601 * New option -crl_download in several openssl utilities to download CRLs
5602 from CRLDP extension in certificates.
5606 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5610 * New function X509_CRL_diff to generate a delta CRL from the difference
5611 of two full CRLs. Add support to "crl" utility.
5615 * New functions to set lookup_crls function and to retrieve
5616 X509_STORE from X509_STORE_CTX.
5620 * Print out deprecated issuer and subject unique ID fields in
5625 * Extend OCSP I/O functions so they can be used for simple general purpose
5626 HTTP as well as OCSP. New wrapper function which can be used to download
5627 CRLs using the OCSP API.
5631 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5635 * `SSL_CONF*` functions. These provide a common framework for application
5636 configuration using configuration files or command lines.
5640 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5641 message callback and prints the results. Needs compile time option
5642 "enable-ssl-trace". New options to s_client and s_server to enable
5647 * New ctrl and macro to retrieve supported points extensions.
5648 Print out extension in s_server and s_client.
5652 * New functions to retrieve certificate signature and signature
5657 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5662 * New Suite B modes for TLS code. These use and enforce the requirements
5663 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5664 only use Suite B curves. The Suite B modes can be set by using the
5665 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5669 * New chain verification flags for Suite B levels of security. Check
5670 algorithms are acceptable when flags are set in X509_verify_cert.
5674 * Make tls1_check_chain return a set of flags indicating checks passed
5675 by a certificate chain. Add additional tests to handle client
5676 certificates: checks for matching certificate type and issuer name
5681 * If an attempt is made to use a signature algorithm not in the peer
5682 preference list abort the handshake. If client has no suitable
5683 signature algorithms in response to a certificate request do not
5684 use the certificate.
5688 * If server EC tmp key is not in client preference list abort handshake.
5692 * Add support for certificate stores in CERT structure. This makes it
5693 possible to have different stores per SSL structure or one store in
5694 the parent SSL_CTX. Include distinct stores for certificate chain
5695 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5696 to build and store a certificate chain in CERT structure: returning
5697 an error if the chain cannot be built: this will allow applications
5698 to test if a chain is correctly configured.
5700 Note: if the CERT based stores are not set then the parent SSL_CTX
5701 store is used to retain compatibility with existing behaviour.
5705 * New function ssl_set_client_disabled to set a ciphersuite disabled
5706 mask based on the current session, check mask when sending client
5707 hello and checking the requested ciphersuite.
5711 * New ctrls to retrieve and set certificate types in a certificate
5712 request message. Print out received values in s_client. If certificate
5713 types is not set with custom values set sensible values based on
5714 supported signature algorithms.
5718 * Support for distinct client and server supported signature algorithms.
5722 * Add certificate callback. If set this is called whenever a certificate
5723 is required by client or server. An application can decide which
5724 certificate chain to present based on arbitrary criteria: for example
5725 supported signature algorithms. Add very simple example to s_server.
5726 This fixes many of the problems and restrictions of the existing client
5727 certificate callback: for example you can now clear an existing
5728 certificate and specify the whole chain.
5732 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5733 the certificate can be used for (if anything). Set valid_flags field
5734 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5735 to have similar checks in it.
5737 Add new "cert_flags" field to CERT structure and include a "strict mode".
5738 This enforces some TLS certificate requirements (such as only permitting
5739 certificate signature algorithms contained in the supported algorithms
5740 extension) which some implementations ignore: this option should be used
5741 with caution as it could cause interoperability issues.
5745 * Update and tidy signature algorithm extension processing. Work out
5746 shared signature algorithms based on preferences and peer algorithms
5747 and print them out in s_client and s_server. Abort handshake if no
5748 shared signature algorithms.
5752 * Add new functions to allow customised supported signature algorithms
5753 for SSL and SSL_CTX structures. Add options to s_client and s_server
5758 * New function SSL_certs_clear() to delete all references to certificates
5759 from an SSL structure. Before this once a certificate had been added
5760 it couldn't be removed.
5764 * Integrate hostname, email address and IP address checking with certificate
5765 verification. New verify options supporting checking in openssl utility.
5769 * Fixes and wildcard matching support to hostname and email checking
5770 functions. Add manual page.
5772 *Florian Weimer (Red Hat Product Security Team)*
5774 * New functions to check a hostname email or IP address against a
5775 certificate. Add options x509 utility to print results of checks against
5780 * Fix OCSP checking.
5782 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5784 * Initial experimental support for explicitly trusted non-root CAs.
5785 OpenSSL still tries to build a complete chain to a root but if an
5786 intermediate CA has a trust setting included that is used. The first
5787 setting is used: whether to trust (e.g., -addtrust option to the x509
5792 * Add -trusted_first option which attempts to find certificates in the
5793 trusted store even if an untrusted chain is also supplied.
5797 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5798 platform support for Linux and Android.
5802 * Support for linux-x32, ILP32 environment in x86_64 framework.
5806 * Experimental multi-implementation support for FIPS capable OpenSSL.
5807 When in FIPS mode the approved implementations are used as normal,
5808 when not in FIPS mode the internal unapproved versions are used instead.
5809 This means that the FIPS capable OpenSSL isn't forced to use the
5810 (often lower performance) FIPS implementations outside FIPS mode.
5814 * Transparently support X9.42 DH parameters when calling
5815 PEM_read_bio_DHparameters. This means existing applications can handle
5816 the new parameter format automatically.
5820 * Initial experimental support for X9.42 DH parameter format: mainly
5821 to support use of 'q' parameter for RFC5114 parameters.
5825 * Add DH parameters from RFC5114 including test data to dhtest.
5829 * Support for automatic EC temporary key parameter selection. If enabled
5830 the most preferred EC parameters are automatically used instead of
5831 hardcoded fixed parameters. Now a server just has to call:
5832 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5833 support ECDH and use the most appropriate parameters.
5837 * Enhance and tidy EC curve and point format TLS extension code. Use
5838 static structures instead of allocation if default values are used.
5839 New ctrls to set curves we wish to support and to retrieve shared curves.
5840 Print out shared curves in s_server. New options to s_server and s_client
5841 to set list of supported curves.
5845 * New ctrls to retrieve supported signature algorithms and
5846 supported curve values as an array of NIDs. Extend openssl utility
5847 to print out received values.
5851 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5852 between NIDs and the more common NIST names such as "P-256". Enhance
5853 ecparam utility and ECC method to recognise the NIST names for curves.
5857 * Enhance SSL/TLS certificate chain handling to support different
5858 chains for each certificate instead of one chain in the parent SSL_CTX.
5862 * Support for fixed DH ciphersuite client authentication: where both
5863 server and client use DH certificates with common parameters.
5867 * Support for fixed DH ciphersuites: those requiring DH server
5872 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5874 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5875 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5876 X509_CINF_get_signature were reverted post internal team review.
5881 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5883 * OCSP Status Request extension unbounded memory growth
5885 A malicious client can send an excessively large OCSP Status Request
5886 extension. If that client continually requests renegotiation, sending a
5887 large OCSP Status Request extension each time, then there will be unbounded
5888 memory growth on the server. This will eventually lead to a Denial Of
5889 Service attack through memory exhaustion. Servers with a default
5890 configuration are vulnerable even if they do not support OCSP. Builds using
5891 the "no-ocsp" build time option are not affected.
5893 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5898 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5901 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5907 * OOB write in MDC2_Update()
5909 An overflow can occur in MDC2_Update() either if called directly or
5910 through the EVP_DigestUpdate() function using MDC2. If an attacker
5911 is able to supply very large amounts of input data after a previous
5912 call to EVP_EncryptUpdate() with a partial block then a length check
5913 can overflow resulting in a heap corruption.
5915 The amount of data needed is comparable to SIZE_MAX which is impractical
5918 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5923 * Malformed SHA512 ticket DoS
5925 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5926 DoS attack where a malformed ticket will result in an OOB read which will
5929 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5930 a custom server callback and ticket lookup mechanism.
5932 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5937 * OOB write in BN_bn2dec()
5939 The function BN_bn2dec() does not check the return value of BN_div_word().
5940 This can cause an OOB write if an application uses this function with an
5941 overly large BIGNUM. This could be a problem if an overly large certificate
5942 or CRL is printed out from an untrusted source. TLS is not affected because
5943 record limits will reject an oversized certificate before it is parsed.
5945 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5950 * OOB read in TS_OBJ_print_bio()
5952 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5953 the total length the OID text representation would use and not the amount
5954 of data written. This will result in OOB reads when large OIDs are
5957 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5962 * Pointer arithmetic undefined behaviour
5964 Avoid some undefined pointer arithmetic
5966 A common idiom in the codebase is to check limits in the following manner:
5969 Where "p" points to some malloc'd data of SIZE bytes and
5972 "len" here could be from some externally supplied data (e.g. from a TLS
5975 The rules of C pointer arithmetic are such that "p + len" is only well
5976 defined where len <= SIZE. Therefore the above idiom is actually
5977 undefined behaviour.
5979 For example this could cause problems if some malloc implementation
5980 provides an address for "p" such that "p + len" actually overflows for
5981 values of len that are too big and therefore p + len < limit.
5983 This issue was reported to OpenSSL by Guido Vranken
5988 * Constant time flag not preserved in DSA signing
5990 Operations in the DSA signing algorithm should run in constant time in
5991 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5992 implementation means that a non-constant time codepath is followed for
5993 certain operations. This has been demonstrated through a cache-timing
5994 attack to be sufficient for an attacker to recover the private DSA key.
5996 This issue was reported by César Pereida (Aalto University), Billy Brumley
5997 (Tampere University of Technology), and Yuval Yarom (The University of
5998 Adelaide and NICTA).
6003 * DTLS buffered message DoS
6005 In a DTLS connection where handshake messages are delivered out-of-order
6006 those messages that OpenSSL is not yet ready to process will be buffered
6007 for later use. Under certain circumstances, a flaw in the logic means that
6008 those messages do not get removed from the buffer even though the handshake
6009 has been completed. An attacker could force up to approx. 15 messages to
6010 remain in the buffer when they are no longer required. These messages will
6011 be cleared when the DTLS connection is closed. The default maximum size for
6012 a message is 100k. Therefore the attacker could force an additional 1500k
6013 to be consumed per connection. By opening many simulataneous connections an
6014 attacker could cause a DoS attack through memory exhaustion.
6016 This issue was reported to OpenSSL by Quan Luo.
6021 * DTLS replay protection DoS
6023 A flaw in the DTLS replay attack protection mechanism means that records
6024 that arrive for future epochs update the replay protection "window" before
6025 the MAC for the record has been validated. This could be exploited by an
6026 attacker by sending a record for the next epoch (which does not have to
6027 decrypt or have a valid MAC), with a very large sequence number. This means
6028 that all subsequent legitimate packets are dropped causing a denial of
6029 service for a specific DTLS connection.
6031 This issue was reported to OpenSSL by the OCAP audit team.
6036 * Certificate message OOB reads
6038 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6039 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6040 theoretical DoS risk but this has not been observed in practice on common
6043 The messages affected are client certificate, client certificate request
6044 and server certificate. As a result the attack can only be performed
6045 against a client or a server which enables client authentication.
6047 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6052 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6054 * Prevent padding oracle in AES-NI CBC MAC check
6056 A MITM attacker can use a padding oracle attack to decrypt traffic
6057 when the connection uses an AES CBC cipher and the server support
6060 This issue was introduced as part of the fix for Lucky 13 padding
6061 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6062 constant time by making sure that always the same bytes are read and
6063 compared against either the MAC or padding bytes. But it no longer
6064 checked that there was enough data to have both the MAC and padding
6067 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6072 * Fix EVP_EncodeUpdate overflow
6074 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6075 Base64 encoding of binary data. If an attacker is able to supply very large
6076 amounts of input data then a length check can overflow resulting in a heap
6079 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6080 the `PEM_write_bio*` family of functions. These are mainly used within the
6081 OpenSSL command line applications, so any application which processes data
6082 from an untrusted source and outputs it as a PEM file should be considered
6083 vulnerable to this issue. User applications that call these APIs directly
6084 with large amounts of untrusted data may also be vulnerable.
6086 This issue was reported by Guido Vranken.
6091 * Fix EVP_EncryptUpdate overflow
6093 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6094 is able to supply very large amounts of input data after a previous call to
6095 EVP_EncryptUpdate() with a partial block then a length check can overflow
6096 resulting in a heap corruption. Following an analysis of all OpenSSL
6097 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6098 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6099 the first called function after an EVP_EncryptInit(), and therefore that
6100 specific call must be safe. The second form is where the length passed to
6101 EVP_EncryptUpdate() can be seen from the code to be some small value and
6102 therefore there is no possibility of an overflow. Since all instances are
6103 one of these two forms, it is believed that there can be no overflows in
6104 internal code due to this problem. It should be noted that
6105 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6106 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6107 of these calls have also been analysed too and it is believed there are no
6108 instances in internal usage where an overflow could occur.
6110 This issue was reported by Guido Vranken.
6115 * Prevent ASN.1 BIO excessive memory allocation
6117 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6118 a short invalid encoding can casuse allocation of large amounts of memory
6119 potentially consuming excessive resources or exhausting memory.
6121 Any application parsing untrusted data through d2i BIO functions is
6122 affected. The memory based functions such as d2i_X509() are *not* affected.
6123 Since the memory based functions are used by the TLS library, TLS
6124 applications are not affected.
6126 This issue was reported by Brian Carpenter.
6133 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6134 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6135 in arbitrary stack data being returned in the buffer.
6137 This issue was reported by Guido Vranken.
6142 * Modify behavior of ALPN to invoke callback after SNI/servername
6143 callback, such that updates to the SSL_CTX affect ALPN.
6147 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6152 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6153 methods are enabled and ssl2 is disabled the methods return NULL.
6157 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6159 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6160 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6161 provide any "EXPORT" or "LOW" strength ciphers.
6165 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6166 is by default disabled at build-time. Builds that are not configured with
6167 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6168 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6169 will need to explicitly call either of:
6171 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6173 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6175 as appropriate. Even if either of those is used, or the application
6176 explicitly uses the version-specific SSLv2_method() or its client and
6177 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6178 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6179 ciphers, and SSLv2 56-bit DES are no longer available.
6184 * Fix a double-free in DSA code
6186 A double free bug was discovered when OpenSSL parses malformed DSA private
6187 keys and could lead to a DoS attack or memory corruption for applications
6188 that receive DSA private keys from untrusted sources. This scenario is
6191 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6197 * Disable SRP fake user seed to address a server memory leak.
6199 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6201 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6202 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6203 was changed to ignore the "fake user" SRP seed, even if the seed
6206 Users should use SRP_VBASE_get1_by_user instead. Note that in
6207 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6208 also that even though configuring the SRP seed attempts to hide
6209 invalid usernames by continuing the handshake with fake
6210 credentials, this behaviour is not constant time and no strong
6211 guarantees are made that the handshake is indistinguishable from
6212 that of a valid user.
6217 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6219 In the BN_hex2bn function the number of hex digits is calculated using an
6220 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6221 large values of `i` this can result in `bn_expand` not allocating any
6222 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6223 field as NULL leading to a subsequent NULL ptr deref. For very large values
6224 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6225 In this case memory is allocated to the internal BIGNUM data field, but it
6226 is insufficiently sized leading to heap corruption. A similar issue exists
6227 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6228 is ever called by user applications with very large untrusted hex/dec data.
6229 This is anticipated to be a rare occurrence.
6231 All OpenSSL internal usage of these functions use data that is not expected
6232 to be untrusted, e.g. config file data or application command line
6233 arguments. If user developed applications generate config file data based
6234 on untrusted data then it is possible that this could also lead to security
6235 consequences. This is also anticipated to be rare.
6237 This issue was reported to OpenSSL by Guido Vranken.
6242 * Fix memory issues in `BIO_*printf` functions
6244 The internal `fmtstr` function used in processing a "%s" format string in
6245 the `BIO_*printf` functions could overflow while calculating the length of a
6246 string and cause an OOB read when printing very long strings.
6248 Additionally the internal `doapr_outch` function can attempt to write to an
6249 OOB memory location (at an offset from the NULL pointer) in the event of a
6250 memory allocation failure. In 1.0.2 and below this could be caused where
6251 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6252 could be in processing a very long "%s" format string. Memory leaks can
6255 The first issue may mask the second issue dependent on compiler behaviour.
6256 These problems could enable attacks where large amounts of untrusted data
6257 is passed to the `BIO_*printf` functions. If applications use these functions
6258 in this way then they could be vulnerable. OpenSSL itself uses these
6259 functions when printing out human-readable dumps of ASN.1 data. Therefore
6260 applications that print this data could be vulnerable if the data is from
6261 untrusted sources. OpenSSL command line applications could also be
6262 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6263 as command line arguments.
6265 Libssl is not considered directly vulnerable. Additionally certificates etc
6266 received via remote connections via libssl are also unlikely to be able to
6267 trigger these issues because of message size limits enforced within libssl.
6269 This issue was reported to OpenSSL Guido Vranken.
6274 * Side channel attack on modular exponentiation
6276 A side-channel attack was found which makes use of cache-bank conflicts on
6277 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6278 of RSA keys. The ability to exploit this issue is limited as it relies on
6279 an attacker who has control of code in a thread running on the same
6280 hyper-threaded core as the victim thread which is performing decryptions.
6282 This issue was reported to OpenSSL by Yuval Yarom, The University of
6283 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6284 Nadia Heninger, University of Pennsylvania with more information at
6285 <http://cachebleed.info>.
6290 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6291 if no keysize is specified with default_bits. This fixes an
6292 omission in an earlier change that changed all RSA/DSA key generation
6293 commands to use 2048 bits by default.
6297 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6299 * Protection for DH small subgroup attacks
6301 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6302 switched on by default and cannot be disabled. This could have some
6307 * SSLv2 doesn't block disabled ciphers
6309 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6310 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6311 been disabled, provided that the SSLv2 protocol was not also disabled via
6314 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6315 and Sebastian Schinzel.
6320 * Reject DH handshakes with parameters shorter than 1024 bits.
6324 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6326 * Certificate verify crash with missing PSS parameter
6328 The signature verification routines will crash with a NULL pointer
6329 dereference if presented with an ASN.1 signature using the RSA PSS
6330 algorithm and absent mask generation function parameter. Since these
6331 routines are used to verify certificate signature algorithms this can be
6332 used to crash any certificate verification operation and exploited in a
6333 DoS attack. Any application which performs certificate verification is
6334 vulnerable including OpenSSL clients and servers which enable client
6337 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6342 * X509_ATTRIBUTE memory leak
6344 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6345 memory. This structure is used by the PKCS#7 and CMS routines so any
6346 application which reads PKCS#7 or CMS data from untrusted sources is
6347 affected. SSL/TLS is not affected.
6349 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6355 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6356 This changes the decoding behaviour for some invalid messages,
6357 though the change is mostly in the more lenient direction, and
6358 legacy behaviour is preserved as much as possible.
6362 * In DSA_generate_parameters_ex, if the provided seed is too short,
6363 use a random seed, as already documented.
6365 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6367 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6369 * Alternate chains certificate forgery
6371 During certificate verfification, OpenSSL will attempt to find an
6372 alternative certificate chain if the first attempt to build such a chain
6373 fails. An error in the implementation of this logic can mean that an
6374 attacker could cause certain checks on untrusted certificates to be
6375 bypassed, such as the CA flag, enabling them to use a valid leaf
6376 certificate to act as a CA and "issue" an invalid certificate.
6378 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6384 * Race condition handling PSK identify hint
6386 If PSK identity hints are received by a multi-threaded client then
6387 the values are wrongly updated in the parent SSL_CTX structure. This can
6388 result in a race condition potentially leading to a double free of the
6394 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6396 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6397 incompatibility in the handling of HMAC. The previous ABI has now been
6400 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6402 * Malformed ECParameters causes infinite loop
6404 When processing an ECParameters structure OpenSSL enters an infinite loop
6405 if the curve specified is over a specially malformed binary polynomial
6408 This can be used to perform denial of service against any
6409 system which processes public keys, certificate requests or
6410 certificates. This includes TLS clients and TLS servers with
6411 client authentication enabled.
6413 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6418 * Exploitable out-of-bounds read in X509_cmp_time
6420 X509_cmp_time does not properly check the length of the ASN1_TIME
6421 string and can read a few bytes out of bounds. In addition,
6422 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6425 An attacker can use this to craft malformed certificates and CRLs of
6426 various sizes and potentially cause a segmentation fault, resulting in
6427 a DoS on applications that verify certificates or CRLs. TLS clients
6428 that verify CRLs are affected. TLS clients and servers with client
6429 authentication enabled may be affected if they use custom verification
6432 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6433 independently by Hanno Böck.
6438 * PKCS7 crash with missing EnvelopedContent
6440 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6441 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6442 with missing content and trigger a NULL pointer dereference on parsing.
6444 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6445 structures from untrusted sources are affected. OpenSSL clients and
6446 servers are not affected.
6448 This issue was reported to OpenSSL by Michal Zalewski (Google).
6453 * CMS verify infinite loop with unknown hash function
6455 When verifying a signedData message the CMS code can enter an infinite loop
6456 if presented with an unknown hash function OID. This can be used to perform
6457 denial of service against any system which verifies signedData messages using
6459 This issue was reported to OpenSSL by Johannes Bauer.
6464 * Race condition handling NewSessionTicket
6466 If a NewSessionTicket is received by a multi-threaded client when attempting to
6467 reuse a previous ticket then a race condition can occur potentially leading to
6468 a double free of the ticket data.
6473 * Reject DH handshakes with parameters shorter than 768 bits.
6475 *Kurt Roeckx and Emilia Kasper*
6477 * dhparam: generate 2048-bit parameters by default.
6479 *Kurt Roeckx and Emilia Kasper*
6481 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6483 * Segmentation fault in ASN1_TYPE_cmp fix
6485 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6486 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6487 certificate signature algorithm consistency this can be used to crash any
6488 certificate verification operation and exploited in a DoS attack. Any
6489 application which performs certificate verification is vulnerable including
6490 OpenSSL clients and servers which enable client authentication.
6495 * ASN.1 structure reuse memory corruption fix
6497 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6498 memory corruption via an invalid write. Such reuse is and has been
6499 strongly discouraged and is believed to be rare.
6501 Applications that parse structures containing CHOICE or ANY DEFINED BY
6502 components may be affected. Certificate parsing (d2i_X509 and related
6503 functions) are however not affected. OpenSSL clients and servers are
6509 * PKCS7 NULL pointer dereferences fix
6511 The PKCS#7 parsing code does not handle missing outer ContentInfo
6512 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6513 missing content and trigger a NULL pointer dereference on parsing.
6515 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6516 otherwise parse PKCS#7 structures from untrusted sources are
6517 affected. OpenSSL clients and servers are not affected.
6519 This issue was reported to OpenSSL by Michal Zalewski (Google).
6524 * DoS via reachable assert in SSLv2 servers fix
6526 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6527 servers that both support SSLv2 and enable export cipher suites by sending
6528 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6530 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6531 (OpenSSL development team).
6536 * Use After Free following d2i_ECPrivatekey error fix
6538 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6539 could cause a use after free condition. This, in turn, could cause a double
6540 free in several private key parsing functions (such as d2i_PrivateKey
6541 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6542 for applications that receive EC private keys from untrusted
6543 sources. This scenario is considered rare.
6545 This issue was discovered by the BoringSSL project and fixed in their
6551 * X509_to_X509_REQ NULL pointer deref fix
6553 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6554 the certificate key is invalid. This function is rarely used in practice.
6556 This issue was discovered by Brian Carpenter.
6561 * Removed the export ciphers from the DEFAULT ciphers
6565 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6567 * Build fixes for the Windows and OpenVMS platforms
6569 *Matt Caswell and Richard Levitte*
6571 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6573 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6574 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6575 dereference. This could lead to a Denial Of Service attack. Thanks to
6576 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6581 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6582 dtls1_buffer_record function under certain conditions. In particular this
6583 could occur if an attacker sent repeated DTLS records with the same
6584 sequence number but for the next epoch. The memory leak could be exploited
6585 by an attacker in a Denial of Service attack through memory exhaustion.
6586 Thanks to Chris Mueller for reporting this issue.
6591 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6592 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6593 method would be set to NULL which could later result in a NULL pointer
6594 dereference. Thanks to Frank Schmirler for reporting this issue.
6599 * Abort handshake if server key exchange message is omitted for ephemeral
6602 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6603 reporting this issue.
6608 * Remove non-export ephemeral RSA code on client and server. This code
6609 violated the TLS standard by allowing the use of temporary RSA keys in
6610 non-export ciphersuites and could be used by a server to effectively
6611 downgrade the RSA key length used to a value smaller than the server
6612 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6613 INRIA or reporting this issue.
6618 * Fixed issue where DH client certificates are accepted without verification.
6619 An OpenSSL server will accept a DH certificate for client authentication
6620 without the certificate verify message. This effectively allows a client to
6621 authenticate without the use of a private key. This only affects servers
6622 which trust a client certificate authority which issues certificates
6623 containing DH keys: these are extremely rare and hardly ever encountered.
6624 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6630 * Ensure that the session ID context of an SSL is updated when its
6631 SSL_CTX is updated via SSL_set_SSL_CTX.
6633 The session ID context is typically set from the parent SSL_CTX,
6634 and can vary with the CTX.
6638 * Fix various certificate fingerprint issues.
6640 By using non-DER or invalid encodings outside the signed portion of a
6641 certificate the fingerprint can be changed without breaking the signature.
6642 Although no details of the signed portion of the certificate can be changed
6643 this can cause problems with some applications: e.g. those using the
6644 certificate fingerprint for blacklists.
6646 1. Reject signatures with non zero unused bits.
6648 If the BIT STRING containing the signature has non zero unused bits reject
6649 the signature. All current signature algorithms require zero unused bits.
6651 2. Check certificate algorithm consistency.
6653 Check the AlgorithmIdentifier inside TBS matches the one in the
6654 certificate signature. NB: this will result in signature failure
6655 errors for some broken certificates.
6657 Thanks to Konrad Kraszewski from Google for reporting this issue.
6659 3. Check DSA/ECDSA signatures use DER.
6661 Re-encode DSA/ECDSA signatures and compare with the original received
6662 signature. Return an error if there is a mismatch.
6664 This will reject various cases including garbage after signature
6665 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6666 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6667 (negative or with leading zeroes).
6669 Further analysis was conducted and fixes were developed by Stephen Henson
6670 of the OpenSSL core team.
6676 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6677 results on some platforms, including x86_64. This bug occurs at random
6678 with a very low probability, and is not known to be exploitable in any
6679 way, though its exact impact is difficult to determine. Thanks to Pieter
6680 Wuille (Blockstream) who reported this issue and also suggested an initial
6681 fix. Further analysis was conducted by the OpenSSL development team and
6682 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6683 the OpenSSL core team.
6688 * Do not resume sessions on the server if the negotiated protocol
6689 version does not match the session's version. Resuming with a different
6690 version, while not strictly forbidden by the RFC, is of questionable
6691 sanity and breaks all known clients.
6693 *David Benjamin, Emilia Käsper*
6695 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6696 early CCS messages during renegotiation. (Note that because
6697 renegotiation is encrypted, this early CCS was not exploitable.)
6701 * Tighten client-side session ticket handling during renegotiation:
6702 ensure that the client only accepts a session ticket if the server sends
6703 the extension anew in the ServerHello. Previously, a TLS client would
6704 reuse the old extension state and thus accept a session ticket if one was
6705 announced in the initial ServerHello.
6707 Similarly, ensure that the client requires a session ticket if one
6708 was advertised in the ServerHello. Previously, a TLS client would
6709 ignore a missing NewSessionTicket message.
6713 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6717 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6718 sends a carefully crafted handshake message, to cause OpenSSL to fail
6719 to free up to 64k of memory causing a memory leak. This could be
6720 exploited in a Denial Of Service attack. This issue affects OpenSSL
6721 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6722 whether SRTP is used or configured. Implementations of OpenSSL that
6723 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6725 The fix was developed by the OpenSSL team.
6730 * Session Ticket Memory Leak.
6732 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6733 integrity of that ticket is first verified. In the event of a session
6734 ticket integrity check failing, OpenSSL will fail to free memory
6735 causing a memory leak. By sending a large number of invalid session
6736 tickets an attacker could exploit this issue in a Denial Of Service
6742 * Build option no-ssl3 is incomplete.
6744 When OpenSSL is configured with "no-ssl3" as a build option, servers
6745 could accept and complete a SSL 3.0 handshake, and clients could be
6746 configured to send them.
6749 *Akamai and the OpenSSL team*
6751 * Add support for TLS_FALLBACK_SCSV.
6752 Client applications doing fallback retries should call
6753 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6756 *Adam Langley, Bodo Moeller*
6758 * Add additional DigestInfo checks.
6760 Re-encode DigestInto in DER and check against the original when
6761 verifying RSA signature: this will reject any improperly encoded
6762 DigestInfo structures.
6764 Note: this is a precautionary measure and no attacks are currently known.
6768 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6770 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6771 SRP code can be overrun an internal buffer. Add sanity check that
6772 g, A, B < N to SRP code.
6774 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6775 Group for discovering this issue.
6780 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6781 TLS 1.0 instead of higher protocol versions when the ClientHello message
6782 is badly fragmented. This allows a man-in-the-middle attacker to force a
6783 downgrade to TLS 1.0 even if both the server and the client support a
6784 higher protocol version, by modifying the client's TLS records.
6786 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6787 researching this issue.
6792 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6793 to a denial of service attack. A malicious server can crash the client
6794 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6795 ciphersuite and sending carefully crafted handshake messages.
6797 Thanks to Felix Gröbert (Google) for discovering and researching this
6803 * By sending carefully crafted DTLS packets an attacker could cause openssl
6804 to leak memory. This can be exploited through a Denial of Service attack.
6805 Thanks to Adam Langley for discovering and researching this issue.
6810 * An attacker can force openssl to consume large amounts of memory whilst
6811 processing DTLS handshake messages. This can be exploited through a
6812 Denial of Service attack.
6813 Thanks to Adam Langley for discovering and researching this issue.
6818 * An attacker can force an error condition which causes openssl to crash
6819 whilst processing DTLS packets due to memory being freed twice. This
6820 can be exploited through a Denial of Service attack.
6821 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6827 * If a multithreaded client connects to a malicious server using a resumed
6828 session and the server sends an ec point format extension it could write
6829 up to 255 bytes to freed memory.
6831 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6837 * A malicious server can crash an OpenSSL client with a null pointer
6838 dereference (read) by specifying an SRP ciphersuite even though it was not
6839 properly negotiated with the client. This can be exploited through a
6840 Denial of Service attack.
6842 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6843 discovering and researching this issue.
6848 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6849 X509_name_oneline, X509_name_print_ex et al. to leak some information
6850 from the stack. Applications may be affected if they echo pretty printing
6851 output to the attacker.
6853 Thanks to Ivan Fratric (Google) for discovering this issue.
6856 *Emilia Käsper, and Steve Henson*
6858 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6859 for corner cases. (Certain input points at infinity could lead to
6860 bogus results, with non-infinity inputs mapped to infinity too.)
6864 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6866 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6867 handshake can force the use of weak keying material in OpenSSL
6868 SSL/TLS clients and servers.
6870 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6871 researching this issue. ([CVE-2014-0224])
6873 *KIKUCHI Masashi, Steve Henson*
6875 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6876 OpenSSL DTLS client the code can be made to recurse eventually crashing
6879 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6882 *Imre Rad, Steve Henson*
6884 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6885 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6886 client or server. This is potentially exploitable to run arbitrary
6887 code on a vulnerable client or server.
6889 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6891 *Jüri Aedla, Steve Henson*
6893 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6894 are subject to a denial of service attack.
6896 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6897 this issue. ([CVE-2014-3470])
6899 *Felix Gröbert, Ivan Fratric, Steve Henson*
6901 * Harmonize version and its documentation. -f flag is used to display
6904 *mancha <mancha1@zoho.com>*
6906 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6907 in i2d_ECPrivateKey.
6909 *mancha <mancha1@zoho.com>*
6911 * Fix some double frees. These are not thought to be exploitable.
6913 *mancha <mancha1@zoho.com>*
6915 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6917 * A missing bounds check in the handling of the TLS heartbeat extension
6918 can be used to reveal up to 64k of memory to a connected client or
6921 Thanks for Neel Mehta of Google Security for discovering this bug and to
6922 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6923 preparing the fix ([CVE-2014-0160])
6925 *Adam Langley, Bodo Moeller*
6927 * Fix for the attack described in the paper "Recovering OpenSSL
6928 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6929 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6930 <http://eprint.iacr.org/2014/140>
6932 Thanks to Yuval Yarom and Naomi Benger for discovering this
6933 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6935 *Yuval Yarom and Naomi Benger*
6937 * TLS pad extension: draft-agl-tls-padding-03
6939 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6940 TLS client Hello record length value would otherwise be > 255 and
6941 less that 512 pad with a dummy extension containing zeroes so it
6942 is at least 512 bytes long.
6944 *Adam Langley, Steve Henson*
6946 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6948 * Fix for TLS record tampering bug. A carefully crafted invalid
6949 handshake could crash OpenSSL with a NULL pointer exception.
6950 Thanks to Anton Johansson for reporting this issues.
6953 * Keep original DTLS digest and encryption contexts in retransmission
6954 structures so we can use the previous session parameters if they need
6955 to be resent. ([CVE-2013-6450])
6959 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6960 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6961 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6962 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6963 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6964 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6966 *Rob Stradling, Adam Langley*
6968 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6970 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6971 supporting platforms or when small records were transferred.
6973 *Andy Polyakov, Steve Henson*
6975 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6977 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6979 This addresses the flaw in CBC record processing discovered by
6980 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6981 at: <http://www.isg.rhul.ac.uk/tls/>
6983 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6984 Security Group at Royal Holloway, University of London
6985 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6986 Emilia Käsper for the initial patch.
6989 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6991 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6992 ciphersuites which can be exploited in a denial of service attack.
6993 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6994 and detecting this bug and to Wolfgang Ettlinger
6995 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7000 * Return an error when checking OCSP signatures when key is NULL.
7001 This fixes a DoS attack. ([CVE-2013-0166])
7005 * Make openssl verify return errors.
7007 *Chris Palmer <palmer@google.com> and Ben Laurie*
7009 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7010 the right response is stapled. Also change SSL_get_certificate()
7011 so it returns the certificate actually sent.
7012 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7014 *Rob Stradling <rob.stradling@comodo.com>*
7016 * Fix possible deadlock when decoding public keys.
7020 * Don't use TLS 1.0 record version number in initial client hello
7025 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7027 * Sanity check record length before skipping explicit IV in TLS
7028 1.2, 1.1 and DTLS to fix DoS attack.
7030 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7031 fuzzing as a service testing platform.
7036 * Initialise tkeylen properly when encrypting CMS messages.
7037 Thanks to Solar Designer of Openwall for reporting this issue.
7041 * In FIPS mode don't try to use composite ciphers as they are not
7046 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7048 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7049 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7050 mean any application compiled against OpenSSL 1.0.0 headers setting
7051 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7052 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7053 0x10000000L Any application which was previously compiled against
7054 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7055 will need to be recompiled as a result. Letting be results in
7056 inability to disable specifically TLS 1.1 and in client context,
7057 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7061 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7062 disable just protocol X, but all protocols above X *if* there are
7063 protocols *below* X still enabled. In more practical terms it means
7064 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7065 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7066 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7071 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7073 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7074 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7075 in CRYPTO_realloc_clean.
7077 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7078 issue and to Adam Langley <agl@chromium.org> for fixing it.
7081 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7083 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7087 * Workarounds for some broken servers that "hang" if a client hello
7088 record length exceeds 255 bytes.
7090 1. Do not use record version number > TLS 1.0 in initial client
7091 hello: some (but not all) hanging servers will now work.
7092 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7093 the number of ciphers sent in the client hello. This should be
7094 set to an even number, such as 50, for example by passing:
7095 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7096 Most broken servers should now work.
7097 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7098 TLS 1.2 client support entirely.
7102 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7106 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7108 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7109 STRING form instead of a DigestInfo.
7113 * The format used for MDC2 RSA signatures is inconsistent between EVP
7114 and the RSA_sign/RSA_verify functions. This was made more apparent when
7115 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7116 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7117 the correct format in RSA_verify so both forms transparently work.
7121 * Some servers which support TLS 1.0 can choke if we initially indicate
7122 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7123 encrypted premaster secret. As a workaround use the maximum permitted
7124 client version in client hello, this should keep such servers happy
7125 and still work with previous versions of OpenSSL.
7129 * Add support for TLS/DTLS heartbeats.
7131 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7133 * Add support for SCTP.
7135 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7137 * Improved PRNG seeding for VOS.
7139 *Paul Green <Paul.Green@stratus.com>*
7141 * Extensive assembler packs updates, most notably:
7143 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7144 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7145 - x86_64: bit-sliced AES implementation;
7146 - ARM: NEON support, contemporary platforms optimizations;
7147 - s390x: z196 support;
7148 - `*`: GHASH and GF(2^m) multiplication implementations;
7152 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7153 (removal of unnecessary code)
7155 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7157 * Add TLS key material exporter from RFC 5705.
7161 * Add DTLS-SRTP negotiation from RFC 5764.
7165 * Add Next Protocol Negotiation,
7166 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7167 disabled with a no-npn flag to config or Configure. Code donated
7170 *Adam Langley <agl@google.com> and Ben Laurie*
7172 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7173 NIST-P256, NIST-P521, with constant-time single point multiplication on
7174 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7175 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7176 Code made available under Apache License version 2.0.
7178 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7179 line to include this in your build of OpenSSL, and run "make depend" (or
7180 "make update"). This enables the following EC_METHODs:
7182 EC_GFp_nistp224_method()
7183 EC_GFp_nistp256_method()
7184 EC_GFp_nistp521_method()
7186 EC_GROUP_new_by_curve_name() will automatically use these (while
7187 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7190 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7192 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7193 all platforms. Move ssize_t definition from e_os.h to the public
7194 header file e_os2.h as it now appears in public header file cms.h
7198 * New -sigopt option to the ca, req and x509 utilities. Additional
7199 signature parameters can be passed using this option and in
7204 * Add RSA PSS signing function. This will generate and set the
7205 appropriate AlgorithmIdentifiers for PSS based on those in the
7206 corresponding EVP_MD_CTX structure. No application support yet.
7210 * Support for companion algorithm specific ASN1 signing routines.
7211 New function ASN1_item_sign_ctx() signs a pre-initialised
7212 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7213 the appropriate parameters.
7217 * Add new algorithm specific ASN1 verification initialisation function
7218 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7219 handling will be the same no matter what EVP_PKEY_METHOD is used.
7220 Add a PSS handler to support verification of PSS signatures: checked
7221 against a number of sample certificates.
7225 * Add signature printing for PSS. Add PSS OIDs.
7227 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7229 * Add algorithm specific signature printing. An individual ASN1 method
7230 can now print out signatures instead of the standard hex dump.
7232 More complex signatures (e.g. PSS) can print out more meaningful
7233 information. Include DSA version that prints out the signature
7238 * Password based recipient info support for CMS library: implementing
7243 * Split password based encryption into PBES2 and PBKDF2 functions. This
7244 neatly separates the code into cipher and PBE sections and is required
7245 for some algorithms that split PBES2 into separate pieces (such as
7246 password based CMS).
7250 * Session-handling fixes:
7251 - Fix handling of connections that are resuming with a session ID,
7252 but also support Session Tickets.
7253 - Fix a bug that suppressed issuing of a new ticket if the client
7254 presented a ticket with an expired session.
7255 - Try to set the ticket lifetime hint to something reasonable.
7256 - Make tickets shorter by excluding irrelevant information.
7257 - On the client side, don't ignore renewed tickets.
7259 *Adam Langley, Bodo Moeller (Google)*
7261 * Fix PSK session representation.
7265 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7267 This work was sponsored by Intel.
7271 * Add GCM support to TLS library. Some custom code is needed to split
7272 the IV between the fixed (from PRF) and explicit (from TLS record)
7273 portions. This adds all GCM ciphersuites supported by RFC5288 and
7274 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7275 add a special AESGCM string for GCM only.
7279 * Expand range of ctrls for AES GCM. Permit setting invocation
7280 field on decrypt and retrieval of invocation field only on encrypt.
7284 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7285 As required by RFC5289 these ciphersuites cannot be used if for
7286 versions of TLS earlier than 1.2.
7290 * For FIPS capable OpenSSL interpret a NULL default public key method
7291 as unset and return the appropriate default but do *not* set the default.
7292 This means we can return the appropriate method in applications that
7293 switch between FIPS and non-FIPS modes.
7297 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7298 ENGINE is used then we cannot handle that in the FIPS module so we
7299 keep original code iff non-FIPS operations are allowed.
7303 * Add -attime option to openssl utilities.
7305 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7307 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7311 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7312 FIPS EC methods unconditionally for now.
7316 * New build option no-ec2m to disable characteristic 2 code.
7320 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7321 all cases can be covered as some introduce binary incompatibilities.
7325 * Redirect RSA operations to FIPS module including keygen,
7326 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7330 * Add similar low level API blocking to ciphers.
7334 * Low level digest APIs are not approved in FIPS mode: any attempt
7335 to use these will cause a fatal error. Applications that *really* want
7336 to use them can use the `private_*` version instead.
7340 * Redirect cipher operations to FIPS module for FIPS builds.
7344 * Redirect digest operations to FIPS module for FIPS builds.
7348 * Update build system to add "fips" flag which will link in fipscanister.o
7349 for static and shared library builds embedding a signature if needed.
7353 * Output TLS supported curves in preference order instead of numerical
7354 order. This is currently hardcoded for the highest order curves first.
7355 This should be configurable so applications can judge speed vs strength.
7359 * Add TLS v1.2 server support for client authentication.
7363 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7368 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7369 FIPS modules versions.
7373 * Add TLS v1.2 client side support for client authentication. Keep cache
7374 of handshake records longer as we don't know the hash algorithm to use
7375 until after the certificate request message is received.
7379 * Initial TLS v1.2 client support. Add a default signature algorithms
7380 extension including all the algorithms we support. Parse new signature
7381 format in client key exchange. Relax some ECC signing restrictions for
7382 TLS v1.2 as indicated in RFC5246.
7386 * Add server support for TLS v1.2 signature algorithms extension. Switch
7387 to new signature format when needed using client digest preference.
7388 All server ciphersuites should now work correctly in TLS v1.2. No client
7389 support yet and no support for client certificates.
7393 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7394 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7395 ciphersuites. At present only RSA key exchange ciphersuites work with
7396 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7397 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7398 and version checking.
7402 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7403 with this defined it will not be affected by any changes to ssl internal
7404 structures. Add several utility functions to allow openssl application
7405 to work with OPENSSL_NO_SSL_INTERN defined.
7409 * A long standing patch to add support for SRP from EdelWeb (Peter
7410 Sylvester and Christophe Renou) was integrated.
7411 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7412 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7415 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7419 * Permit abbreviated handshakes when renegotiating using the function
7420 SSL_renegotiate_abbreviated().
7422 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7424 * Add call to ENGINE_register_all_complete() to
7425 ENGINE_load_builtin_engines(), so some implementations get used
7426 automatically instead of needing explicit application support.
7430 * Add support for TLS key exporter as described in RFC5705.
7432 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7434 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7435 a few changes are required:
7437 Add SSL_OP_NO_TLSv1_1 flag.
7438 Add TLSv1_1 methods.
7439 Update version checking logic to handle version 1.1.
7440 Add explicit IV handling (ported from DTLS code).
7441 Add command line options to s_client/s_server.
7448 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7450 * X509_ATTRIBUTE memory leak
7452 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7453 memory. This structure is used by the PKCS#7 and CMS routines so any
7454 application which reads PKCS#7 or CMS data from untrusted sources is
7455 affected. SSL/TLS is not affected.
7457 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7463 * Race condition handling PSK identify hint
7465 If PSK identity hints are received by a multi-threaded client then
7466 the values are wrongly updated in the parent SSL_CTX structure. This can
7467 result in a race condition potentially leading to a double free of the
7473 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7475 * Malformed ECParameters causes infinite loop
7477 When processing an ECParameters structure OpenSSL enters an infinite loop
7478 if the curve specified is over a specially malformed binary polynomial
7481 This can be used to perform denial of service against any
7482 system which processes public keys, certificate requests or
7483 certificates. This includes TLS clients and TLS servers with
7484 client authentication enabled.
7486 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7491 * Exploitable out-of-bounds read in X509_cmp_time
7493 X509_cmp_time does not properly check the length of the ASN1_TIME
7494 string and can read a few bytes out of bounds. In addition,
7495 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7498 An attacker can use this to craft malformed certificates and CRLs of
7499 various sizes and potentially cause a segmentation fault, resulting in
7500 a DoS on applications that verify certificates or CRLs. TLS clients
7501 that verify CRLs are affected. TLS clients and servers with client
7502 authentication enabled may be affected if they use custom verification
7505 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7506 independently by Hanno Böck.
7511 * PKCS7 crash with missing EnvelopedContent
7513 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7514 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7515 with missing content and trigger a NULL pointer dereference on parsing.
7517 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7518 structures from untrusted sources are affected. OpenSSL clients and
7519 servers are not affected.
7521 This issue was reported to OpenSSL by Michal Zalewski (Google).
7526 * CMS verify infinite loop with unknown hash function
7528 When verifying a signedData message the CMS code can enter an infinite loop
7529 if presented with an unknown hash function OID. This can be used to perform
7530 denial of service against any system which verifies signedData messages using
7532 This issue was reported to OpenSSL by Johannes Bauer.
7537 * Race condition handling NewSessionTicket
7539 If a NewSessionTicket is received by a multi-threaded client when attempting to
7540 reuse a previous ticket then a race condition can occur potentially leading to
7541 a double free of the ticket data.
7546 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7548 * Segmentation fault in ASN1_TYPE_cmp fix
7550 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7551 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7552 certificate signature algorithm consistency this can be used to crash any
7553 certificate verification operation and exploited in a DoS attack. Any
7554 application which performs certificate verification is vulnerable including
7555 OpenSSL clients and servers which enable client authentication.
7560 * ASN.1 structure reuse memory corruption fix
7562 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7563 memory corruption via an invalid write. Such reuse is and has been
7564 strongly discouraged and is believed to be rare.
7566 Applications that parse structures containing CHOICE or ANY DEFINED BY
7567 components may be affected. Certificate parsing (d2i_X509 and related
7568 functions) are however not affected. OpenSSL clients and servers are
7574 * PKCS7 NULL pointer dereferences fix
7576 The PKCS#7 parsing code does not handle missing outer ContentInfo
7577 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7578 missing content and trigger a NULL pointer dereference on parsing.
7580 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7581 otherwise parse PKCS#7 structures from untrusted sources are
7582 affected. OpenSSL clients and servers are not affected.
7584 This issue was reported to OpenSSL by Michal Zalewski (Google).
7589 * DoS via reachable assert in SSLv2 servers fix
7591 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7592 servers that both support SSLv2 and enable export cipher suites by sending
7593 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7595 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7596 (OpenSSL development team).
7601 * Use After Free following d2i_ECPrivatekey error fix
7603 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7604 could cause a use after free condition. This, in turn, could cause a double
7605 free in several private key parsing functions (such as d2i_PrivateKey
7606 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7607 for applications that receive EC private keys from untrusted
7608 sources. This scenario is considered rare.
7610 This issue was discovered by the BoringSSL project and fixed in their
7616 * X509_to_X509_REQ NULL pointer deref fix
7618 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7619 the certificate key is invalid. This function is rarely used in practice.
7621 This issue was discovered by Brian Carpenter.
7626 * Removed the export ciphers from the DEFAULT ciphers
7630 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7632 * Build fixes for the Windows and OpenVMS platforms
7634 *Matt Caswell and Richard Levitte*
7636 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7638 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7639 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7640 dereference. This could lead to a Denial Of Service attack. Thanks to
7641 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7646 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7647 dtls1_buffer_record function under certain conditions. In particular this
7648 could occur if an attacker sent repeated DTLS records with the same
7649 sequence number but for the next epoch. The memory leak could be exploited
7650 by an attacker in a Denial of Service attack through memory exhaustion.
7651 Thanks to Chris Mueller for reporting this issue.
7656 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7657 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7658 method would be set to NULL which could later result in a NULL pointer
7659 dereference. Thanks to Frank Schmirler for reporting this issue.
7664 * Abort handshake if server key exchange message is omitted for ephemeral
7667 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7668 reporting this issue.
7673 * Remove non-export ephemeral RSA code on client and server. This code
7674 violated the TLS standard by allowing the use of temporary RSA keys in
7675 non-export ciphersuites and could be used by a server to effectively
7676 downgrade the RSA key length used to a value smaller than the server
7677 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7678 INRIA or reporting this issue.
7683 * Fixed issue where DH client certificates are accepted without verification.
7684 An OpenSSL server will accept a DH certificate for client authentication
7685 without the certificate verify message. This effectively allows a client to
7686 authenticate without the use of a private key. This only affects servers
7687 which trust a client certificate authority which issues certificates
7688 containing DH keys: these are extremely rare and hardly ever encountered.
7689 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7695 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7696 results on some platforms, including x86_64. This bug occurs at random
7697 with a very low probability, and is not known to be exploitable in any
7698 way, though its exact impact is difficult to determine. Thanks to Pieter
7699 Wuille (Blockstream) who reported this issue and also suggested an initial
7700 fix. Further analysis was conducted by the OpenSSL development team and
7701 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7702 the OpenSSL core team.
7707 * Fix various certificate fingerprint issues.
7709 By using non-DER or invalid encodings outside the signed portion of a
7710 certificate the fingerprint can be changed without breaking the signature.
7711 Although no details of the signed portion of the certificate can be changed
7712 this can cause problems with some applications: e.g. those using the
7713 certificate fingerprint for blacklists.
7715 1. Reject signatures with non zero unused bits.
7717 If the BIT STRING containing the signature has non zero unused bits reject
7718 the signature. All current signature algorithms require zero unused bits.
7720 2. Check certificate algorithm consistency.
7722 Check the AlgorithmIdentifier inside TBS matches the one in the
7723 certificate signature. NB: this will result in signature failure
7724 errors for some broken certificates.
7726 Thanks to Konrad Kraszewski from Google for reporting this issue.
7728 3. Check DSA/ECDSA signatures use DER.
7730 Reencode DSA/ECDSA signatures and compare with the original received
7731 signature. Return an error if there is a mismatch.
7733 This will reject various cases including garbage after signature
7734 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7735 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7736 (negative or with leading zeroes).
7738 Further analysis was conducted and fixes were developed by Stephen Henson
7739 of the OpenSSL core team.
7745 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7747 * Session Ticket Memory Leak.
7749 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7750 integrity of that ticket is first verified. In the event of a session
7751 ticket integrity check failing, OpenSSL will fail to free memory
7752 causing a memory leak. By sending a large number of invalid session
7753 tickets an attacker could exploit this issue in a Denial Of Service
7759 * Build option no-ssl3 is incomplete.
7761 When OpenSSL is configured with "no-ssl3" as a build option, servers
7762 could accept and complete a SSL 3.0 handshake, and clients could be
7763 configured to send them.
7766 *Akamai and the OpenSSL team*
7768 * Add support for TLS_FALLBACK_SCSV.
7769 Client applications doing fallback retries should call
7770 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7773 *Adam Langley, Bodo Moeller*
7775 * Add additional DigestInfo checks.
7777 Reencode DigestInto in DER and check against the original when
7778 verifying RSA signature: this will reject any improperly encoded
7779 DigestInfo structures.
7781 Note: this is a precautionary measure and no attacks are currently known.
7785 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7787 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7788 to a denial of service attack. A malicious server can crash the client
7789 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7790 ciphersuite and sending carefully crafted handshake messages.
7792 Thanks to Felix Gröbert (Google) for discovering and researching this
7798 * By sending carefully crafted DTLS packets an attacker could cause openssl
7799 to leak memory. This can be exploited through a Denial of Service attack.
7800 Thanks to Adam Langley for discovering and researching this issue.
7805 * An attacker can force openssl to consume large amounts of memory whilst
7806 processing DTLS handshake messages. This can be exploited through a
7807 Denial of Service attack.
7808 Thanks to Adam Langley for discovering and researching this issue.
7813 * An attacker can force an error condition which causes openssl to crash
7814 whilst processing DTLS packets due to memory being freed twice. This
7815 can be exploited through a Denial of Service attack.
7816 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7822 * If a multithreaded client connects to a malicious server using a resumed
7823 session and the server sends an ec point format extension it could write
7824 up to 255 bytes to freed memory.
7826 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7832 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7833 X509_name_oneline, X509_name_print_ex et al. to leak some information
7834 from the stack. Applications may be affected if they echo pretty printing
7835 output to the attacker.
7837 Thanks to Ivan Fratric (Google) for discovering this issue.
7840 *Emilia Käsper, and Steve Henson*
7842 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7843 for corner cases. (Certain input points at infinity could lead to
7844 bogus results, with non-infinity inputs mapped to infinity too.)
7848 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7850 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7851 handshake can force the use of weak keying material in OpenSSL
7852 SSL/TLS clients and servers.
7854 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7855 researching this issue. ([CVE-2014-0224])
7857 *KIKUCHI Masashi, Steve Henson*
7859 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7860 OpenSSL DTLS client the code can be made to recurse eventually crashing
7863 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7866 *Imre Rad, Steve Henson*
7868 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7869 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7870 client or server. This is potentially exploitable to run arbitrary
7871 code on a vulnerable client or server.
7873 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7875 *Jüri Aedla, Steve Henson*
7877 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7878 are subject to a denial of service attack.
7880 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7881 this issue. ([CVE-2014-3470])
7883 *Felix Gröbert, Ivan Fratric, Steve Henson*
7885 * Harmonize version and its documentation. -f flag is used to display
7888 *mancha <mancha1@zoho.com>*
7890 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7891 in i2d_ECPrivateKey.
7893 *mancha <mancha1@zoho.com>*
7895 * Fix some double frees. These are not thought to be exploitable.
7897 *mancha <mancha1@zoho.com>*
7899 * Fix for the attack described in the paper "Recovering OpenSSL
7900 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7901 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7902 <http://eprint.iacr.org/2014/140>
7904 Thanks to Yuval Yarom and Naomi Benger for discovering this
7905 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7907 *Yuval Yarom and Naomi Benger*
7909 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7911 * Keep original DTLS digest and encryption contexts in retransmission
7912 structures so we can use the previous session parameters if they need
7913 to be resent. ([CVE-2013-6450])
7917 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7918 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7919 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7920 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7921 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7922 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7924 *Rob Stradling, Adam Langley*
7926 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7928 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7930 This addresses the flaw in CBC record processing discovered by
7931 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7932 at: <http://www.isg.rhul.ac.uk/tls/>
7934 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7935 Security Group at Royal Holloway, University of London
7936 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7937 Emilia Käsper for the initial patch.
7940 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7942 * Return an error when checking OCSP signatures when key is NULL.
7943 This fixes a DoS attack. ([CVE-2013-0166])
7947 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7948 the right response is stapled. Also change SSL_get_certificate()
7949 so it returns the certificate actually sent.
7950 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7951 (This is a backport)
7953 *Rob Stradling <rob.stradling@comodo.com>*
7955 * Fix possible deadlock when decoding public keys.
7959 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7961 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7964 * Sanity check record length before skipping explicit IV in DTLS
7967 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7968 fuzzing as a service testing platform.
7973 * Initialise tkeylen properly when encrypting CMS messages.
7974 Thanks to Solar Designer of Openwall for reporting this issue.
7978 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7980 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7981 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7982 in CRYPTO_realloc_clean.
7984 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7985 issue and to Adam Langley <agl@chromium.org> for fixing it.
7988 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7990 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7992 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7993 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7994 content decryption and always return the same error. Note: this attack
7995 needs on average 2^20 messages so it only affects automated senders. The
7996 old behaviour can be re-enabled in the CMS code by setting the
7997 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7998 an MMA defence is not necessary.
7999 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8000 this issue. ([CVE-2012-0884])
8004 * Fix CVE-2011-4619: make sure we really are receiving a
8005 client hello before rejecting multiple SGC restarts. Thanks to
8006 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8010 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8012 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8013 Thanks to Antonio Martin, Enterprise Secure Access Research and
8014 Development, Cisco Systems, Inc. for discovering this bug and
8015 preparing a fix. ([CVE-2012-0050])
8019 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8021 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8022 of the Vaudenay padding oracle attack on CBC mode encryption
8023 which enables an efficient plaintext recovery attack against
8024 the OpenSSL implementation of DTLS. Their attack exploits timing
8025 differences arising during decryption processing. A research
8026 paper describing this attack can be found at:
8027 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8028 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8029 Security Group at Royal Holloway, University of London
8030 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8031 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8032 for preparing the fix. ([CVE-2011-4108])
8034 *Robin Seggelmann, Michael Tuexen*
8036 * Clear bytes used for block padding of SSL 3.0 records.
8039 *Adam Langley (Google)*
8041 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8042 Kadianakis <desnacked@gmail.com> for discovering this issue and
8043 Adam Langley for preparing the fix. ([CVE-2011-4619])
8045 *Adam Langley (Google)*
8047 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8049 *Andrey Kulikov <amdeich@gmail.com>*
8051 * Prevent malformed RFC3779 data triggering an assertion failure.
8052 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8053 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8055 *Rob Austein <sra@hactrn.net>*
8057 * Improved PRNG seeding for VOS.
8059 *Paul Green <Paul.Green@stratus.com>*
8061 * Fix ssl_ciph.c set-up race.
8063 *Adam Langley (Google)*
8065 * Fix spurious failures in ecdsatest.c.
8067 *Emilia Käsper (Google)*
8069 * Fix the BIO_f_buffer() implementation (which was mixing different
8070 interpretations of the `..._len` fields).
8072 *Adam Langley (Google)*
8074 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8075 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8076 threads won't reuse the same blinding coefficients.
8078 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8079 lock to call BN_BLINDING_invert_ex, and avoids one use of
8080 BN_BLINDING_update for each BN_BLINDING structure (previously,
8081 the last update always remained unused).
8083 *Emilia Käsper (Google)*
8085 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8087 *Bob Buckholz (Google)*
8089 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8091 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8092 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8094 *Kaspar Brand <ossl@velox.ch>*
8096 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8097 for multi-threaded use of ECDH. ([CVE-2011-3210])
8099 *Adam Langley (Google)*
8101 * Fix x509_name_ex_d2i memory leak on bad inputs.
8105 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8106 signature public key algorithm by using OID xref utilities instead.
8107 Before this you could only use some ECC ciphersuites with SHA1 only.
8111 * Add protection against ECDSA timing attacks as mentioned in the paper
8112 by Billy Bob Brumley and Nicola Tuveri, see:
8113 <http://eprint.iacr.org/2011/232.pdf>
8115 *Billy Bob Brumley and Nicola Tuveri*
8117 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8119 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8121 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8123 * Fix bug in string printing code: if *any* escaping is enabled we must
8124 escape the escape character (backslash) or the resulting string is
8129 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8131 * Disable code workaround for ancient and obsolete Netscape browsers
8132 and servers: an attacker can use it in a ciphersuite downgrade attack.
8133 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8137 * Fixed J-PAKE implementation error, originally discovered by
8138 Sebastien Martini, further info and confirmation from Stefan
8139 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8143 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8145 * Fix extension code to avoid race conditions which can result in a buffer
8146 overrun vulnerability: resumed sessions must not be modified as they can
8147 be shared by multiple threads. CVE-2010-3864
8151 * Fix WIN32 build system to correctly link an ENGINE directory into
8156 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8158 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8161 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8163 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8165 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8166 context. The operation can be customised via the ctrl mechanism in
8167 case ENGINEs want to include additional functionality.
8171 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8175 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8176 output hashes compatible with older versions of OpenSSL.
8178 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8180 * Fix compression algorithm handling: if resuming a session use the
8181 compression algorithm of the resumed session instead of determining
8182 it from client hello again. Don't allow server to change algorithm.
8186 * Add load_crls() function to commands tidying load_certs() too. Add option
8187 to verify utility to allow additional CRLs to be included.
8191 * Update OCSP request code to permit adding custom headers to the request:
8192 some responders need this.
8196 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8199 *Julia Lawall <julia@diku.dk>*
8201 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8202 needlessly dereferenced structures, used obsolete functions and
8203 didn't handle all updated verify codes correctly.
8207 * Disable MD2 in the default configuration.
8211 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8212 indicate the initial BIO being pushed or popped. This makes it possible
8213 to determine whether the BIO is the one explicitly called or as a result
8214 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8215 it handles reference counts correctly and doesn't zero out the I/O bio
8216 when it is not being explicitly popped. WARNING: applications which
8217 included workarounds for the old buggy behaviour will need to be modified
8218 or they could free up already freed BIOs.
8222 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8223 renaming to all platforms (within the 0.9.8 branch, this was
8224 done conditionally on Netware platforms to avoid a name clash).
8226 *Guenter <lists@gknw.net>*
8228 * Add ECDHE and PSK support to DTLS.
8230 *Michael Tuexen <tuexen@fh-muenster.de>*
8232 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8237 * Add "missing" function EVP_MD_flags() (without this the only way to
8238 retrieve a digest flags is by accessing the structure directly. Update
8239 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8240 or cipher is registered as in the "from" argument. Print out all
8241 registered digests in the dgst usage message instead of manually
8242 attempting to work them out.
8246 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8247 this allows the use of compression and extensions. Change default cipher
8248 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8249 by default unless an application cipher string requests it.
8253 * Alter match criteria in PKCS12_parse(). It used to try to use local
8254 key ids to find matching certificates and keys but some PKCS#12 files
8255 don't follow the (somewhat unwritten) rules and this strategy fails.
8256 Now just gather all certificates together and the first private key
8257 then look for the first certificate that matches the key.
8261 * Support use of registered digest and cipher names for dgst and cipher
8262 commands instead of having to add each one as a special case. So now
8269 openssl dgst -sha256 foo
8271 and this works for ENGINE based algorithms too.
8275 * Update Gost ENGINE to support parameter files.
8277 *Victor B. Wagner <vitus@cryptocom.ru>*
8279 * Support GeneralizedTime in ca utility.
8281 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8283 * Enhance the hash format used for certificate directory links. The new
8284 form uses the canonical encoding (meaning equivalent names will work
8285 even if they aren't identical) and uses SHA1 instead of MD5. This form
8286 is incompatible with the older format and as a result c_rehash should
8287 be used to rebuild symbolic links.
8291 * Make PKCS#8 the default write format for private keys, replacing the
8292 traditional format. This form is standardised, more secure and doesn't
8293 include an implicit MD5 dependency.
8297 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8298 committed to OpenSSL should pass this lot as a minimum.
8302 * Add session ticket override functionality for use by EAP-FAST.
8304 *Jouni Malinen <j@w1.fi>*
8306 * Modify HMAC functions to return a value. Since these can be implemented
8307 in an ENGINE errors can occur.
8311 * Type-checked OBJ_bsearch_ex.
8315 * Type-checked OBJ_bsearch. Also some constification necessitated
8316 by type-checking. Still to come: TXT_DB, bsearch(?),
8317 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8322 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8323 seconds to a tm structure directly, instead of going through OS
8324 specific date routines. This avoids any issues with OS routines such
8325 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8326 and X509_time_adj_ex() to cover the extended range. The existing
8327 X509_time_adj() is still usable and will no longer have any date issues.
8331 * Delta CRL support. New use deltas option which will attempt to locate
8332 and search any appropriate delta CRLs available.
8334 This work was sponsored by Google.
8338 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8339 code and add additional score elements. Validate alternate CRL paths
8340 as part of the CRL checking and indicate a new error "CRL path validation
8341 error" in this case. Applications wanting additional details can use
8342 the verify callback and check the new "parent" field. If this is not
8343 NULL CRL path validation is taking place. Existing applications won't
8344 see this because it requires extended CRL support which is off by
8347 This work was sponsored by Google.
8351 * Support for freshest CRL extension.
8353 This work was sponsored by Google.
8357 * Initial indirect CRL support. Currently only supported in the CRLs
8358 passed directly and not via lookup. Process certificate issuer
8359 CRL entry extension and lookup CRL entries by bother issuer name
8360 and serial number. Check and process CRL issuer entry in IDP extension.
8362 This work was sponsored by Google.
8366 * Add support for distinct certificate and CRL paths. The CRL issuer
8367 certificate is validated separately in this case. Only enabled if
8368 an extended CRL support flag is set: this flag will enable additional
8369 CRL functionality in future.
8371 This work was sponsored by Google.
8375 * Add support for policy mappings extension.
8377 This work was sponsored by Google.
8381 * Fixes to pathlength constraint, self issued certificate handling,
8382 policy processing to align with RFC3280 and PKITS tests.
8384 This work was sponsored by Google.
8388 * Support for name constraints certificate extension. DN, email, DNS
8389 and URI types are currently supported.
8391 This work was sponsored by Google.
8395 * To cater for systems that provide a pointer-based thread ID rather
8396 than numeric, deprecate the current numeric thread ID mechanism and
8397 replace it with a structure and associated callback type. This
8398 mechanism allows a numeric "hash" to be extracted from a thread ID in
8399 either case, and on platforms where pointers are larger than 'long',
8400 mixing is done to help ensure the numeric 'hash' is usable even if it
8401 can't be guaranteed unique. The default mechanism is to use "&errno"
8402 as a pointer-based thread ID to distinguish between threads.
8404 Applications that want to provide their own thread IDs should now use
8405 CRYPTO_THREADID_set_callback() to register a callback that will call
8406 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8408 Note that ERR_remove_state() is now deprecated, because it is tied
8409 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8410 to free the current thread's error state should be replaced by
8411 ERR_remove_thread_state(NULL).
8413 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8414 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8415 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8416 application was previously providing a numeric thread callback that
8417 was inappropriate for distinguishing threads, then uniqueness might
8418 have been obtained with &errno that happened immediately in the
8419 intermediate development versions of OpenSSL; this is no longer the
8420 case, the numeric thread callback will now override the automatic use
8423 *Geoff Thorpe, with help from Bodo Moeller*
8425 * Initial support for different CRL issuing certificates. This covers a
8426 simple case where the self issued certificates in the chain exist and
8427 the real CRL issuer is higher in the existing chain.
8429 This work was sponsored by Google.
8433 * Removed effectively defunct crypto/store from the build.
8437 * Revamp of STACK to provide stronger type-checking. Still to come:
8438 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8439 ASN1_STRING, CONF_VALUE.
8443 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8444 RAM on SSL connections. This option can save about 34k per idle SSL.
8448 * Revamp of LHASH to provide stronger type-checking. Still to come:
8449 STACK, TXT_DB, bsearch, qsort.
8453 * Initial support for Cryptographic Message Syntax (aka CMS) based
8454 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8455 support for data, signedData, compressedData, digestedData and
8456 encryptedData, envelopedData types included. Scripts to check against
8457 RFC4134 examples draft and interop and consistency checks of many
8458 content types and variants.
8462 * Add options to enc utility to support use of zlib compression BIO.
8466 * Extend mk1mf to support importing of options and assembly language
8467 files from Configure script, currently only included in VC-WIN32.
8468 The assembly language rules can now optionally generate the source
8469 files from the associated perl scripts.
8473 * Implement remaining functionality needed to support GOST ciphersuites.
8474 Interop testing has been performed using CryptoPro implementations.
8476 *Victor B. Wagner <vitus@cryptocom.ru>*
8478 * s390x assembler pack.
8482 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8487 * Implement Opaque PRF Input TLS extension as specified in
8488 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8489 official specification yet and no extension type assignment by
8490 IANA exists, this extension (for now) will have to be explicitly
8491 enabled when building OpenSSL by providing the extension number
8492 to use. For example, specify an option
8494 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8496 to the "config" or "Configure" script to enable the extension,
8497 assuming extension number 0x9527 (which is a completely arbitrary
8498 and unofficial assignment based on the MD5 hash of the Internet
8499 Draft). Note that by doing so, you potentially lose
8500 interoperability with other TLS implementations since these might
8501 be using the same extension number for other purposes.
8503 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8504 opaque PRF input value to use in the handshake. This will create
8505 an internal copy of the length-'len' string at 'src', and will
8506 return non-zero for success.
8508 To get more control and flexibility, provide a callback function
8511 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8512 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8516 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8519 Callback function 'cb' will be called in handshakes, and is
8520 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8521 Argument 'arg' is for application purposes (the value as given to
8522 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8523 be provided to the callback function). The callback function
8524 has to return non-zero to report success: usually 1 to use opaque
8525 PRF input just if possible, or 2 to enforce use of the opaque PRF
8526 input. In the latter case, the library will abort the handshake
8527 if opaque PRF input is not successfully negotiated.
8529 Arguments 'peerinput' and 'len' given to the callback function
8530 will always be NULL and 0 in the case of a client. A server will
8531 see the client's opaque PRF input through these variables if
8532 available (NULL and 0 otherwise). Note that if the server
8533 provides an opaque PRF input, the length must be the same as the
8534 length of the client's opaque PRF input.
8536 Note that the callback function will only be called when creating
8537 a new session (session resumption can resume whatever was
8538 previously negotiated), and will not be called in SSL 2.0
8539 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8540 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8541 for applications that need to enforce opaque PRF input.
8545 * Update ssl code to support digests other than SHA1+MD5 for handshake
8548 *Victor B. Wagner <vitus@cryptocom.ru>*
8550 * Add RFC4507 support to OpenSSL. This includes the corrections in
8551 RFC4507bis. The encrypted ticket format is an encrypted encoded
8552 SSL_SESSION structure, that way new session features are automatically
8555 If a client application caches session in an SSL_SESSION structure
8556 support is transparent because tickets are now stored in the encoded
8559 The SSL_CTX structure automatically generates keys for ticket
8560 protection in servers so again support should be possible
8561 with no application modification.
8563 If a client or server wishes to disable RFC4507 support then the option
8564 SSL_OP_NO_TICKET can be set.
8566 Add a TLS extension debugging callback to allow the contents of any client
8567 or server extensions to be examined.
8569 This work was sponsored by Google.
8573 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8574 OpenSSL should now compile cleanly on gcc 4.2
8576 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8578 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8579 support including streaming MAC support: this is required for GOST
8580 ciphersuite support.
8582 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8584 * Add option -stream to use PKCS#7 streaming in smime utility. New
8585 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8586 to output in BER and PEM format.
8590 * Experimental support for use of HMAC via EVP_PKEY interface. This
8591 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8592 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8593 ENGINE support for HMAC keys which are unextractable. New -mac and
8594 -macopt options to dgst utility.
8598 * New option -sigopt to dgst utility. Update dgst to use
8599 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8600 alternative signing parameters such as X9.31 or PSS in the dgst
8605 * Change ssl_cipher_apply_rule(), the internal function that does
8606 the work each time a ciphersuite string requests enabling
8607 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8608 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8609 the order of disabled ciphersuites such that those ciphersuites
8610 that most recently went from enabled to disabled not only stay
8611 in order with respect to each other, but also have higher priority
8612 than other disabled ciphersuites the next time ciphersuites are
8615 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8616 the same ciphersuites as with "HIGH" alone, but in a specific
8617 order where the PSK ciphersuites come first (since they are the
8618 most recently disabled ciphersuites when "HIGH" is parsed).
8620 Also, change ssl_create_cipher_list() (using this new
8621 functionality) such that between otherwise identical
8622 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8627 * Change ssl_create_cipher_list() so that it automatically
8628 arranges the ciphersuites in reasonable order before starting
8629 to process the rule string. Thus, the definition for "DEFAULT"
8630 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8631 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8632 This makes it much easier to arrive at a reasonable default order
8633 in applications for which anonymous ciphers are OK (meaning
8634 that you can't actually use DEFAULT).
8636 *Bodo Moeller; suggested by Victor Duchovni*
8638 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8639 processing) into multiple integers instead of setting
8640 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8641 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8642 (These masks as well as the individual bit definitions are hidden
8643 away into the non-exported interface ssl/ssl_locl.h, so this
8644 change to the definition of the SSL_CIPHER structure shouldn't
8645 affect applications.) This give us more bits for each of these
8646 categories, so there is no longer a need to coagulate AES128 and
8647 AES256 into a single algorithm bit, and to coagulate Camellia128
8648 and Camellia256 into a single algorithm bit, which has led to all
8651 Thus, among other things, the kludge introduced in 0.9.7m and
8652 0.9.8e for masking out AES256 independently of AES128 or masking
8653 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8655 With the change, we also introduce new ciphersuite aliases that
8656 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8661 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8662 Use the leftmost N bytes of the signature input if the input is
8663 larger than the prime q (with N being the size in bytes of q).
8667 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8668 it yet and it is largely untested.
8672 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8676 * Initial incomplete changes to avoid need for function casts in OpenSSL
8677 some compilers (gcc 4.2 and later) reject their use. Safestack is
8678 reimplemented. Update ASN1 to avoid use of legacy functions.
8682 * Win32/64 targets are linked with Winsock2.
8686 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8687 to external functions. This can be used to increase CRL handling
8688 efficiency especially when CRLs are very large by (for example) storing
8689 the CRL revoked certificates in a database.
8693 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8694 new CRLs added to a directory can be used. New command line option
8695 -verify_return_error to s_client and s_server. This causes real errors
8696 to be returned by the verify callback instead of carrying on no matter
8697 what. This reflects the way a "real world" verify callback would behave.
8701 * GOST engine, supporting several GOST algorithms and public key formats.
8702 Kindly donated by Cryptocom.
8706 * Partial support for Issuing Distribution Point CRL extension. CRLs
8707 partitioned by DP are handled but no indirect CRL or reason partitioning
8708 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8709 selected via a scoring technique which handles IDP and AKID in CRLs.
8713 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8714 will ultimately be used for all verify operations: this will remove the
8715 X509_STORE dependency on certificate verification and allow alternative
8716 lookup methods. X509_STORE based implementations of these two callbacks.
8720 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8721 Modify get_crl() to find a valid (unexpired) CRL if possible.
8725 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8726 this would be called X509_CRL_cmp() but that name is already used by
8727 a function that just compares CRL issuer names. Cache several CRL
8728 extensions in X509_CRL structure and cache CRLDP in X509.
8732 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8733 this maps equivalent X509_NAME structures into a consistent structure.
8734 Name comparison can then be performed rapidly using memcmp().
8738 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8743 * Allow digests to supply their own micalg string for S/MIME type using
8744 the ctrl EVP_MD_CTRL_MICALG.
8748 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8749 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8750 ctrl. It can then customise the structure before and/or after signing
8755 * New function OBJ_add_sigid() to allow application defined signature OIDs
8756 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8757 to free up any added signature OIDs.
8761 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8762 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8763 digest and cipher tables. New options added to openssl utility:
8764 list-message-digest-algorithms and list-cipher-algorithms.
8768 * Change the array representation of binary polynomials: the list
8769 of degrees of non-zero coefficients is now terminated with -1.
8770 Previously it was terminated with 0, which was also part of the
8771 value; thus, the array representation was not applicable to
8772 polynomials where t^0 has coefficient zero. This change makes
8773 the array representation useful in a more general context.
8777 * Various modifications and fixes to SSL/TLS cipher string
8778 handling. For ECC, the code now distinguishes between fixed ECDH
8779 with RSA certificates on the one hand and with ECDSA certificates
8780 on the other hand, since these are separate ciphersuites. The
8781 unused code for Fortezza ciphersuites has been removed.
8783 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8784 (not "ECDHE"). For consistency with the code for DH
8785 certificates, use of ECDH certificates is now considered ECDH
8786 authentication, not RSA or ECDSA authentication (the latter is
8787 merely the CA's signing algorithm and not actively used in the
8790 The temporary ciphersuite alias "ECCdraft" is no longer
8791 available, and ECC ciphersuites are no longer excluded from "ALL"
8792 and "DEFAULT". The following aliases now exist for RFC 4492
8793 ciphersuites, most of these by analogy with the DH case:
8795 kECDHr - ECDH cert, signed with RSA
8796 kECDHe - ECDH cert, signed with ECDSA
8797 kECDH - ECDH cert (signed with either RSA or ECDSA)
8798 kEECDH - ephemeral ECDH
8799 ECDH - ECDH cert or ephemeral ECDH
8805 AECDH - anonymous ECDH
8806 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8810 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8811 Use correct micalg parameters depending on digest(s) in signed message.
8815 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8816 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8820 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8821 an engine to register a method. Add ENGINE lookups for methods and
8822 functional reference processing.
8826 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8827 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8832 * New -resign option to smime utility. This adds one or more signers
8833 to an existing PKCS#7 signedData structure. Also -md option to use an
8834 alternative message digest algorithm for signing.
8838 * Tidy up PKCS#7 routines and add new functions to make it easier to
8839 create PKCS7 structures containing multiple signers. Update smime
8840 application to support multiple signers.
8844 * New -macalg option to pkcs12 utility to allow setting of an alternative
8849 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8850 Reorganize PBE internals to lookup from a static table using NIDs,
8851 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8852 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8853 PRF which will be automatically used with PBES2.
8857 * Replace the algorithm specific calls to generate keys in "req" with the
8862 * Update PKCS#7 enveloped data routines to use new API. This is now
8863 supported by any public key method supporting the encrypt operation. A
8864 ctrl is added to allow the public key algorithm to examine or modify
8865 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8870 * Add a ctrl to asn1 method to allow a public key algorithm to express
8871 a default digest type to use. In most cases this will be SHA1 but some
8872 algorithms (such as GOST) need to specify an alternative digest. The
8873 return value indicates how strong the preference is 1 means optional and
8874 2 is mandatory (that is it is the only supported type). Modify
8875 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8876 use the default md. Update openssl utilities to use the default digest
8877 type for signing if it is not explicitly indicated.
8881 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8882 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8883 signing method from the key type. This effectively removes the link
8884 between digests and public key types.
8888 * Add an OID cross reference table and utility functions. Its purpose is to
8889 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8890 rsaEncryption. This will allow some of the algorithm specific hackery
8891 needed to use the correct OID to be removed.
8895 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8896 structures for PKCS7_sign(). They are now set up by the relevant public
8901 * Add provisional EC pkey method with support for ECDSA and ECDH.
8905 * Add support for key derivation (agreement) in the API, DH method and
8910 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8911 public and private key formats. As a side effect these add additional
8912 command line functionality not previously available: DSA signatures can be
8913 generated and verified using pkeyutl and DH key support and generation in
8920 *Oliver Tappe <zooey@hirschkaefer.de>*
8922 * New make target "install_html_docs" installs HTML renditions of the
8925 *Oliver Tappe <zooey@hirschkaefer.de>*
8927 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8928 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8929 support key and parameter generation and add initial key generation
8930 functionality for RSA.
8934 * Add functions for main EVP_PKEY_method operations. The undocumented
8935 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8936 `EVP_PKEY_{encrypt,decrypt}_old`.
8940 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8941 key API, doesn't do much yet.
8945 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8946 public key algorithms. New option to openssl utility:
8947 "list-public-key-algorithms" to print out info.
8951 * Implement the Supported Elliptic Curves Extension for
8952 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8956 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8957 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8961 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8962 utilities such as rsa, dsa, dsaparam etc except they process any key
8967 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8968 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8969 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8974 * Initial support for pluggable public key ASN1.
8975 De-spaghettify the public key ASN1 handling. Move public and private
8976 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8977 algorithm specific handling to a single module within the relevant
8978 algorithm directory. Add functions to allow (near) opaque processing
8979 of public and private key structures.
8983 * Implement the Supported Point Formats Extension for
8984 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8988 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8989 for the psk identity [hint] and the psk callback functions to the
8990 SSL_SESSION, SSL and SSL_CTX structure.
8993 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8997 SSL_CTX_use_psk_identity_hint
8998 SSL_get_psk_identity_hint
8999 SSL_get_psk_identity
9000 SSL_use_psk_identity_hint
9002 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9004 * Add RFC 3161 compliant time stamp request creation, response generation
9005 and response verification functionality.
9007 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9009 * Add initial support for TLS extensions, specifically for the server_name
9010 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9011 have new members for a host name. The SSL data structure has an
9012 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9013 stored in that context to allow for session resumption, even after the
9014 SSL has been switched to a new SSL_CTX in reaction to a client's
9015 server_name extension.
9017 New functions (subject to change):
9019 SSL_get_servername()
9020 SSL_get_servername_type()
9023 New CTRL codes and macros (subject to change):
9025 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9026 - SSL_CTX_set_tlsext_servername_callback()
9027 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9028 - SSL_CTX_set_tlsext_servername_arg()
9029 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9031 openssl s_client has a new '-servername ...' option.
9033 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9034 '-key2 ...', '-servername_fatal' (subject to change). This allows
9035 testing the HostName extension for a specific single host name ('-cert'
9036 and '-key' remain fallbacks for handshakes without HostName
9037 negotiation). If the unrecognized_name alert has to be sent, this by
9038 default is a warning; it becomes fatal with the '-servername_fatal'
9041 *Peter Sylvester, Remy Allais, Christophe Renou*
9043 * Whirlpool hash implementation is added.
9047 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9048 bn(64,32). Because of instruction set limitations it doesn't have
9049 any negative impact on performance. This was done mostly in order
9050 to make it possible to share assembler modules, such as bn_mul_mont
9051 implementations, between 32- and 64-bit builds without hassle.
9055 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9056 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9061 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9062 dedicated Montgomery multiplication procedure, is introduced.
9063 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9064 "64-bit" performance on certain 32-bit targets.
9068 * New option SSL_OP_NO_COMP to disable use of compression selectively
9069 in SSL structures. New SSL ctrl to set maximum send fragment size.
9070 Save memory by setting the I/O buffer sizes dynamically instead of
9071 using the maximum available value.
9075 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9076 in addition to the text details.
9080 * Very, very preliminary EXPERIMENTAL support for printing of general
9081 ASN1 structures. This currently produces rather ugly output and doesn't
9082 handle several customised structures at all.
9086 * Integrated support for PVK file format and some related formats such
9087 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9088 these in the 'rsa' and 'dsa' utilities.
9092 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9096 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9097 place for the (very old) "NETSCAPE" format certificates which are now
9098 handled using new ASN1 code equivalents.
9102 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9103 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9104 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9108 * Modify CRL distribution points extension code to print out previously
9109 unsupported fields. Enhance extension setting code to allow setting of
9114 * Add print and set support for Issuing Distribution Point CRL extension.
9118 * Change 'Configure' script to enable Camellia by default.
9125 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9127 * When rejecting SSL/TLS records due to an incorrect version number, never
9128 update s->server with a new major version number. As of
9129 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9130 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9131 the previous behavior could result in a read attempt at NULL when
9132 receiving specific incorrect SSL/TLS records once record payload
9133 protection is active. ([CVE-2010-0740])
9135 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9137 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9138 could be crashed if the relevant tables were not present (e.g. chrooted).
9140 *Tomas Hoger <thoger@redhat.com>*
9142 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9144 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9146 *Martin Olsson, Neel Mehta*
9148 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9149 accommodate for stack sorting, always a write lock!).
9153 * On some versions of WIN32 Heap32Next is very slow. This can cause
9154 excessive delays in the RAND_poll(): over a minute. As a workaround
9155 include a time check in the inner Heap32Next loop too.
9159 * The code that handled flushing of data in SSL/TLS originally used the
9160 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9161 the problem outlined in PR#1949. The fix suggested there however can
9162 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9163 of Apache). So instead simplify the code to flush unconditionally.
9164 This should be fine since flushing with no data to flush is a no op.
9168 * Handle TLS versions 2.0 and later properly and correctly use the
9169 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9170 off ancient servers have a habit of sticking around for a while...
9174 * Modify compression code so it frees up structures without using the
9175 ex_data callbacks. This works around a problem where some applications
9176 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9177 restarting) then use compression (e.g. SSL with compression) later.
9178 This results in significant per-connection memory leaks and
9179 has caused some security issues including CVE-2008-1678 and
9184 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9185 change when encrypting or decrypting.
9189 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9190 connect and renegotiate with servers which do not support RI.
9191 Until RI is more widely deployed this option is enabled by default.
9195 * Add "missing" ssl ctrls to clear options and mode.
9199 * If client attempts to renegotiate and doesn't support RI respond with
9200 a no_renegotiation alert as required by RFC5746. Some renegotiating
9201 TLS clients will continue a connection gracefully when they receive
9202 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9203 waiting for a server hello which it will never receive. Now we treat a
9204 received no_renegotiation alert as a fatal error. This is because
9205 applications requesting a renegotiation might well expect it to succeed
9206 and would have no code in place to handle the server denying it so the
9207 only safe thing to do is to terminate the connection.
9211 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9212 peer supports secure renegotiation and 0 otherwise. Print out peer
9213 renegotiation support in s_client/s_server.
9217 * Replace the highly broken and deprecated SPKAC certification method with
9218 the updated NID creation version. This should correctly handle UTF8.
9222 * Implement RFC5746. Re-enable renegotiation but require the extension
9223 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9224 turns out to be a bad idea. It has been replaced by
9225 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9226 SSL_CTX_set_options(). This is really not recommended unless you
9227 know what you are doing.
9229 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9231 * Fixes to stateless session resumption handling. Use initial_ctx when
9232 issuing and attempting to decrypt tickets in case it has changed during
9233 servername handling. Use a non-zero length session ID when attempting
9234 stateless session resumption: this makes it possible to determine if
9235 a resumption has occurred immediately after receiving server hello
9236 (several places in OpenSSL subtly assume this) instead of later in
9241 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9242 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9243 fixes for a few places where the return code is not checked
9246 *Julia Lawall <julia@diku.dk>*
9248 * Add --strict-warnings option to Configure script to include devteam
9249 warnings in other configurations.
9253 * Add support for --libdir option and LIBDIR variable in makefiles. This
9254 makes it possible to install openssl libraries in locations which
9255 have names other than "lib", for example "/usr/lib64" which some
9258 *Steve Henson, based on patch from Jeremy Utley*
9260 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9261 X690 8.9.12 and can produce some misleading textual output of OIDs.
9263 *Steve Henson, reported by Dan Kaminsky*
9265 * Delete MD2 from algorithm tables. This follows the recommendation in
9266 several standards that it is not used in new applications due to
9267 several cryptographic weaknesses. For binary compatibility reasons
9268 the MD2 API is still compiled in by default.
9272 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9277 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9278 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9281 *Guenter <lists@gknw.net>*
9283 * Fix the server certificate chain building code to use X509_verify_cert(),
9284 it used to have an ad-hoc builder which was unable to cope with anything
9285 other than a simple chain.
9287 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9289 * Don't check self signed certificate signatures in X509_verify_cert()
9290 by default (a flag can override this): it just wastes time without
9291 adding any security. As a useful side effect self signed root CAs
9292 with non-FIPS digests are now usable in FIPS mode.
9296 * In dtls1_process_out_of_seq_message() the check if the current message
9297 is already buffered was missing. For every new message was memory
9298 allocated, allowing an attacker to perform an denial of service attack
9299 with sending out of seq handshake messages until there is no memory
9300 left. Additionally every future message was buffered, even if the
9301 sequence number made no sense and would be part of another handshake.
9302 So only messages with sequence numbers less than 10 in advance will be
9303 buffered. ([CVE-2009-1378])
9305 *Robin Seggelmann, discovered by Daniel Mentz*
9307 * Records are buffered if they arrive with a future epoch to be
9308 processed after finishing the corresponding handshake. There is
9309 currently no limitation to this buffer allowing an attacker to perform
9310 a DOS attack with sending records with future epochs until there is no
9311 memory left. This patch adds the pqueue_size() function to determine
9312 the size of a buffer and limits the record buffer to 100 entries.
9315 *Robin Seggelmann, discovered by Daniel Mentz*
9317 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9318 parent structure is freed. ([CVE-2009-1379])
9322 * Handle non-blocking I/O properly in SSL_shutdown() call.
9324 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9326 * Add `2.5.4.*` OIDs
9328 *Ilya O. <vrghost@gmail.com>*
9330 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9332 * Disable renegotiation completely - this fixes a severe security
9333 problem ([CVE-2009-3555]) at the cost of breaking all
9334 renegotiation. Renegotiation can be re-enabled by setting
9335 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9336 run-time. This is really not recommended unless you know what
9341 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9343 * Don't set val to NULL when freeing up structures, it is freed up by
9344 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9345 zeroing past the valid field. ([CVE-2009-0789])
9347 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9349 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9350 checked correctly. This would allow some invalid signed attributes to
9351 appear to verify correctly. ([CVE-2009-0591])
9353 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9355 * Reject UniversalString and BMPString types with invalid lengths. This
9356 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9357 a legal length. ([CVE-2009-0590])
9361 * Set S/MIME signing as the default purpose rather than setting it
9362 unconditionally. This allows applications to override it at the store
9367 * Permit restricted recursion of ASN1 strings. This is needed in practice
9368 to handle some structures.
9372 * Improve efficiency of mem_gets: don't search whole buffer each time
9375 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9377 * New -hex option for openssl rand.
9381 * Print out UTF8String and NumericString when parsing ASN1.
9385 * Support NumericString type for name components.
9389 * Allow CC in the environment to override the automatically chosen
9390 compiler. Note that nothing is done to ensure flags work with the
9395 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9397 * Properly check EVP_VerifyFinal() and similar return values
9400 *Ben Laurie, Bodo Moeller, Google Security Team*
9402 * Enable TLS extensions by default.
9406 * Allow the CHIL engine to be loaded, whether the application is
9407 multithreaded or not. (This does not release the developer from the
9408 obligation to set up the dynamic locking callbacks.)
9410 *Sander Temme <sander@temme.net>*
9412 * Use correct exit code if there is an error in dgst command.
9414 *Steve Henson; problem pointed out by Roland Dirlewanger*
9416 * Tweak Configure so that you need to say "experimental-jpake" to enable
9417 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9421 * Add experimental JPAKE support, including demo authentication in
9422 s_client and s_server.
9426 * Set the comparison function in v3_addr_canonize().
9428 *Rob Austein <sra@hactrn.net>*
9430 * Add support for XMPP STARTTLS in s_client.
9432 *Philip Paeps <philip@freebsd.org>*
9434 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9435 to ensure that even with this option, only ciphersuites in the
9436 server's preference list will be accepted. (Note that the option
9437 applies only when resuming a session, so the earlier behavior was
9438 just about the algorithm choice for symmetric cryptography.)
9442 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9444 * Fix NULL pointer dereference if a DTLS server received
9445 ChangeCipherSpec as first record ([CVE-2009-1386]).
9449 * Fix a state transition in s3_srvr.c and d1_srvr.c
9450 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9454 * The fix in 0.9.8c that supposedly got rid of unsafe
9455 double-checked locking was incomplete for RSA blinding,
9456 addressing just one layer of what turns out to have been
9457 doubly unsafe triple-checked locking.
9459 So now fix this for real by retiring the MONT_HELPER macro
9460 in crypto/rsa/rsa_eay.c.
9462 *Bodo Moeller; problem pointed out by Marius Schilder*
9464 * Various precautionary measures:
9466 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9468 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9469 (NB: This would require knowledge of the secret session ticket key
9470 to exploit, in which case you'd be SOL either way.)
9472 - Change bn_nist.c so that it will properly handle input BIGNUMs
9473 outside the expected range.
9475 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9478 *Neel Mehta, Bodo Moeller*
9480 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9481 the load fails. Useful for distros.
9483 *Ben Laurie and the FreeBSD team*
9485 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9489 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9493 * Expand ENGINE to support engine supplied SSL client certificate functions.
9495 This work was sponsored by Logica.
9499 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9500 keystores. Support for SSL/TLS client authentication too.
9501 Not compiled unless enable-capieng specified to Configure.
9503 This work was sponsored by Logica.
9507 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9508 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9509 attribute creation routines such as certificate requests and PKCS#12
9514 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9516 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9517 handshake which could lead to a client crash as found using the
9518 Codenomicon TLS test suite ([CVE-2008-1672])
9520 *Steve Henson, Mark Cox*
9522 * Fix double free in TLS server name extensions which could lead to
9523 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9527 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9529 Clear the error queue to ensure that error entries left from
9530 older function calls do not interfere with the correct operation.
9532 *Lutz Jaenicke, Erik de Castro Lopo*
9534 * Remove root CA certificates of commercial CAs:
9536 The OpenSSL project does not recommend any specific CA and does not
9537 have any policy with respect to including or excluding any CA.
9538 Therefore it does not make any sense to ship an arbitrary selection
9539 of root CA certificates with the OpenSSL software.
9543 * RSA OAEP patches to fix two separate invalid memory reads.
9544 The first one involves inputs when 'lzero' is greater than
9545 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9546 before the beginning of from). The second one involves inputs where
9547 the 'db' section contains nothing but zeroes (there is a one-byte
9548 invalid read after the end of 'db').
9550 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9552 * Partial backport from 0.9.9-dev:
9554 Introduce bn_mul_mont (dedicated Montgomery multiplication
9555 procedure) as a candidate for BIGNUM assembler implementation.
9556 While 0.9.9-dev uses assembler for various architectures, only
9557 x86_64 is available by default here in the 0.9.8 branch, and
9558 32-bit x86 is available through a compile-time setting.
9560 To try the 32-bit x86 assembler implementation, use Configure
9561 option "enable-montasm" (which exists only for this backport).
9563 As "enable-montasm" for 32-bit x86 disclaims code stability
9564 anyway, in this constellation we activate additional code
9565 backported from 0.9.9-dev for further performance improvements,
9566 namely BN_from_montgomery_word. (To enable this otherwise,
9567 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9569 *Andy Polyakov (backport partially by Bodo Moeller)*
9571 * Add TLS session ticket callback. This allows an application to set
9572 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9573 values. This is useful for key rollover for example where several key
9574 sets may exist with different names.
9578 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9579 This was broken until now in 0.9.8 releases, such that the only way
9580 a registered ENGINE could be used (assuming it initialises
9581 successfully on the host) was to explicitly set it as the default
9582 for the relevant algorithms. This is in contradiction with 0.9.7
9583 behaviour and the documentation. With this fix, when an ENGINE is
9584 registered into a given algorithm's table of implementations, the
9585 'uptodate' flag is reset so that auto-discovery will be used next
9586 time a new context for that algorithm attempts to select an
9589 *Ian Lister (tweaked by Geoff Thorpe)*
9591 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9592 implementation in the following ways:
9594 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9597 Lack of BER streaming support means one pass streaming processing is
9598 only supported if data is detached: setting the streaming flag is
9599 ignored for embedded content.
9601 CMS support is disabled by default and must be explicitly enabled
9602 with the enable-cms configuration option.
9606 * Update the GMP engine glue to do direct copies between BIGNUM and
9607 mpz_t when openssl and GMP use the same limb size. Otherwise the
9608 existing "conversion via a text string export" trick is still used.
9610 *Paul Sheer <paulsheer@gmail.com>*
9612 * Zlib compression BIO. This is a filter BIO which compressed and
9613 uncompresses any data passed through it.
9617 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9618 RFC3394 compatible AES key wrapping.
9622 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9623 sets string data without copying. X509_ALGOR_set0() and
9624 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9625 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9626 from an X509_ATTRIBUTE structure optionally checking it occurs only
9627 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9632 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9633 to get the expected BN_FLG_CONSTTIME behavior.
9635 *Bodo Moeller (Google)*
9639 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9640 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9641 - added some more tests to do_tests.pl
9642 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9643 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9644 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9645 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9646 - various changes to netware.pl to enable gcc-cross builds on Win32
9648 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9649 - various changes to fix missing prototype warnings
9650 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9651 - added AES, WHIRLPOOL and CPUID assembler code to build files
9652 - added missing AES assembler make rules to mk1mf.pl
9653 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9655 *Guenter Knauf <eflash@gmx.net>*
9657 * Implement certificate status request TLS extension defined in RFC3546.
9658 A client can set the appropriate parameters and receive the encoded
9659 OCSP response via a callback. A server can query the supplied parameters
9660 and set the encoded OCSP response in the callback. Add simplified examples
9661 to s_client and s_server.
9665 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9668 + Binary incompatibility of ssl_ctx_st structure
9669 + DTLS interoperation with non-compliant servers
9670 + Don't call get_session_cb() without proposed session
9671 + Fix ia64 assembler code
9673 *Andy Polyakov, Steve Henson*
9675 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9677 * DTLS Handshake overhaul. There were longstanding issues with
9678 OpenSSL DTLS implementation, which were making it impossible for
9679 RFC 4347 compliant client to communicate with OpenSSL server.
9680 Unfortunately just fixing these incompatibilities would "cut off"
9681 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9682 server keeps tolerating non RFC compliant syntax. The opposite is
9683 not true, 0.9.8f client can not communicate with earlier server.
9684 This update even addresses CVE-2007-4995.
9688 * Changes to avoid need for function casts in OpenSSL: some compilers
9689 (gcc 4.2 and later) reject their use.
9690 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9693 * Add RFC4507 support to OpenSSL. This includes the corrections in
9694 RFC4507bis. The encrypted ticket format is an encrypted encoded
9695 SSL_SESSION structure, that way new session features are automatically
9698 If a client application caches session in an SSL_SESSION structure
9699 support is transparent because tickets are now stored in the encoded
9702 The SSL_CTX structure automatically generates keys for ticket
9703 protection in servers so again support should be possible
9704 with no application modification.
9706 If a client or server wishes to disable RFC4507 support then the option
9707 SSL_OP_NO_TICKET can be set.
9709 Add a TLS extension debugging callback to allow the contents of any client
9710 or server extensions to be examined.
9712 This work was sponsored by Google.
9716 * Add initial support for TLS extensions, specifically for the server_name
9717 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9718 have new members for a host name. The SSL data structure has an
9719 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9720 stored in that context to allow for session resumption, even after the
9721 SSL has been switched to a new SSL_CTX in reaction to a client's
9722 server_name extension.
9724 New functions (subject to change):
9726 SSL_get_servername()
9727 SSL_get_servername_type()
9730 New CTRL codes and macros (subject to change):
9732 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9733 - SSL_CTX_set_tlsext_servername_callback()
9734 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9735 - SSL_CTX_set_tlsext_servername_arg()
9736 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9738 openssl s_client has a new '-servername ...' option.
9740 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9741 '-key2 ...', '-servername_fatal' (subject to change). This allows
9742 testing the HostName extension for a specific single host name ('-cert'
9743 and '-key' remain fallbacks for handshakes without HostName
9744 negotiation). If the unrecognized_name alert has to be sent, this by
9745 default is a warning; it becomes fatal with the '-servername_fatal'
9748 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9750 * Add AES and SSE2 assembly language support to VC++ build.
9754 * Mitigate attack on final subtraction in Montgomery reduction.
9758 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9759 (which previously caused an internal error).
9763 * Squeeze another 10% out of IGE mode when in != out.
9767 * AES IGE mode speedup.
9769 *Dean Gaudet (Google)*
9771 * Add the Korean symmetric 128-bit cipher SEED (see
9772 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9773 add SEED ciphersuites from RFC 4162:
9775 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9776 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9777 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9778 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9780 To minimize changes between patchlevels in the OpenSSL 0.9.8
9781 series, SEED remains excluded from compilation unless OpenSSL
9782 is configured with 'enable-seed'.
9784 *KISA, Bodo Moeller*
9786 * Mitigate branch prediction attacks, which can be practical if a
9787 single processor is shared, allowing a spy process to extract
9788 information. For detailed background information, see
9789 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9790 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9791 and Necessary Software Countermeasures"). The core of the change
9792 are new versions BN_div_no_branch() and
9793 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9794 respectively, which are slower, but avoid the security-relevant
9795 conditional branches. These are automatically called by BN_div()
9796 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9797 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9798 remove a conditional branch.
9800 BN_FLG_CONSTTIME is the new name for the previous
9801 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9802 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9803 in the exponent causes BN_mod_exp_mont() to use the alternative
9804 implementation in BN_mod_exp_mont_consttime().) The old name
9805 remains as a deprecated alias.
9807 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9808 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9809 constant-time implementations for more than just exponentiation.
9810 Here too the old name is kept as a deprecated alias.
9812 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9813 the BN_BLINDING structure gets an independent copy of the
9814 modulus. This means that the previous `BIGNUM *m` argument to
9815 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9816 essentially becomes `const BIGNUM *m`, although we can't actually
9817 change this in the header file before 0.9.9. It allows
9818 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9819 enable BN_FLG_CONSTTIME.
9821 *Matthew D Wood (Intel Corp)*
9823 * In the SSL/TLS server implementation, be strict about session ID
9824 context matching (which matters if an application uses a single
9825 external cache for different purposes). Previously,
9826 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9827 set. This did ensure strict client verification, but meant that,
9828 with applications using a single external cache for quite
9829 different requirements, clients could circumvent ciphersuite
9830 restrictions for a given session ID context by starting a session
9831 in a different context.
9835 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9836 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9837 authentication-only ciphersuites.
9841 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9842 not complete and could lead to a possible single byte overflow
9843 ([CVE-2007-5135]) [Ben Laurie]
9845 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9847 * Since AES128 and AES256 (and similarly Camellia128 and
9848 Camellia256) share a single mask bit in the logic of
9849 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9850 kludge to work properly if AES128 is available and AES256 isn't
9851 (or if Camellia128 is available and Camellia256 isn't).
9855 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9856 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9857 When a point or a seed is encoded in a BIT STRING, we need to
9858 prevent the removal of trailing zero bits to get the proper DER
9859 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9860 of a NamedBitList, for which trailing 0 bits need to be removed.)
9864 * Have SSL/TLS server implementation tolerate "mismatched" record
9865 protocol version while receiving ClientHello even if the
9866 ClientHello is fragmented. (The server can't insist on the
9867 particular protocol version it has chosen before the ServerHello
9868 message has informed the client about his choice.)
9872 * Add RFC 3779 support.
9874 *Rob Austein for ARIN, Ben Laurie*
9876 * Load error codes if they are not already present instead of using a
9877 static variable. This allows them to be cleanly unloaded and reloaded.
9878 Improve header file function name parsing.
9882 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9883 or CAPABILITY handshake as required by RFCs.
9887 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9889 * Introduce limits to prevent malicious keys being able to
9890 cause a denial of service. ([CVE-2006-2940])
9892 *Steve Henson, Bodo Moeller*
9894 * Fix ASN.1 parsing of certain invalid structures that can result
9895 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9897 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9898 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9900 * Fix SSL client code which could crash if connecting to a
9901 malicious SSLv2 server. ([CVE-2006-4343])
9903 *Tavis Ormandy and Will Drewry, Google Security Team*
9905 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9906 match only those. Before that, "AES256-SHA" would be interpreted
9907 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9908 the same strength classification in 0.9.7h) as we currently only
9909 have a single AES bit in the ciphersuite description bitmap.
9910 That change, however, also applied to ciphersuite strings such as
9911 "RC4-MD5" that intentionally matched multiple ciphersuites --
9912 namely, SSL 2.0 ciphersuites in addition to the more common ones
9913 from SSL 3.0/TLS 1.0.
9915 So we change the selection algorithm again: Naming an explicit
9916 ciphersuite selects this one ciphersuite, and any other similar
9917 ciphersuite (same bitmap) from *other* protocol versions.
9918 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9919 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9921 Since SSL 2.0 does not have any ciphersuites for which the
9922 128/256 bit distinction would be relevant, this works for now.
9923 The proper fix will be to use different bits for AES128 and
9924 AES256, which would have avoided the problems from the beginning;
9925 however, bits are scarce, so we can only do this in a new release
9926 (not just a patchlevel) when we can change the SSL_CIPHER
9927 definition to split the single 'unsigned long mask' bitmap into
9928 multiple values to extend the available space.
9932 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9934 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9935 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9937 * Add AES IGE and biIGE modes.
9941 * Change the Unix randomness entropy gathering to use poll() when
9942 possible instead of select(), since the latter has some
9943 undesirable limitations.
9945 *Darryl Miles via Richard Levitte and Bodo Moeller*
9947 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9948 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9949 cannot be implicitly activated as part of, e.g., the "AES" alias.
9950 However, please upgrade to OpenSSL 0.9.9[-dev] for
9951 non-experimental use of the ECC ciphersuites to get TLS extension
9952 support, which is required for curve and point format negotiation
9953 to avoid potential handshake problems.
9957 * Disable rogue ciphersuites:
9959 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9960 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9961 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9963 The latter two were purportedly from
9964 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9967 Also deactivate the remaining ciphersuites from
9968 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9969 unofficial, and the ID has long expired.
9973 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9974 dual-core machines) and other potential thread-safety issues.
9978 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9979 versions), which is now available for royalty-free use
9980 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9981 Also, add Camellia TLS ciphersuites from RFC 4132.
9983 To minimize changes between patchlevels in the OpenSSL 0.9.8
9984 series, Camellia remains excluded from compilation unless OpenSSL
9985 is configured with 'enable-camellia'.
9989 * Disable the padding bug check when compression is in use. The padding
9990 bug check assumes the first packet is of even length, this is not
9991 necessarily true if compression is enabled and can result in false
9992 positives causing handshake failure. The actual bug test is ancient
9993 code so it is hoped that implementations will either have fixed it by
9994 now or any which still have the bug do not support compression.
9998 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10000 * When applying a cipher rule check to see if string match is an explicit
10001 cipher suite and only match that one cipher suite if it is.
10005 * Link in manifests for VC++ if needed.
10007 *Austin Ziegler <halostatue@gmail.com>*
10009 * Update support for ECC-based TLS ciphersuites according to
10010 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10011 TLS extensions, which are supported starting with the 0.9.9
10012 branch, not in the OpenSSL 0.9.8 branch).
10016 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10017 opaque EVP_CIPHER_CTX handling.
10021 * Fixes and enhancements to zlib compression code. We now only use
10022 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10023 to conform with the standards mentioned here:
10024 <http://www.zlib.net/DLL_FAQ.txt>
10025 Static zlib linking now works on Windows and the new --with-zlib-include
10026 --with-zlib-lib options to Configure can be used to supply the location
10027 of the headers and library. Gracefully handle case where zlib library
10032 * Several fixes and enhancements to the OID generation code. The old code
10033 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10034 handle numbers larger than ULONG_MAX, truncated printing and had a
10035 non standard OBJ_obj2txt() behaviour.
10039 * Add support for building of engines under engine/ as shared libraries
10040 under VC++ build system.
10044 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10045 Hopefully, we will not see any false combination of paths any more.
10049 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10051 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10052 (part of SSL_OP_ALL). This option used to disable the
10053 countermeasure against man-in-the-middle protocol-version
10054 rollback in the SSL 2.0 server implementation, which is a bad
10055 idea. ([CVE-2005-2969])
10057 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10058 for Information Security, National Institute of Advanced Industrial
10059 Science and Technology [AIST], Japan)*
10061 * Add two function to clear and return the verify parameter flags.
10065 * Keep cipherlists sorted in the source instead of sorting them at
10066 runtime, thus removing the need for a lock.
10070 * Avoid some small subgroup attacks in Diffie-Hellman.
10072 *Nick Mathewson and Ben Laurie*
10074 * Add functions for well-known primes.
10078 * Extended Windows CE support.
10080 *Satoshi Nakamura and Andy Polyakov*
10082 * Initialize SSL_METHOD structures at compile time instead of during
10083 runtime, thus removing the need for a lock.
10087 * Make PKCS7_decrypt() work even if no certificate is supplied by
10088 attempting to decrypt each encrypted key in turn. Add support to
10093 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10095 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10098 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10102 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10103 key into the same file any more.
10107 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10111 * Add -utf8 command line and config file option to 'ca'.
10113 *Stefan <stf@udoma.org*
10115 * Removed the macro des_crypt(), as it seems to conflict with some
10116 libraries. Use DES_crypt().
10120 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10121 involves renaming the source and generated shared-libs for
10122 both. The engines will accept the corrected or legacy ids
10123 ('ncipher' and '4758_cca' respectively) when binding. NB,
10124 this only applies when building 'shared'.
10126 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10128 * Add attribute functions to EVP_PKEY structure. Modify
10129 PKCS12_create() to recognize a CSP name attribute and
10130 use it. Make -CSP option work again in pkcs12 utility.
10134 * Add new functionality to the bn blinding code:
10135 - automatic re-creation of the BN_BLINDING parameters after
10136 a fixed number of uses (currently 32)
10137 - add new function for parameter creation
10138 - introduce flags to control the update behaviour of the
10139 BN_BLINDING parameters
10140 - hide BN_BLINDING structure
10141 Add a second BN_BLINDING slot to the RSA structure to improve
10142 performance when a single RSA object is shared among several
10147 * Add support for DTLS.
10149 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10151 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10152 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10156 * Remove buggy and incomplete DH cert support from
10157 ssl/ssl_rsa.c and ssl/s3_both.c
10161 * Use SHA-1 instead of MD5 as the default digest algorithm for
10162 the `apps/openssl` commands.
10166 * Compile clean with "-Wall -Wmissing-prototypes
10167 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10168 DEBUG_SAFESTACK must also be set.
10172 * Change ./Configure so that certain algorithms can be disabled by default.
10173 The new counterpiece to "no-xxx" is "enable-xxx".
10175 The patented RC5 and MDC2 algorithms will now be disabled unless
10176 "enable-rc5" and "enable-mdc2", respectively, are specified.
10178 (IDEA remains enabled despite being patented. This is because IDEA
10179 is frequently required for interoperability, and there is no license
10180 fee for non-commercial use. As before, "no-idea" can be used to
10181 avoid this algorithm.)
10185 * Add processing of proxy certificates (see RFC 3820). This work was
10186 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10187 EGEE (Enabling Grids for E-science in Europe).
10191 * RC4 performance overhaul on modern architectures/implementations, such
10192 as Intel P4, IA-64 and AMD64.
10196 * New utility extract-section.pl. This can be used specify an alternative
10197 section number in a pod file instead of having to treat each file as
10198 a separate case in Makefile. This can be done by adding two lines to the
10201 =for comment openssl_section:XXX
10203 The blank line is mandatory.
10207 * New arguments -certform, -keyform and -pass for s_client and s_server
10208 to allow alternative format key and certificate files and passphrase
10213 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10214 update associated structures and add various utility functions.
10216 Add new policy related verify parameters, include policy checking in
10217 standard verify code. Enhance 'smime' application with extra parameters
10218 to support policy checking and print out.
10222 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10223 Nehemiah processors. These extensions support AES encryption in hardware
10224 as well as RNG (though RNG support is currently disabled).
10226 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10228 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10232 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10234 *Andy Polyakov and a number of other people*
10236 * Improved PowerPC platform support. Most notably BIGNUM assembler
10237 implementation contributed by IBM.
10239 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10241 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10242 exponent rather than 'unsigned long'. There is a corresponding change to
10243 the new 'rsa_keygen' element of the RSA_METHOD structure.
10245 *Jelte Jansen, Geoff Thorpe*
10247 * Functionality for creating the initial serial number file is now
10248 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10250 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10251 number file to 1, which is bound to cause problems. To avoid
10252 the problems while respecting compatibility between different 0.9.7
10253 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10254 CA.pl for serial number initialization. With the new release 0.9.8,
10255 we can fix the problem directly in the 'ca' utility.)
10259 * Reduced header interdependencies by declaring more opaque objects in
10260 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10261 give fewer recursive includes, which could break lazy source code - so
10262 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10263 developers should define this symbol when building and using openssl to
10264 ensure they track the recommended behaviour, interfaces, [etc], but
10265 backwards-compatible behaviour prevails when this isn't defined.
10269 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10273 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10274 This will generate a random key of the appropriate length based on the
10275 cipher context. The EVP_CIPHER can provide its own random key generation
10276 routine to support keys of a specific form. This is used in the des and
10277 3des routines to generate a key of the correct parity. Update S/MIME
10278 code to use new functions and hence generate correct parity DES keys.
10279 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10280 valid (weak or incorrect parity).
10284 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10285 as looking them up. This is useful when the verified structure may contain
10286 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10287 present unless the new PKCS7_NO_CRL flag is asserted.
10291 * Extend ASN1 oid configuration module. It now additionally accepts the
10294 shortName = some long name, 1.2.3.4
10298 * Reimplemented the BN_CTX implementation. There is now no more static
10299 limitation on the number of variables it can handle nor the depth of the
10300 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10301 information can now expand as required, and rather than having a single
10302 static array of bignums, BN_CTX now uses a linked-list of such arrays
10303 allowing it to expand on demand whilst maintaining the usefulness of
10304 BN_CTX's "bundling".
10308 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10309 to allow all RSA operations to function using a single BN_CTX.
10313 * Preliminary support for certificate policy evaluation and checking. This
10314 is initially intended to pass the tests outlined in "Conformance Testing
10315 of Relying Party Client Certificate Path Processing Logic" v1.07.
10319 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10320 remained unused and not that useful. A variety of other little bignum
10321 tweaks and fixes have also been made continuing on from the audit (see
10326 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10327 associated ASN1, EVP and SSL functions and old ASN1 macros.
10331 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10332 and this should never fail. So the return value from the use of
10333 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10334 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10338 * BN_CTX_get() should return zero-valued bignums, providing the same
10339 initialised value as BN_new().
10341 *Geoff Thorpe, suggested by Ulf Möller*
10343 * Support for inhibitAnyPolicy certificate extension.
10347 * An audit of the BIGNUM code is underway, for which debugging code is
10348 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10349 is considered valid when processing BIGNUMs, and causes execution to
10350 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10351 further steps are taken to deliberately pollute unused data in BIGNUM
10352 structures to try and expose faulty code further on. For now, openssl will
10353 (in its default mode of operation) continue to tolerate the inconsistent
10354 forms that it has tolerated in the past, but authors and packagers should
10355 consider trying openssl and their own applications when compiled with
10356 these debugging symbols defined. It will help highlight potential bugs in
10357 their own code, and will improve the test coverage for OpenSSL itself. At
10358 some point, these tighter rules will become openssl's default to improve
10359 maintainability, though the assert()s and other overheads will remain only
10360 in debugging configurations. See bn.h for more details.
10362 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10364 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10365 that can only be obtained through BN_CTX_new() (which implicitly
10366 initialises it). The presence of this function only made it possible
10367 to overwrite an existing structure (and cause memory leaks).
10371 * Because of the callback-based approach for implementing LHASH as a
10372 template type, lh_insert() adds opaque objects to hash-tables and
10373 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10374 to clean up those corresponding objects before destroying the hash table
10375 (and losing the object pointers). So some over-zealous constifications in
10376 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10377 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10378 prototyped to have "const" restrictions on the object pointers they are
10379 given (and so aren't required to cast them away any more).
10383 * The tmdiff.h API was so ugly and minimal that our own timing utility
10384 (speed) prefers to use its own implementation. The two implementations
10385 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10386 its object type properly exposed (MS_TM) instead of casting to/from
10387 `char *`. This may still change yet if someone realises MS_TM and
10389 aren't necessarily the greatest nomenclatures - but this is what was used
10390 internally to the implementation so I've used that for now.
10394 * Ensure that deprecated functions do not get compiled when
10395 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10396 the self-tests were still using deprecated key-generation functions so
10397 these have been updated also.
10401 * Reorganise PKCS#7 code to separate the digest location functionality
10402 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10403 New function PKCS7_set_digest() to set the digest type for PKCS#7
10404 digestedData type. Add additional code to correctly generate the
10405 digestedData type and add support for this type in PKCS7 initialization
10410 * New function PKCS7_set0_type_other() this initializes a PKCS7
10411 structure of type "other".
10415 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10416 sure the loop does correctly stop and breaking ("division by zero")
10417 modulus operations are not performed. The (pre-generated) prime
10418 table crypto/bn/bn_prime.h was already correct, but it could not be
10419 re-generated on some platforms because of the "division by zero"
10420 situation in the script.
10422 *Ralf S. Engelschall*
10424 * Update support for ECC-based TLS ciphersuites according to
10425 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10426 SHA-1 now is only used for "small" curves (where the
10427 representation of a field element takes up to 24 bytes); for
10428 larger curves, the field element resulting from ECDH is directly
10429 used as premaster secret.
10431 *Douglas Stebila (Sun Microsystems Laboratories)*
10433 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10434 curve secp160r1 to the tests.
10436 *Douglas Stebila (Sun Microsystems Laboratories)*
10438 * Add the possibility to load symbols globally with DSO.
10440 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10442 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10443 control of the error stack.
10447 * Add support for STORE in ENGINE.
10451 * Add the STORE type. The intention is to provide a common interface
10452 to certificate and key stores, be they simple file-based stores, or
10453 HSM-type store, or LDAP stores, or...
10454 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10458 * Add a generic structure called OPENSSL_ITEM. This can be used to
10459 pass a list of arguments to any function as well as provide a way
10460 for a function to pass data back to the caller.
10464 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10465 works like BUF_strdup() but can be used to duplicate a portion of
10466 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10471 * Add the function sk_find_ex() which works like sk_find(), but will
10472 return an index to an element even if an exact match couldn't be
10473 found. The index is guaranteed to point at the element where the
10474 searched-for key would be inserted to preserve sorting order.
10478 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10479 takes an extra flags argument for optional functionality. Currently,
10480 the following flags are defined:
10482 OBJ_BSEARCH_VALUE_ON_NOMATCH
10483 This one gets OBJ_bsearch_ex() to return a pointer to the first
10484 element where the comparing function returns a negative or zero
10487 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10488 This one gets OBJ_bsearch_ex() to return a pointer to the first
10489 element where the comparing function returns zero. This is useful
10490 if there are more than one element where the comparing function
10495 * Make it possible to create self-signed certificates with 'openssl ca'
10496 in such a way that the self-signed certificate becomes part of the
10497 CA database and uses the same mechanisms for serial number generation
10498 as all other certificate signing. The new flag '-selfsign' enables
10499 this functionality. Adapt CA.sh and CA.pl.in.
10503 * Add functionality to check the public key of a certificate request
10504 against a given private. This is useful to check that a certificate
10505 request can be signed by that key (self-signing).
10509 * Make it possible to have multiple active certificates with the same
10510 subject in the CA index file. This is done only if the keyword
10511 'unique_subject' is set to 'no' in the main CA section (default
10512 if 'CA_default') of the configuration file. The value is saved
10513 with the database itself in a separate index attribute file,
10514 named like the index file with '.attr' appended to the name.
10518 * Generate multi-valued AVAs using '+' notation in config files for
10523 * Support for nameConstraints certificate extension.
10527 * Support for policyConstraints certificate extension.
10531 * Support for policyMappings certificate extension.
10535 * Make sure the default DSA_METHOD implementation only uses its
10536 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10537 and change its own handlers to be NULL so as to remove unnecessary
10538 indirection. This lets alternative implementations fallback to the
10539 default implementation more easily.
10543 * Support for directoryName in GeneralName related extensions
10548 * Make it possible to link applications using Makefile.shared.
10549 Make that possible even when linking against static libraries!
10553 * Support for single pass processing for S/MIME signing. This now
10554 means that S/MIME signing can be done from a pipe, in addition
10555 cleartext signing (multipart/signed type) is effectively streaming
10556 and the signed data does not need to be all held in memory.
10558 This is done with a new flag PKCS7_STREAM. When this flag is set
10559 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10560 is done after the data is output (and digests calculated) in
10561 SMIME_write_PKCS7().
10565 * Add full support for -rpath/-R, both in shared libraries and
10566 applications, at least on the platforms where it's known how
10571 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10572 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10573 will now compute a table of multiples of the generator that
10574 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10575 faster (notably in the case of a single point multiplication,
10576 scalar * generator).
10578 *Nils Larsch, Bodo Moeller*
10580 * IPv6 support for certificate extensions. The various extensions
10581 which use the IP:a.b.c.d can now take IPv6 addresses using the
10582 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10587 * Added an ENGINE that implements RSA by performing private key
10588 exponentiations with the GMP library. The conversions to and from
10589 GMP's mpz_t format aren't optimised nor are any montgomery forms
10590 cached, and on x86 it appears OpenSSL's own performance has caught up.
10591 However there are likely to be other architectures where GMP could
10592 provide a boost. This ENGINE is not built in by default, but it can be
10593 specified at Configure time and should be accompanied by the necessary
10594 linker additions, eg;
10595 ./config -DOPENSSL_USE_GMP -lgmp
10599 * "openssl engine" will not display ENGINE/DSO load failure errors when
10600 testing availability of engines with "-t" - the old behaviour is
10601 produced by increasing the feature's verbosity with "-tt".
10605 * ECDSA routines: under certain error conditions uninitialized BN objects
10606 could be freed. Solution: make sure initialization is performed early
10607 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10612 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10613 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10614 software implementations. For DSA and DH, parameter generation can
10615 also be overridden by providing the appropriate method callbacks.
10619 * Change the "progress" mechanism used in key-generation and
10620 primality testing to functions that take a new BN_GENCB pointer in
10621 place of callback/argument pairs. The new API functions have `_ex`
10622 postfixes and the older functions are reimplemented as wrappers for
10623 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10624 declarations of the old functions to help (graceful) attempts to
10625 migrate to the new functions. Also, the new key-generation API
10626 functions operate on a caller-supplied key-structure and return
10627 success/failure rather than returning a key or NULL - this is to
10628 help make "keygen" another member function of RSA_METHOD etc.
10630 Example for using the new callback interface:
10632 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10633 void *my_arg = ...;
10636 BN_GENCB_set(&my_cb, my_callback, my_arg);
10638 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10639 /* For the meaning of a, b in calls to my_callback(), see the
10640 * documentation of the function that calls the callback.
10641 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10642 * my_callback should return 1 if it wants BN_is_prime_ex()
10643 * to continue, or 0 to stop.
10648 * Change the ZLIB compression method to be stateful, and make it
10649 available to TLS with the number defined in
10650 draft-ietf-tls-compression-04.txt.
10654 * Add the ASN.1 structures and functions for CertificatePair, which
10655 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10657 CertificatePair ::= SEQUENCE {
10658 forward [0] Certificate OPTIONAL,
10659 reverse [1] Certificate OPTIONAL,
10660 -- at least one of the pair shall be present -- }
10662 Also implement the PEM functions to read and write certificate
10663 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10665 This needed to be defined, mostly for the sake of the LDAP
10666 attribute crossCertificatePair, but may prove useful elsewhere as
10671 * Make it possible to inhibit symlinking of shared libraries in
10672 Makefile.shared, for Cygwin's sake.
10676 * Extend the BIGNUM API by creating a function
10677 void BN_set_negative(BIGNUM *a, int neg);
10678 and a macro that behave like
10679 int BN_is_negative(const BIGNUM *a);
10681 to avoid the need to access 'a->neg' directly in applications.
10685 * Implement fast modular reduction for pseudo-Mersenne primes
10686 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10687 EC_GROUP_new_curve_GFp() will now automatically use this
10690 *Nils Larsch <nla@trustcenter.de>*
10692 * Add new lock type (CRYPTO_LOCK_BN).
10696 * Change the ENGINE framework to automatically load engines
10697 dynamically from specific directories unless they could be
10698 found to already be built in or loaded. Move all the
10699 current engines except for the cryptodev one to a new
10700 directory engines/.
10701 The engines in engines/ are built as shared libraries if
10702 the "shared" options was given to ./Configure or ./config.
10703 Otherwise, they are inserted in libcrypto.a.
10704 /usr/local/ssl/engines is the default directory for dynamic
10705 engines, but that can be overridden at configure time through
10706 the usual use of --prefix and/or --openssldir, and at run
10707 time with the environment variable OPENSSL_ENGINES.
10709 *Geoff Thorpe and Richard Levitte*
10711 * Add Makefile.shared, a helper makefile to build shared
10712 libraries. Adapt Makefile.org.
10716 * Add version info to Win32 DLLs.
10718 *Peter 'Luna' Runestig" <peter@runestig.com>*
10720 * Add new 'medium level' PKCS#12 API. Certificates and keys
10721 can be added using this API to created arbitrary PKCS#12
10722 files while avoiding the low level API.
10724 New options to PKCS12_create(), key or cert can be NULL and
10725 will then be omitted from the output file. The encryption
10726 algorithm NIDs can be set to -1 for no encryption, the mac
10727 iteration count can be set to 0 to omit the mac.
10729 Enhance pkcs12 utility by making the -nokeys and -nocerts
10730 options work when creating a PKCS#12 file. New option -nomac
10731 to omit the mac, NONE can be set for an encryption algorithm.
10732 New code is modified to use the enhanced PKCS12_create()
10733 instead of the low level API.
10737 * Extend ASN1 encoder to support indefinite length constructed
10738 encoding. This can output sequences tags and octet strings in
10739 this form. Modify pk7_asn1.c to support indefinite length
10740 encoding. This is experimental and needs additional code to
10741 be useful, such as an ASN1 bio and some enhanced streaming
10744 Extend template encode functionality so that tagging is passed
10745 down to the template encoder.
10749 * Let 'openssl req' fail if an argument to '-newkey' is not
10750 recognized instead of using RSA as a default.
10754 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10755 As these are not official, they are not included in "ALL";
10756 the "ECCdraft" ciphersuite group alias can be used to select them.
10758 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10760 * Add ECDH engine support.
10762 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10764 * Add ECDH in new directory crypto/ecdh/.
10766 *Douglas Stebila (Sun Microsystems Laboratories)*
10768 * Let BN_rand_range() abort with an error after 100 iterations
10769 without success (which indicates a broken PRNG).
10773 * Change BN_mod_sqrt() so that it verifies that the input value
10774 is really the square of the return value. (Previously,
10775 BN_mod_sqrt would show GIGO behaviour.)
10779 * Add named elliptic curves over binary fields from X9.62, SECG,
10780 and WAP/WTLS; add OIDs that were still missing.
10782 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10784 * Extend the EC library for elliptic curves over binary fields
10785 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10788 EC_GF2m_simple_method
10792 EC_GROUP_new_curve_GF2m
10793 EC_GROUP_set_curve_GF2m
10794 EC_GROUP_get_curve_GF2m
10795 EC_POINT_set_affine_coordinates_GF2m
10796 EC_POINT_get_affine_coordinates_GF2m
10797 EC_POINT_set_compressed_coordinates_GF2m
10799 Point compression for binary fields is disabled by default for
10800 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10803 As binary polynomials are represented as BIGNUMs, various members
10804 of the EC_GROUP and EC_POINT data structures can be shared
10805 between the implementations for prime fields and binary fields;
10806 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10807 are essentially identical to their `..._GFp` counterparts.
10808 (For simplicity, the `..._GFp` prefix has been dropped from
10809 various internal method names.)
10811 An internal 'field_div' method (similar to 'field_mul' and
10812 'field_sqr') has been added; this is used only for binary fields.
10814 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10816 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10817 through methods ('mul', 'precompute_mult').
10819 The generic implementations (now internally called 'ec_wNAF_mul'
10820 and 'ec_wNAF_precomputed_mult') remain the default if these
10821 methods are undefined.
10823 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10825 * New function EC_GROUP_get_degree, which is defined through
10826 EC_METHOD. For curves over prime fields, this returns the bit
10827 length of the modulus.
10829 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10831 * New functions EC_GROUP_dup, EC_POINT_dup.
10832 (These simply call ..._new and ..._copy).
10834 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10836 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10837 Polynomials are represented as BIGNUMs (where the sign bit is not
10838 used) in the following functions [macros]:
10841 BN_GF2m_sub [= BN_GF2m_add]
10842 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10843 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10844 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10846 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10847 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10848 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10849 BN_GF2m_cmp [= BN_ucmp]
10851 (Note that only the 'mod' functions are actually for fields GF(2^m).
10852 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10854 For some functions, an the irreducible polynomial defining a
10855 field can be given as an 'unsigned int[]' with strictly
10856 decreasing elements giving the indices of those bits that are set;
10857 i.e., p[] represents the polynomial
10858 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10860 p[0] > p[1] > ... > p[k] = 0.
10861 This applies to the following functions:
10864 BN_GF2m_mod_mul_arr
10865 BN_GF2m_mod_sqr_arr
10866 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10867 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10868 BN_GF2m_mod_exp_arr
10869 BN_GF2m_mod_sqrt_arr
10870 BN_GF2m_mod_solve_quad_arr
10874 Conversion can be performed by the following functions:
10879 bntest.c has additional tests for binary polynomial arithmetic.
10881 Two implementations for BN_GF2m_mod_div() are available.
10882 The default algorithm simply uses BN_GF2m_mod_inv() and
10883 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10884 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10885 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10887 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10889 * Add new error code 'ERR_R_DISABLED' that can be used when some
10890 functionality is disabled at compile-time.
10892 *Douglas Stebila <douglas.stebila@sun.com>*
10894 * Change default behaviour of 'openssl asn1parse' so that more
10895 information is visible when viewing, e.g., a certificate:
10897 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10898 mode the content of non-printable OCTET STRINGs is output in a
10899 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10900 avoid the appearance of a printable string.
10902 *Nils Larsch <nla@trustcenter.de>*
10904 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10906 EC_GROUP_set_asn1_flag()
10907 EC_GROUP_get_asn1_flag()
10908 EC_GROUP_set_point_conversion_form()
10909 EC_GROUP_get_point_conversion_form()
10910 These control ASN1 encoding details:
10911 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10912 has been set to OPENSSL_EC_NAMED_CURVE.
10913 - Points are encoded in uncompressed form by default; options for
10914 asn1_for are as for point2oct, namely
10915 POINT_CONVERSION_COMPRESSED
10916 POINT_CONVERSION_UNCOMPRESSED
10917 POINT_CONVERSION_HYBRID
10919 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10921 EC_GROUP_set_seed()
10922 EC_GROUP_get0_seed()
10923 EC_GROUP_get_seed_len()
10924 This is used only for ASN1 purposes (so far).
10926 *Nils Larsch <nla@trustcenter.de>*
10928 * Add 'field_type' member to EC_METHOD, which holds the NID
10929 of the appropriate field type OID. The new function
10930 EC_METHOD_get_field_type() returns this value.
10932 *Nils Larsch <nla@trustcenter.de>*
10935 EC_POINT_point2bn()
10936 EC_POINT_bn2point()
10937 EC_POINT_point2hex()
10938 EC_POINT_hex2point()
10939 providing useful interfaces to EC_POINT_point2oct() and
10940 EC_POINT_oct2point().
10942 *Nils Larsch <nla@trustcenter.de>*
10944 * Change internals of the EC library so that the functions
10945 EC_GROUP_set_generator()
10946 EC_GROUP_get_generator()
10947 EC_GROUP_get_order()
10948 EC_GROUP_get_cofactor()
10949 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10950 to methods, which would lead to unnecessary code duplication when
10951 adding different types of curves.
10953 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10955 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10956 arithmetic, and such that modified wNAFs are generated
10957 (which avoid length expansion in many cases).
10961 * Add a function EC_GROUP_check_discriminant() (defined via
10962 EC_METHOD) that verifies that the curve discriminant is non-zero.
10964 Add a function EC_GROUP_check() that makes some sanity tests
10965 on a EC_GROUP, its generator and order. This includes
10966 EC_GROUP_check_discriminant().
10968 *Nils Larsch <nla@trustcenter.de>*
10970 * Add ECDSA in new directory crypto/ecdsa/.
10972 Add applications 'openssl ecparam' and 'openssl ecdsa'
10973 (these are based on 'openssl dsaparam' and 'openssl dsa').
10975 ECDSA support is also included in various other files across the
10976 library. Most notably,
10977 - 'openssl req' now has a '-newkey ecdsa:file' option;
10978 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10979 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10980 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10981 them suitable for ECDSA where domain parameters must be
10982 extracted before the specific public key;
10983 - ECDSA engine support has been added.
10985 *Nils Larsch <nla@trustcenter.de>*
10987 * Include some named elliptic curves, and add OIDs from X9.62,
10988 SECG, and WAP/WTLS. Each curve can be obtained from the new
10990 EC_GROUP_new_by_curve_name(),
10991 and the list of available named curves can be obtained with
10992 EC_get_builtin_curves().
10993 Also add a 'curve_name' member to EC_GROUP objects, which can be
10995 EC_GROUP_set_curve_name()
10996 EC_GROUP_get_curve_name()
10998 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11000 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11001 was actually never needed) and in BN_mul(). The removal in BN_mul()
11002 required a small change in bn_mul_part_recursive() and the addition
11003 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11004 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11005 bn_sub_words() and bn_add_words() except they take arrays with
11010 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11012 * Cleanse PEM buffers before freeing them since they may contain
11015 *Benjamin Bennett <ben@psc.edu>*
11017 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11018 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11019 authentication-only ciphersuites.
11023 * Since AES128 and AES256 share a single mask bit in the logic of
11024 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11025 kludge to work properly if AES128 is available and AES256 isn't.
11029 * Expand security boundary to match 1.1.1 module.
11033 * Remove redundant features: hash file source, editing of test vectors
11034 modify fipsld to use external fips_premain.c signature.
11038 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11039 run algorithm test programs.
11043 * Make algorithm test programs more tolerant of whitespace.
11047 * Have SSL/TLS server implementation tolerate "mismatched" record
11048 protocol version while receiving ClientHello even if the
11049 ClientHello is fragmented. (The server can't insist on the
11050 particular protocol version it has chosen before the ServerHello
11051 message has informed the client about his choice.)
11055 * Load error codes if they are not already present instead of using a
11056 static variable. This allows them to be cleanly unloaded and reloaded.
11060 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11062 * Introduce limits to prevent malicious keys being able to
11063 cause a denial of service. ([CVE-2006-2940])
11065 *Steve Henson, Bodo Moeller*
11067 * Fix ASN.1 parsing of certain invalid structures that can result
11068 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11070 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11071 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11073 * Fix SSL client code which could crash if connecting to a
11074 malicious SSLv2 server. ([CVE-2006-4343])
11076 *Tavis Ormandy and Will Drewry, Google Security Team*
11078 * Change ciphersuite string processing so that an explicit
11079 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11080 will no longer include "AES128-SHA"), and any other similar
11081 ciphersuite (same bitmap) from *other* protocol versions (so that
11082 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11083 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11084 changes from 0.9.8b and 0.9.8d.
11088 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11090 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11091 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11093 * Change the Unix randomness entropy gathering to use poll() when
11094 possible instead of select(), since the latter has some
11095 undesirable limitations.
11097 *Darryl Miles via Richard Levitte and Bodo Moeller*
11099 * Disable rogue ciphersuites:
11101 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11102 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11103 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11105 The latter two were purportedly from
11106 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11109 Also deactivate the remaining ciphersuites from
11110 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11111 unofficial, and the ID has long expired.
11115 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11116 dual-core machines) and other potential thread-safety issues.
11120 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11122 * Adapt fipsld and the build system to link against the validated FIPS
11123 module in FIPS mode.
11127 * Fixes for VC++ 2005 build under Windows.
11131 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11132 from a Windows bash shell such as MSYS. It is autodetected from the
11133 "config" script when run from a VC++ environment. Modify standard VC++
11134 build to use fipscanister.o from the GNU make build.
11138 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11140 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11141 The value now differs depending on if you build for FIPS or not.
11142 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11143 safely run with a non-FIPSed libcrypto, as it may crash because of
11144 the difference induced by this change.
11148 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11150 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11151 (part of SSL_OP_ALL). This option used to disable the
11152 countermeasure against man-in-the-middle protocol-version
11153 rollback in the SSL 2.0 server implementation, which is a bad
11154 idea. ([CVE-2005-2969])
11156 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11157 for Information Security, National Institute of Advanced Industrial
11158 Science and Technology [AIST, Japan)]*
11160 * Minimal support for X9.31 signatures and PSS padding modes. This is
11161 mainly for FIPS compliance and not fully integrated at this stage.
11165 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11166 the exponentiation using a fixed-length exponent. (Otherwise,
11167 the information leaked through timing could expose the secret key
11168 after many signatures; cf. Bleichenbacher's attack on DSA with
11173 * Make a new fixed-window mod_exp implementation the default for
11174 RSA, DSA, and DH private-key operations so that the sequence of
11175 squares and multiplies and the memory access pattern are
11176 independent of the particular secret key. This will mitigate
11177 cache-timing and potential related attacks.
11179 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11180 and this is automatically used by BN_mod_exp_mont() if the new flag
11181 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11182 will use this BN flag for private exponents unless the flag
11183 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11184 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11186 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11188 * Change the client implementation for SSLv23_method() and
11189 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11190 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11191 (Previously, the SSL 2.0 backwards compatible Client Hello
11192 message format would be used even with SSL_OP_NO_SSLv2.)
11196 * Add support for smime-type MIME parameter in S/MIME messages which some
11201 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11202 a threadsafe manner. Modify rsa code to use new function and add calls
11203 to dsa and dh code (which had race conditions before).
11207 * Include the fixed error library code in the C error file definitions
11208 instead of fixing them up at runtime. This keeps the error code
11209 structures constant.
11213 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11215 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11218 * Fixes for newer kerberos headers. NB: the casts are needed because
11219 the 'length' field is signed on one version and unsigned on another
11220 with no (?) obvious way to tell the difference, without these VC++
11221 complains. Also the "definition" of FAR (blank) is no longer included
11222 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11223 some needed definitions.
11227 * Undo Cygwin change.
11231 * Added support for proxy certificates according to RFC 3820.
11232 Because they may be a security thread to unaware applications,
11233 they must be explicitly allowed in run-time. See
11234 docs/HOWTO/proxy_certificates.txt for further information.
11238 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11240 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11241 server and client random values. Previously
11242 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11243 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11245 This change has negligible security impact because:
11247 1. Server and client random values still have 24 bytes of pseudo random
11250 2. Server and client random values are sent in the clear in the initial
11253 3. The master secret is derived using the premaster secret (48 bytes in
11254 size for static RSA ciphersuites) as well as client server and random
11257 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11260 *Stephen Henson, reported by UK NISCC*
11262 * Use Windows randomness collection on Cygwin.
11266 * Fix hang in EGD/PRNGD query when communication socket is closed
11267 prematurely by EGD/PRNGD.
11269 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11271 * Prompt for pass phrases when appropriate for PKCS12 input format.
11275 * Back-port of selected performance improvements from development
11276 branch, as well as improved support for PowerPC platforms.
11280 * Add lots of checks for memory allocation failure, error codes to indicate
11281 failure and freeing up memory if a failure occurs.
11283 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11285 * Add new -passin argument to dgst.
11289 * Perform some character comparisons of different types in X509_NAME_cmp:
11290 this is needed for some certificates that re-encode DNs into UTF8Strings
11291 (in violation of RFC3280) and can't or won't issue name rollover
11296 * Make an explicit check during certificate validation to see that
11297 the CA setting in each certificate on the chain is correct. As a
11298 side effect always do the following basic checks on extensions,
11299 not just when there's an associated purpose to the check:
11301 - if there is an unhandled critical extension (unless the user
11302 has chosen to ignore this fault)
11303 - if the path length has been exceeded (if one is set at all)
11304 - that certain extensions fit the associated purpose (if one has
11309 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11311 * Avoid a race condition when CRLs are checked in a multi threaded
11312 environment. This would happen due to the reordering of the revoked
11313 entries during signature checking and serial number lookup. Now the
11314 encoding is cached and the serial number sort performed under a lock.
11315 Add new STACK function sk_is_sorted().
11319 * Add Delta CRL to the extension code.
11323 * Various fixes to s3_pkt.c so alerts are sent properly.
11325 *David Holmes <d.holmes@f5.com>*
11327 * Reduce the chances of duplicate issuer name and serial numbers (in
11328 violation of RFC3280) using the OpenSSL certificate creation utilities.
11329 This is done by creating a random 64 bit value for the initial serial
11330 number when a serial number file is created or when a self signed
11331 certificate is created using 'openssl req -x509'. The initial serial
11332 number file is created using 'openssl x509 -next_serial' in CA.pl
11333 rather than being initialized to 1.
11337 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11339 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11340 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11342 *Joe Orton, Steve Henson*
11344 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11347 *Joe Orton, Steve Henson*
11349 * Make it possible to have multiple active certificates with the same
11350 subject in the CA index file. This is done only if the keyword
11351 'unique_subject' is set to 'no' in the main CA section (default
11352 if 'CA_default') of the configuration file. The value is saved
11353 with the database itself in a separate index attribute file,
11354 named like the index file with '.attr' appended to the name.
11358 * X509 verify fixes. Disable broken certificate workarounds when
11359 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11360 keyUsage extension present. Don't accept CRLs with unhandled critical
11361 extensions: since verify currently doesn't process CRL extensions this
11362 rejects a CRL with *any* critical extensions. Add new verify error codes
11367 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11368 A clarification of RFC2560 will require the use of OCTET STRINGs and
11369 some implementations cannot handle the current raw format. Since OpenSSL
11370 copies and compares OCSP nonces as opaque blobs without any attempt at
11371 parsing them this should not create any compatibility issues.
11375 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11376 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11377 this HMAC (and other) operations are several times slower than OpenSSL
11382 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11384 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11386 * Use the correct content when signing type "other".
11390 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11392 * Fix various bugs revealed by running the NISCC test suite:
11394 Stop out of bounds reads in the ASN1 code when presented with
11395 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11397 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11399 If verify callback ignores invalid public key errors don't try to check
11400 certificate signature with the NULL public key.
11404 * New -ignore_err option in ocsp application to stop the server
11405 exiting on the first error in a request.
11409 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11410 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11415 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11416 extra data after the compression methods not only for TLS 1.0
11417 but also for SSL 3.0 (as required by the specification).
11419 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11421 * Change X509_certificate_type() to mark the key as exported/exportable
11422 when it's 512 *bits* long, not 512 bytes.
11426 * Change AES_cbc_encrypt() so it outputs exact multiple of
11427 blocks during encryption.
11431 * Various fixes to base64 BIO and non blocking I/O. On write
11432 flushes were not handled properly if the BIO retried. On read
11433 data was not being buffered properly and had various logic bugs.
11434 This also affects blocking I/O when the data being decoded is a
11439 * Various S/MIME bugfixes and compatibility changes:
11440 output correct application/pkcs7 MIME type if
11441 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11442 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11443 of files as .eml work). Correctly handle very long lines in MIME
11448 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11450 * Countermeasure against the Klima-Pokorny-Rosa extension of
11451 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11452 a protocol version number mismatch like a decryption error
11453 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11457 * Turn on RSA blinding by default in the default implementation
11458 to avoid a timing attack. Applications that don't want it can call
11459 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11460 They would be ill-advised to do so in most cases.
11462 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11464 * Change RSA blinding code so that it works when the PRNG is not
11465 seeded (in this case, the secret RSA exponent is abused as
11466 an unpredictable seed -- if it is not unpredictable, there
11467 is no point in blinding anyway). Make RSA blinding thread-safe
11468 by remembering the creator's thread ID in rsa->blinding and
11469 having all other threads use local one-time blinding factors
11470 (this requires more computation than sharing rsa->blinding, but
11471 avoids excessive locking; and if an RSA object is not shared
11472 between threads, blinding will still be very fast).
11476 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11477 ENGINE as defaults for all supported algorithms irrespective of
11478 the 'flags' parameter. 'flags' is now honoured, so applications
11479 should make sure they are passing it correctly.
11483 * Target "mingw" now allows native Windows code to be generated in
11484 the Cygwin environment as well as with the MinGW compiler.
11488 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11490 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11491 via timing by performing a MAC computation even if incorrect
11492 block cipher padding has been found. This is a countermeasure
11493 against active attacks where the attacker has to distinguish
11494 between bad padding and a MAC verification error. ([CVE-2003-0078])
11496 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11497 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11498 Martin Vuagnoux (EPFL, Ilion)*
11500 * Make the no-err option work as intended. The intention with no-err
11501 is not to have the whole error stack handling routines removed from
11502 libcrypto, it's only intended to remove all the function name and
11503 reason texts, thereby removing some of the footprint that may not
11504 be interesting if those errors aren't displayed anyway.
11506 NOTE: it's still possible for any application or module to have its
11507 own set of error texts inserted. The routines are there, just not
11508 used by default when no-err is given.
11512 * Add support for FreeBSD on IA64.
11514 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11516 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11517 Kerberos function mit_des_cbc_cksum(). Before this change,
11518 the value returned by DES_cbc_cksum() was like the one from
11519 mit_des_cbc_cksum(), except the bytes were swapped.
11521 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11523 * Allow an application to disable the automatic SSL chain building.
11524 Before this a rather primitive chain build was always performed in
11525 ssl3_output_cert_chain(): an application had no way to send the
11526 correct chain if the automatic operation produced an incorrect result.
11528 Now the chain builder is disabled if either:
11530 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11532 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11534 The reasoning behind this is that an application would not want the
11535 auto chain building to take place if extra chain certificates are
11536 present and it might also want a means of sending no additional
11537 certificates (for example the chain has two certificates and the
11542 * Add the possibility to build without the ENGINE framework.
11544 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11546 * Under Win32 gmtime() can return NULL: check return value in
11547 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11551 * DSA routines: under certain error conditions uninitialized BN objects
11552 could be freed. Solution: make sure initialization is performed early
11553 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11554 Nils Larsch <nla@trustcenter.de> via PR#459)
11558 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11559 checked on reconnect on the client side, therefore session resumption
11560 could still fail with a "ssl session id is different" error. This
11561 behaviour is masked when SSL_OP_ALL is used due to
11562 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11563 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11564 followup to PR #377.
11568 * IA-32 assembler support enhancements: unified ELF targets, support
11569 for SCO/Caldera platforms, fix for Cygwin shared build.
11573 * Add support for FreeBSD on sparc64. As a consequence, support for
11574 FreeBSD on non-x86 processors is separate from x86 processors on
11575 the config script, much like the NetBSD support.
11577 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11579 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11581 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11584 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11585 code (06) was taken as the first octet of the session ID and the last
11586 octet was ignored consequently. As a result SSLv2 client side session
11587 caching could not have worked due to the session ID mismatch between
11589 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11594 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11595 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11600 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11601 seems that in spite of existing for more than a year, many application
11602 author have done nothing to provide the necessary callbacks, which
11603 means that this particular engine will not work properly anywhere.
11604 This is a very unfortunate situation which forces us, in the name
11605 of usability, to give the hw_ncipher.c a static lock, which is part
11607 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11608 appear in 0.9.8 or later. We EXPECT application authors to have
11609 dealt properly with this when 0.9.8 is released (unless we actually
11610 make such changes in the libcrypto locking code that changes will
11611 have to be made anyway).
11615 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11616 octets have been read, EOF or an error occurs. Without this change
11617 some truncated ASN1 structures will not produce an error.
11621 * Disable Heimdal support, since it hasn't been fully implemented.
11622 Still give the possibility to force the use of Heimdal, but with
11623 warnings and a request that patches get sent to openssl-dev.
11627 * Add the VC-CE target, introduce the WINCE sysname, and add
11628 INSTALL.WCE and appropriate conditionals to make it build.
11630 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11632 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11633 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11634 edit numbers of the version.
11636 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11638 * Introduce safe string copy and catenation functions
11639 (BUF_strlcpy() and BUF_strlcat()).
11641 *Ben Laurie (CHATS) and Richard Levitte*
11643 * Avoid using fixed-size buffers for one-line DNs.
11645 *Ben Laurie (CHATS)*
11647 * Add BUF_MEM_grow_clean() to avoid information leakage when
11648 resizing buffers containing secrets, and use where appropriate.
11650 *Ben Laurie (CHATS)*
11652 * Avoid using fixed size buffers for configuration file location.
11654 *Ben Laurie (CHATS)*
11656 * Avoid filename truncation for various CA files.
11658 *Ben Laurie (CHATS)*
11660 * Use sizeof in preference to magic numbers.
11662 *Ben Laurie (CHATS)*
11664 * Avoid filename truncation in cert requests.
11666 *Ben Laurie (CHATS)*
11668 * Add assertions to check for (supposedly impossible) buffer
11671 *Ben Laurie (CHATS)*
11673 * Don't cache truncated DNS entries in the local cache (this could
11674 potentially lead to a spoofing attack).
11676 *Ben Laurie (CHATS)*
11678 * Fix various buffers to be large enough for hex/decimal
11679 representations in a platform independent manner.
11681 *Ben Laurie (CHATS)*
11683 * Add CRYPTO_realloc_clean() to avoid information leakage when
11684 resizing buffers containing secrets, and use where appropriate.
11686 *Ben Laurie (CHATS)*
11688 * Add BIO_indent() to avoid much slightly worrying code to do
11691 *Ben Laurie (CHATS)*
11693 * Convert sprintf()/BIO_puts() to BIO_printf().
11695 *Ben Laurie (CHATS)*
11697 * buffer_gets() could terminate with the buffer only half
11700 *Ben Laurie (CHATS)*
11702 * Add assertions to prevent user-supplied crypto functions from
11703 overflowing internal buffers by having large block sizes, etc.
11705 *Ben Laurie (CHATS)*
11707 * New OPENSSL_assert() macro (similar to assert(), but enabled
11710 *Ben Laurie (CHATS)*
11712 * Eliminate unused copy of key in RC4.
11714 *Ben Laurie (CHATS)*
11716 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11718 *Ben Laurie (CHATS)*
11720 * Fix off-by-one error in EGD path.
11722 *Ben Laurie (CHATS)*
11724 * If RANDFILE path is too long, ignore instead of truncating.
11726 *Ben Laurie (CHATS)*
11728 * Eliminate unused and incorrectly sized X.509 structure
11731 *Ben Laurie (CHATS)*
11733 * Eliminate unused and dangerous function knumber().
11735 *Ben Laurie (CHATS)*
11737 * Eliminate unused and dangerous structure, KSSL_ERR.
11739 *Ben Laurie (CHATS)*
11741 * Protect against overlong session ID context length in an encoded
11742 session object. Since these are local, this does not appear to be
11745 *Ben Laurie (CHATS)*
11747 * Change from security patch (see 0.9.6e below) that did not affect
11748 the 0.9.6 release series:
11750 Remote buffer overflow in SSL3 protocol - an attacker could
11751 supply an oversized master key in Kerberos-enabled versions.
11754 *Ben Laurie (CHATS)*
11756 * Change the SSL kerb5 codes to match RFC 2712.
11760 * Make -nameopt work fully for req and add -reqopt switch.
11762 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11764 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11766 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11768 * Make sure tests can be performed even if the corresponding algorithms
11769 have been removed entirely. This was also the last step to make
11770 OpenSSL compilable with DJGPP under all reasonable conditions.
11772 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11774 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11775 to allow version independent disabling of normally unselected ciphers,
11776 which may be activated as a side-effect of selecting a single cipher.
11778 (E.g., cipher list string "RSA" enables ciphersuites that are left
11779 out of "ALL" because they do not provide symmetric encryption.
11780 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11782 *Lutz Jaenicke, Bodo Moeller*
11784 * Add appropriate support for separate platform-dependent build
11785 directories. The recommended way to make a platform-dependent
11786 build directory is the following (tested on Linux), maybe with
11789 # Place yourself outside of the OpenSSL source tree. In
11790 # this example, the environment variable OPENSSL_SOURCE
11791 # is assumed to contain the absolute OpenSSL source directory.
11792 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11793 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11794 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11795 mkdir -p `dirname $F`
11796 ln -s $OPENSSL_SOURCE/$F $F
11799 To be absolutely sure not to disturb the source tree, a "make clean"
11800 is a good thing. If it isn't successful, don't worry about it,
11801 it probably means the source directory is very clean.
11805 * Make sure any ENGINE control commands make local copies of string
11806 pointers passed to them whenever necessary. Otherwise it is possible
11807 the caller may have overwritten (or deallocated) the original string
11808 data when a later ENGINE operation tries to use the stored values.
11810 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11812 * Improve diagnostics in file reading and command-line digests.
11814 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11816 * Add AES modes CFB and OFB to the object database. Correct an
11817 error in AES-CFB decryption.
11821 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11822 allows existing EVP_CIPHER_CTX structures to be reused after
11823 calling `EVP_*Final()`. This behaviour is used by encryption
11824 BIOs and some applications. This has the side effect that
11825 applications must explicitly clean up cipher contexts with
11826 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11830 * Check the values of dna and dnb in bn_mul_recursive before calling
11831 bn_mul_comba (a non zero value means the a or b arrays do not contain
11832 n2 elements) and fallback to bn_mul_normal if either is not zero.
11836 * Fix escaping of non-ASCII characters when using the -subj option
11837 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11841 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11842 form for "surname", serialNumber has no short form.
11843 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11844 therefore remove "mail" short name for "internet 7".
11845 The OID for unique identifiers in X509 certificates is
11846 x500UniqueIdentifier, not uniqueIdentifier.
11847 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11851 * Add an "init" command to the ENGINE config module and auto initialize
11852 ENGINEs. Without any "init" command the ENGINE will be initialized
11853 after all ctrl commands have been executed on it. If init=1 the
11854 ENGINE is initialized at that point (ctrls before that point are run
11855 on the uninitialized ENGINE and after on the initialized one). If
11856 init=0 then the ENGINE will not be initialized at all.
11860 * Fix the 'app_verify_callback' interface so that the user-defined
11861 argument is actually passed to the callback: In the
11862 SSL_CTX_set_cert_verify_callback() prototype, the callback
11863 declaration has been changed from
11866 int (*cb)(X509_STORE_CTX *,void *);
11867 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11868 i=s->ctx->app_verify_callback(&ctx)
11869 has been changed into
11870 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11872 To update applications using SSL_CTX_set_cert_verify_callback(),
11873 a dummy argument can be added to their callback functions.
11875 *D. K. Smetters <smetters@parc.xerox.com>*
11877 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11879 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11881 * Add and OPENSSL_LOAD_CONF define which will cause
11882 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11883 This allows older applications to transparently support certain
11884 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11885 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11886 load the config file and OPENSSL_add_all_algorithms_conf() which will
11887 always load it have also been added.
11891 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11892 Adjust NIDs and EVP layer.
11894 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11896 * Config modules support in openssl utility.
11898 Most commands now load modules from the config file,
11899 though in a few (such as version) this isn't done
11900 because it couldn't be used for anything.
11902 In the case of ca and req the config file used is
11903 the same as the utility itself: that is the -config
11904 command line option can be used to specify an
11909 * Move default behaviour from OPENSSL_config(). If appname is NULL
11910 use "openssl_conf" if filename is NULL use default openssl config file.
11914 * Add an argument to OPENSSL_config() to allow the use of an alternative
11915 config section name. Add a new flag to tolerate a missing config file
11916 and move code to CONF_modules_load_file().
11920 * Support for crypto accelerator cards from Accelerated Encryption
11921 Processing, www.aep.ie. (Use engine 'aep')
11922 The support was copied from 0.9.6c [engine] and adapted/corrected
11923 to work with the new engine framework.
11925 *AEP Inc. and Richard Levitte*
11927 * Support for SureWare crypto accelerator cards from Baltimore
11928 Technologies. (Use engine 'sureware')
11929 The support was copied from 0.9.6c [engine] and adapted
11930 to work with the new engine framework.
11934 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11935 make the newer ENGINE framework commands for the CHIL engine work.
11937 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11939 * Make it possible to produce shared libraries on ReliantUNIX.
11941 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11943 * Add the configuration target debug-linux-ppro.
11944 Make 'openssl rsa' use the general key loading routines
11945 implemented in `apps.c`, and make those routines able to
11946 handle the key format FORMAT_NETSCAPE and the variant
11949 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11951 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11953 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11955 * Add -keyform to rsautl, and document -engine.
11957 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11959 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11960 BIO_R_NO_SUCH_FILE error code rather than the generic
11961 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11965 * Add new functions
11966 ERR_peek_last_error
11967 ERR_peek_last_error_line
11968 ERR_peek_last_error_line_data.
11969 These are similar to
11971 ERR_peek_error_line
11972 ERR_peek_error_line_data,
11973 but report on the latest error recorded rather than the first one
11974 still in the error queue.
11976 *Ben Laurie, Bodo Moeller*
11978 * default_algorithms option in ENGINE config module. This allows things
11980 default_algorithms = ALL
11981 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11985 * Preliminary ENGINE config module.
11989 * New experimental application configuration code.
11993 * Change the AES code to follow the same name structure as all other
11994 symmetric ciphers, and behave the same way. Move everything to
11995 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11997 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11999 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12001 *Ben Laurie and Theo de Raadt*
12003 * Add option to output public keys in req command.
12005 *Massimiliano Pala madwolf@openca.org*
12007 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12008 (up to about 10% better than before for P-192 and P-224).
12012 * New functions/macros
12014 SSL_CTX_set_msg_callback(ctx, cb)
12015 SSL_CTX_set_msg_callback_arg(ctx, arg)
12016 SSL_set_msg_callback(ssl, cb)
12017 SSL_set_msg_callback_arg(ssl, arg)
12019 to request calling a callback function
12021 void cb(int write_p, int version, int content_type,
12022 const void *buf, size_t len, SSL *ssl, void *arg)
12024 whenever a protocol message has been completely received
12025 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12026 protocol version according to which the SSL library interprets
12027 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12028 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12029 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12030 specification (change_cipher_spec(20), alert(21), handshake(22)).
12031 'buf' and 'len' point to the actual message, 'ssl' to the
12032 SSL object, and 'arg' is the application-defined value set by
12033 SSL[_CTX]_set_msg_callback_arg().
12035 'openssl s_client' and 'openssl s_server' have new '-msg' options
12036 to enable a callback that displays all protocol messages.
12040 * Change the shared library support so shared libraries are built as
12041 soon as the corresponding static library is finished, and thereby get
12042 openssl and the test programs linked against the shared library.
12043 This still only happens when the keyword "shard" has been given to
12044 the configuration scripts.
12046 NOTE: shared library support is still an experimental thing, and
12047 backward binary compatibility is still not guaranteed.
12049 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12051 * Add support for Subject Information Access extension.
12053 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12055 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12056 additional bytes when new memory had to be allocated, not just
12057 when reusing an existing buffer.
12061 * New command line and configuration option 'utf8' for the req command.
12062 This allows field values to be specified as UTF8 strings.
12066 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12067 runs for the former and machine-readable output for the latter.
12071 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12072 of the e-mail address in the DN (i.e., it will go into a certificate
12073 extension only). The new configuration file option 'email_in_dn = no'
12074 has the same effect.
12076 *Massimiliano Pala madwolf@openca.org*
12078 * Change all functions with names starting with `des_` to be starting
12079 with `DES_` instead. Add wrappers that are compatible with libdes,
12080 but are named `_ossl_old_des_*`. Finally, add macros that map the
12081 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12082 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12083 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12086 Since we provide two compatibility mappings, the user needs to
12087 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12088 compatibility is desired. The default (i.e., when that macro
12089 isn't defined) is OpenSSL 0.9.6c compatibility.
12091 There are also macros that enable and disable the support of old
12092 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12093 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12094 are defined, the default will apply: to support the old des routines.
12096 In either case, one must include openssl/des.h to get the correct
12097 definitions. Do not try to just include openssl/des_old.h, that
12100 NOTE: This is a major break of an old API into a new one. Software
12101 authors are encouraged to switch to the `DES_` style functions. Some
12102 time in the future, des_old.h and the libdes compatibility functions
12103 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12104 default), and then completely removed.
12108 * Test for certificates which contain unsupported critical extensions.
12109 If such a certificate is found during a verify operation it is
12110 rejected by default: this behaviour can be overridden by either
12111 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12112 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12113 X509_supported_extension() has also been added which returns 1 if a
12114 particular extension is supported.
12118 * Modify the behaviour of EVP cipher functions in similar way to digests
12119 to retain compatibility with existing code.
12123 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12124 compatibility with existing code. In particular the 'ctx' parameter does
12125 not have to be to be initialized before the call to EVP_DigestInit() and
12126 it is tidied up after a call to EVP_DigestFinal(). New function
12127 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12128 EVP_MD_CTX_copy() changed to not require the destination to be
12129 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12130 requires the destination to be valid.
12132 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12133 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12137 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12138 so that complete 'Handshake' protocol structures are kept in memory
12139 instead of overwriting 'msg_type' and 'length' with 'body' data.
12143 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12145 *Massimo Santin via Richard Levitte*
12147 * Major restructuring to the underlying ENGINE code. This includes
12148 reduction of linker bloat, separation of pure "ENGINE" manipulation
12149 (initialisation, etc) from functionality dealing with implementations
12150 of specific crypto interfaces. This change also introduces integrated
12151 support for symmetric ciphers and digest implementations - so ENGINEs
12152 can now accelerate these by providing EVP_CIPHER and EVP_MD
12153 implementations of their own. This is detailed in
12154 [crypto/engine/README.md](crypto/engine/README.md)
12155 as it couldn't be adequately described here. However, there are a few
12156 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12157 were changed in the original introduction of ENGINE code have now
12158 reverted back - the hooking from this code to ENGINE is now a good
12159 deal more passive and at run-time, operations deal directly with
12160 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12161 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12162 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12163 they were not being used by the framework as there is no concept of a
12164 BIGNUM_METHOD and they could not be generalised to the new
12165 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12166 ENGINE_cpy() has been removed as it cannot be consistently defined in
12171 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12175 * Change mkdef.pl to sort symbols that get the same entry number,
12176 and make sure the automatically generated functions `ERR_load_*`
12177 become part of libeay.num as well.
12181 * New function SSL_renegotiate_pending(). This returns true once
12182 renegotiation has been requested (either SSL_renegotiate() call
12183 or HelloRequest/ClientHello received from the peer) and becomes
12184 false once a handshake has been completed.
12185 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12186 sends a HelloRequest, but does not ensure that a handshake takes
12187 place. SSL_renegotiate_pending() is useful for checking if the
12188 client has followed the request.)
12192 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12193 By default, clients may request session resumption even during
12194 renegotiation (if session ID contexts permit); with this option,
12195 session resumption is possible only in the first handshake.
12197 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12198 more bits available for options that should not be part of
12199 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12203 * Add some demos for certificate and certificate request creation.
12207 * Make maximum certificate chain size accepted from the peer application
12208 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12209 "Douglas E. Engert" <deengert@anl.gov>.
12213 * Add support for shared libraries for Unixware-7
12214 (Boyd Lynn Gerber <gerberb@zenez.com>).
12218 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12219 be done prior to destruction. Use this to unload error strings from
12220 ENGINEs that load their own error strings. NB: This adds two new API
12221 functions to "get" and "set" this destroy handler in an ENGINE.
12225 * Alter all existing ENGINE implementations (except "openssl" and
12226 "openbsd") to dynamically instantiate their own error strings. This
12227 makes them more flexible to be built both as statically-linked ENGINEs
12228 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12229 Also, add stub code to each that makes building them as self-contained
12230 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12234 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12235 implementations into applications that are completely implemented in
12236 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12237 commands that can be used to configure what shared-library to load and
12238 to control aspects of the way it is handled. Also, made an update to
12239 the [README-Engine.md](README-Engine.md) file
12240 that brings its information up-to-date and
12241 provides some information and instructions on the "dynamic" ENGINE
12242 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12246 * Make it possible to unload ranges of ERR strings with a new
12247 "ERR_unload_strings" function.
12251 * Add a copy() function to EVP_MD.
12255 * Make EVP_MD routines take a context pointer instead of just the
12256 md_data void pointer.
12260 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12261 that the digest can only process a single chunk of data
12262 (typically because it is provided by a piece of
12263 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12264 is only going to provide a single chunk of data, and hence the
12265 framework needn't accumulate the data for oneshot drivers.
12269 * As with "ERR", make it possible to replace the underlying "ex_data"
12270 functions. This change also alters the storage and management of global
12271 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12272 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12273 index counters. The API functions that use this state have been changed
12274 to take a "class_index" rather than pointers to the class's local STACK
12275 and counter, and there is now an API function to dynamically create new
12276 classes. This centralisation allows us to (a) plug a lot of the
12277 thread-safety problems that existed, and (b) makes it possible to clean
12278 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12279 such data would previously have always leaked in application code and
12280 workarounds were in place to make the memory debugging turn a blind eye
12281 to it. Application code that doesn't use this new function will still
12282 leak as before, but their memory debugging output will announce it now
12283 rather than letting it slide.
12285 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12286 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12287 has a return value to indicate success or failure.
12291 * Make it possible to replace the underlying "ERR" functions such that the
12292 global state (2 LHASH tables and 2 locks) is only used by the "default"
12293 implementation. This change also adds two functions to "get" and "set"
12294 the implementation prior to it being automatically set the first time
12295 any other ERR function takes place. Ie. an application can call "get",
12296 pass the return value to a module it has just loaded, and that module
12297 can call its own "set" function using that value. This means the
12298 module's "ERR" operations will use (and modify) the error state in the
12299 application and not in its own statically linked copy of OpenSSL code.
12303 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12304 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12305 the operation, and provides a more encapsulated way for external code
12306 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12307 to use these functions rather than manually incrementing the counts.
12309 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12313 * Add EVP test program.
12317 * Add symmetric cipher support to ENGINE. Expect the API to change!
12321 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12322 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12323 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12324 These allow a CRL to be built without having to access X509_CRL fields
12325 directly. Modify 'ca' application to use new functions.
12329 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12330 bug workarounds. Rollback attack detection is a security feature.
12331 The problem will only arise on OpenSSL servers when TLSv1 is not
12332 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12333 Software authors not wanting to support TLSv1 will have special reasons
12334 for their choice and can explicitly enable this option.
12336 *Bodo Moeller, Lutz Jaenicke*
12338 * Rationalise EVP so it can be extended: don't include a union of
12339 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12340 (similar to those existing for EVP_CIPHER_CTX).
12345 EVP_MD_CTX_init(&md); /* new function call */
12346 EVP_DigestInit(&md, EVP_sha1());
12347 EVP_DigestUpdate(&md, in, len);
12348 EVP_DigestFinal(&md, out, NULL);
12349 EVP_MD_CTX_cleanup(&md); /* new function call */
12353 * Make DES key schedule conform to the usual scheme, as well as
12354 correcting its structure. This means that calls to DES functions
12355 now have to pass a pointer to a des_key_schedule instead of a
12356 plain des_key_schedule (which was actually always a pointer
12359 des_key_schedule ks;
12361 des_set_key_checked(..., &ks);
12362 des_ncbc_encrypt(..., &ks, ...);
12364 (Note that a later change renames 'des_...' into 'DES_...'.)
12368 * Initial reduction of linker bloat: the use of some functions, such as
12369 PEM causes large amounts of unused functions to be linked in due to
12370 poor organisation. For example pem_all.c contains every PEM function
12371 which has a knock on effect of linking in large amounts of (unused)
12372 ASN1 code. Grouping together similar functions and splitting unrelated
12373 functions prevents this.
12377 * Cleanup of EVP macros.
12381 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12382 correct `_ecb suffix`.
12386 * Add initial OCSP responder support to ocsp application. The
12387 revocation information is handled using the text based index
12388 use by the ca application. The responder can either handle
12389 requests generated internally, supplied in files (for example
12390 via a CGI script) or using an internal minimal server.
12394 * Add configuration choices to get zlib compression for TLS.
12398 * Changes to Kerberos SSL for RFC 2712 compliance:
12399 1. Implemented real KerberosWrapper, instead of just using
12400 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12401 2. Implemented optional authenticator field of KerberosWrapper.
12403 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12404 and authenticator structs; see crypto/krb5/.
12406 Generalized Kerberos calls to support multiple Kerberos libraries.
12407 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12408 via Richard Levitte*
12410 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12411 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12412 values for each of the key sizes rather than having just
12413 parameters (and 'speed' generating keys each time).
12417 * Speed up EVP routines.
12420 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12421 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12422 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12423 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12425 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12426 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12427 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12430 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12432 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12436 * Added the OS2-EMX target.
12438 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12440 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12441 New functions to support `NCONF `routines in extension code.
12442 New function `CONF_set_nconf()`
12443 to allow functions which take an `NCONF` to also handle the old `LHASH`
12444 structure: this means that the old `CONF` compatible routines can be
12445 retained (in particular w.rt. extensions) without having to duplicate the
12446 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12450 * Enhance the general user interface with mechanisms for inner control
12451 and with possibilities to have yes/no kind of prompts.
12455 * Change all calls to low level digest routines in the library and
12456 applications to use EVP. Add missing calls to HMAC_cleanup() and
12457 don't assume HMAC_CTX can be copied using memcpy().
12459 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12461 * Add the possibility to control engines through control names but with
12462 arbitrary arguments instead of just a string.
12463 Change the key loaders to take a UI_METHOD instead of a callback
12464 function pointer. NOTE: this breaks binary compatibility with earlier
12465 versions of OpenSSL [engine].
12466 Adapt the nCipher code for these new conditions and add a card insertion
12471 * Enhance the general user interface with mechanisms to better support
12472 dialog box interfaces, application-defined prompts, the possibility
12473 to use defaults (for example default passwords from somewhere else)
12474 and interrupts/cancellations.
12478 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12479 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12483 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12484 tidy up some unnecessarily weird code in 'sk_new()').
12486 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12488 * Change the key loading routines for ENGINEs to use the same kind
12489 callback (pem_password_cb) as all other routines that need this
12494 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12495 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12496 than this minimum value is recommended.
12500 * New random seeder for OpenVMS, using the system process statistics
12501 that are easily reachable.
12505 * Windows apparently can't transparently handle global
12506 variables defined in DLLs. Initialisations such as:
12508 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12510 won't compile. This is used by the any applications that need to
12511 declare their own ASN1 modules. This was fixed by adding the option
12512 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12513 needed for static libraries under Win32.
12517 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12518 setting of purpose and trust fields. New X509_STORE trust and
12519 purpose functions and tidy up setting in other SSL functions.
12523 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12524 structure. These are inherited by X509_STORE_CTX when it is
12525 initialised. This allows various defaults to be set in the
12526 X509_STORE structure (such as flags for CRL checking and custom
12527 purpose or trust settings) for functions which only use X509_STORE_CTX
12528 internally such as S/MIME.
12530 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12531 trust settings if they are not set in X509_STORE. This allows X509_STORE
12532 purposes and trust (in S/MIME for example) to override any set by default.
12534 Add command line options for CRL checking to smime, s_client and s_server
12539 * Initial CRL based revocation checking. If the CRL checking flag(s)
12540 are set then the CRL is looked up in the X509_STORE structure and
12541 its validity and signature checked, then if the certificate is found
12542 in the CRL the verify fails with a revoked error.
12544 Various new CRL related callbacks added to X509_STORE_CTX structure.
12546 Command line options added to 'verify' application to support this.
12548 This needs some additional work, such as being able to handle multiple
12549 CRLs with different times, extension based lookup (rather than just
12550 by subject name) and ultimately more complete V2 CRL extension
12555 * Add a general user interface API (crypto/ui/). This is designed
12556 to replace things like des_read_password and friends (backward
12557 compatibility functions using this new API are provided).
12558 The purpose is to remove prompting functions from the DES code
12559 section as well as provide for prompting through dialog boxes in
12560 a window system and the like.
12564 * Add "ex_data" support to ENGINE so implementations can add state at a
12565 per-structure level rather than having to store it globally.
12569 * Make it possible for ENGINE structures to be copied when retrieved by
12570 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12571 This causes the "original" ENGINE structure to act like a template,
12572 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12573 operational state can be localised to each ENGINE structure, despite the
12574 fact they all share the same "methods". New ENGINE structures returned in
12575 this case have no functional references and the return value is the single
12576 structural reference. This matches the single structural reference returned
12577 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12582 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12583 needs to match any other type at all we need to manually clear the
12588 * Changes to the "openssl engine" utility to include;
12589 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12590 about an ENGINE's available control commands.
12591 - executing control commands from command line arguments using the
12592 '-pre' and '-post' switches. '-post' is only used if '-t' is
12593 specified and the ENGINE is successfully initialised. The syntax for
12594 the individual commands are colon-separated, for example;
12595 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12599 * New dynamic control command support for ENGINEs. ENGINEs can now
12600 declare their own commands (numbers), names (strings), descriptions,
12601 and input types for run-time discovery by calling applications. A
12602 subset of these commands are implicitly classed as "executable"
12603 depending on their input type, and only these can be invoked through
12604 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12605 can be based on user input, config files, etc). The distinction is
12606 that "executable" commands cannot return anything other than a boolean
12607 result and can only support numeric or string input, whereas some
12608 discoverable commands may only be for direct use through
12609 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12610 pointers, or other custom uses. The "executable" commands are to
12611 support parameterisations of ENGINE behaviour that can be
12612 unambiguously defined by ENGINEs and used consistently across any
12613 OpenSSL-based application. Commands have been added to all the
12614 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12615 control over shared-library paths without source code alterations.
12619 * Changed all ENGINE implementations to dynamically allocate their
12620 ENGINEs rather than declaring them statically. Apart from this being
12621 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12622 this also allows the implementations to compile without using the
12623 internal engine_int.h header.
12627 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12628 'const' value. Any code that should be able to modify a RAND_METHOD
12629 should already have non-const pointers to it (ie. they should only
12630 modify their own ones).
12634 * Made a variety of little tweaks to the ENGINE code.
12635 - "atalla" and "ubsec" string definitions were moved from header files
12636 to C code. "nuron" string definitions were placed in variables
12637 rather than hard-coded - allowing parameterisation of these values
12638 later on via ctrl() commands.
12639 - Removed unused "#if 0"'d code.
12640 - Fixed engine list iteration code so it uses ENGINE_free() to release
12641 structural references.
12642 - Constified the RAND_METHOD element of ENGINE structures.
12643 - Constified various get/set functions as appropriate and added
12644 missing functions (including a catch-all ENGINE_cpy that duplicates
12645 all ENGINE values onto a new ENGINE except reference counts/state).
12646 - Removed NULL parameter checks in get/set functions. Setting a method
12647 or function to NULL is a way of cancelling out a previously set
12648 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12649 and doesn't justify the extra error symbols and code.
12650 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12651 flags from engine_int.h to engine.h.
12652 - Changed prototypes for ENGINE handler functions (init(), finish(),
12653 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12657 * Implement binary inversion algorithm for BN_mod_inverse in addition
12658 to the algorithm using long division. The binary algorithm can be
12659 used only if the modulus is odd. On 32-bit systems, it is faster
12660 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12661 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12662 up to 450 bits. In 64-bit environments, the binary algorithm
12663 appears to be advantageous for much longer moduli; here we use it
12664 for moduli up to 2048 bits.
12668 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12669 could not support the combine flag in choice fields.
12673 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12674 extensions from a certificate request to the certificate.
12678 * Allow multiple 'certopt' and 'nameopt' options to be separated
12679 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12680 file: this allows the display of the certificate about to be
12681 signed to be customised, to allow certain fields to be included
12682 or excluded and extension details. The old system didn't display
12683 multicharacter strings properly, omitted fields not in the policy
12684 and couldn't display additional details such as extensions.
12688 * Function EC_POINTs_mul for multiple scalar multiplication
12689 of an arbitrary number of elliptic curve points
12690 \sum scalars[i]*points[i],
12691 optionally including the generator defined for the EC_GROUP:
12692 scalar*generator + \sum scalars[i]*points[i].
12694 EC_POINT_mul is a simple wrapper function for the typical case
12695 that the point list has just one item (besides the optional
12700 * First EC_METHODs for curves over GF(p):
12702 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12703 operations and provides various method functions that can also
12704 operate with faster implementations of modular arithmetic.
12706 EC_GFp_mont_method() reuses most functions that are part of
12707 EC_GFp_simple_method, but uses Montgomery arithmetic.
12709 *Bodo Moeller; point addition and point doubling
12710 implementation directly derived from source code provided by
12711 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12713 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12714 crypto/ec/ec_lib.c):
12716 Curves are EC_GROUP objects (with an optional group generator)
12717 based on EC_METHODs that are built into the library.
12719 Points are EC_POINT objects based on EC_GROUP objects.
12721 Most of the framework would be able to handle curves over arbitrary
12722 finite fields, but as there are no obvious types for fields other
12723 than GF(p), some functions are limited to that for now.
12727 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12728 that the file contains a complete HTTP response.
12732 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12733 change the def and num file printf format specifier from "%-40sXXX"
12734 to "%-39s XXX". The latter will always guarantee a space after the
12735 field while the former will cause them to run together if the field
12736 is 40 of more characters long.
12740 * Constify the cipher and digest 'method' functions and structures
12741 and modify related functions to take constant EVP_MD and EVP_CIPHER
12746 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12747 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12751 * Modify `EVP_Digest*()` routines so they now return values. Although the
12752 internal software routines can never fail additional hardware versions
12757 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12759 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12760 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12763 ERR_R_NESTED_ASN1_ERROR
12765 ERR_R_MISSING_ASN1_EOS
12766 were 4 .. 9, conflicting with
12767 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12769 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12770 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12772 Add new error code 'ERR_R_INTERNAL_ERROR'.
12776 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12781 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12782 sets the subject name for a new request or supersedes the
12783 subject name in a given request. Formats that can be parsed are
12784 'CN=Some Name, OU=myOU, C=IT'
12786 'CN=Some Name/OU=myOU/C=IT'.
12788 Add options '-batch' and '-verbose' to 'openssl req'.
12790 *Massimiliano Pala <madwolf@hackmasters.net>*
12792 * Introduce the possibility to access global variables through
12793 functions on platform were that's the best way to handle exporting
12794 global variables in shared libraries. To enable this functionality,
12795 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12796 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12797 is normally done by Configure or something similar).
12799 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12800 in the source file (foo.c) like this:
12802 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12803 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12805 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12806 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12808 OPENSSL_DECLARE_GLOBAL(int,foo);
12809 #define foo OPENSSL_GLOBAL_REF(foo)
12810 OPENSSL_DECLARE_GLOBAL(double,bar);
12811 #define bar OPENSSL_GLOBAL_REF(bar)
12813 The #defines are very important, and therefore so is including the
12814 header file everywhere where the defined globals are used.
12816 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12817 of ASN.1 items, but that structure is a bit different.
12819 The largest change is in util/mkdef.pl which has been enhanced with
12820 better and easier to understand logic to choose which symbols should
12821 go into the Windows .def files as well as a number of fixes and code
12822 cleanup (among others, algorithm keywords are now sorted
12823 lexicographically to avoid constant rewrites).
12827 * In BN_div() keep a copy of the sign of 'num' before writing the
12828 result to 'rm' because if rm==num the value will be overwritten
12829 and produce the wrong result if 'num' is negative: this caused
12830 problems with BN_mod() and BN_nnmod().
12834 * Function OCSP_request_verify(). This checks the signature on an
12835 OCSP request and verifies the signer certificate. The signer
12836 certificate is just checked for a generic purpose and OCSP request
12841 * Add OCSP_check_validity() function to check the validity of OCSP
12842 responses. OCSP responses are prepared in real time and may only
12843 be a few seconds old. Simply checking that the current time lies
12844 between thisUpdate and nextUpdate max reject otherwise valid responses
12845 caused by either OCSP responder or client clock inaccuracy. Instead
12846 we allow thisUpdate and nextUpdate to fall within a certain period of
12847 the current time. The age of the response can also optionally be
12848 checked. Two new options -validity_period and -status_age added to
12853 * If signature or public key algorithm is unrecognized print out its
12854 OID rather that just UNKNOWN.
12858 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12859 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12860 ID to be generated from the issuer certificate alone which can then be
12861 passed to OCSP_id_issuer_cmp().
12865 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12866 ASN1 modules to export functions returning ASN1_ITEM pointers
12867 instead of the ASN1_ITEM structures themselves. This adds several
12868 new macros which allow the underlying ASN1 function/structure to
12869 be accessed transparently. As a result code should not use ASN1_ITEM
12870 references directly (such as &X509_it) but instead use the relevant
12871 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12872 use of the new ASN1 code on platforms where exporting structures
12873 is problematical (for example in shared libraries) but exporting
12874 functions returning pointers to structures is not.
12878 * Add support for overriding the generation of SSL/TLS session IDs.
12879 These callbacks can be registered either in an SSL_CTX or per SSL.
12880 The purpose of this is to allow applications to control, if they wish,
12881 the arbitrary values chosen for use as session IDs, particularly as it
12882 can be useful for session caching in multiple-server environments. A
12883 command-line switch for testing this (and any client code that wishes
12884 to use such a feature) has been added to "s_server".
12886 *Geoff Thorpe, Lutz Jaenicke*
12888 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12889 of the form `#if defined(...) || defined(...) || ...` and
12890 `#if !defined(...) && !defined(...) && ...`. This also avoids
12891 the growing number of special cases it was previously handling.
12895 * Make all configuration macros available for application by making
12896 sure they are available in opensslconf.h, by giving them names starting
12897 with `OPENSSL_` to avoid conflicts with other packages and by making
12898 sure e_os2.h will cover all platform-specific cases together with
12900 Additionally, it is now possible to define configuration/platform-
12901 specific names (called "system identities"). In the C code, these
12902 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12903 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12904 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12909 * New option -set_serial to 'req' and 'x509' this allows the serial
12910 number to use to be specified on the command line. Previously self
12911 signed certificates were hard coded with serial number 0 and the
12912 CA options of 'x509' had to use a serial number in a file which was
12917 * New options to 'ca' utility to support V2 CRL entry extensions.
12918 Currently CRL reason, invalidity date and hold instruction are
12919 supported. Add new CRL extensions to V3 code and some new objects.
12923 * New function EVP_CIPHER_CTX_set_padding() this is used to
12924 disable standard block padding (aka PKCS#5 padding) in the EVP
12925 API, which was previously mandatory. This means that the data is
12926 not padded in any way and so the total length much be a multiple
12927 of the block size, otherwise an error occurs.
12931 * Initial (incomplete) OCSP SSL support.
12935 * New function OCSP_parse_url(). This splits up a URL into its host,
12936 port and path components: primarily to parse OCSP URLs. New -url
12937 option to ocsp utility.
12941 * New nonce behavior. The return value of OCSP_check_nonce() now
12942 reflects the various checks performed. Applications can decide
12943 whether to tolerate certain situations such as an absent nonce
12944 in a response when one was present in a request: the ocsp application
12945 just prints out a warning. New function OCSP_add1_basic_nonce()
12946 this is to allow responders to include a nonce in a response even if
12947 the request is nonce-less.
12951 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12952 skipped when using openssl x509 multiple times on a single input file,
12953 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12957 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12958 set string type: to handle setting ASN1_TIME structures. Fix ca
12959 utility to correctly initialize revocation date of CRLs.
12963 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12964 the clients preferred ciphersuites and rather use its own preferences.
12965 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12966 Internet Explorer by ensuring unchanged hash method during stepup.
12967 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12971 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12972 to aes and add a new 'exist' option to print out symbols that don't
12977 * Additional options to ocsp utility to allow flags to be set and
12978 additional certificates supplied.
12982 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12983 OCSP client a number of certificate to only verify the response
12988 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12989 handle the new API. Currently only ECB, CBC modes supported. Add new
12992 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12993 Encryption Standard (AES) Ciphersuites for Transport Layer
12994 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12995 not enabled by default and were not part of the "ALL" ciphersuite
12996 alias because they were not yet official; they could be
12997 explicitly requested by specifying the "AESdraft" ciphersuite
12998 group alias. In the final release of OpenSSL 0.9.7, the group
12999 alias is called "AES" and is part of "ALL".)
13001 *Ben Laurie, Steve Henson, Bodo Moeller*
13003 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13004 request to response.
13008 * Functions for OCSP responders. OCSP_request_onereq_count(),
13009 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13010 extract information from a certificate request. OCSP_response_create()
13011 creates a response and optionally adds a basic response structure.
13012 OCSP_basic_add1_status() adds a complete single response to a basic
13013 response and returns the OCSP_SINGLERESP structure just added (to allow
13014 extensions to be included for example). OCSP_basic_add1_cert() adds a
13015 certificate to a basic response and OCSP_basic_sign() signs a basic
13016 response with various flags. New helper functions ASN1_TIME_check()
13017 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13018 (converts ASN1_TIME to GeneralizedTime).
13022 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13023 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13024 structure from a certificate. X509_pubkey_digest() digests the public_key
13025 contents: this is used in various key identifiers.
13029 * Make sk_sort() tolerate a NULL argument.
13031 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13033 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13034 passed by the function are trusted implicitly. If any of them signed the
13035 response then it is assumed to be valid and is not verified.
13039 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13040 to data. This was previously part of the PKCS7 ASN1 code. This
13041 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13042 *Steve Henson, reported by Kenneth R. Robinette
13043 <support@securenetterm.com>*
13045 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13046 routines: without these tracing memory leaks is very painful.
13047 Fix leaks in PKCS12 and PKCS7 routines.
13051 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13052 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13053 effectively meant GeneralizedTime would never be used. Now it
13054 is initialised to -1 but X509_time_adj() now has to check the value
13055 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13056 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13057 *Steve Henson, reported by Kenneth R. Robinette
13058 <support@securenetterm.com>*
13060 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13061 result in a zero length in the ASN1_INTEGER structure which was
13062 not consistent with the structure when d2i_ASN1_INTEGER() was used
13063 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13064 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13065 where it did not print out a minus for negative ASN1_INTEGER.
13069 * Add summary printout to ocsp utility. The various functions which
13070 convert status values to strings have been renamed to:
13071 OCSP_response_status_str(), OCSP_cert_status_str() and
13072 OCSP_crl_reason_str() and are no longer static. New options
13073 to verify nonce values and to disable verification. OCSP response
13074 printout format cleaned up.
13078 * Add additional OCSP certificate checks. These are those specified
13079 in RFC2560. This consists of two separate checks: the CA of the
13080 certificate being checked must either be the OCSP signer certificate
13081 or the issuer of the OCSP signer certificate. In the latter case the
13082 OCSP signer certificate must contain the OCSP signing extended key
13083 usage. This check is performed by attempting to match the OCSP
13084 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13085 in the OCSP_CERTID structures of the response.
13089 * Initial OCSP certificate verification added to OCSP_basic_verify()
13090 and related routines. This uses the standard OpenSSL certificate
13091 verify routines to perform initial checks (just CA validity) and
13092 to obtain the certificate chain. Then additional checks will be
13093 performed on the chain. Currently the root CA is checked to see
13094 if it is explicitly trusted for OCSP signing. This is used to set
13095 a root CA as a global signing root: that is any certificate that
13096 chains to that CA is an acceptable OCSP signing certificate.
13100 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13101 extensions from a separate configuration file.
13102 As when reading extensions from the main configuration file,
13103 the '-extensions ...' option may be used for specifying the
13106 *Massimiliano Pala <madwolf@comune.modena.it>*
13108 * New OCSP utility. Allows OCSP requests to be generated or
13109 read. The request can be sent to a responder and the output
13110 parsed, outputted or printed in text form. Not complete yet:
13111 still needs to check the OCSP response validity.
13115 * New subcommands for 'openssl ca':
13116 `openssl ca -status <serial>` prints the status of the cert with
13117 the given serial number (according to the index file).
13118 `openssl ca -updatedb` updates the expiry status of certificates
13121 *Massimiliano Pala <madwolf@comune.modena.it>*
13123 * New '-newreq-nodes' command option to CA.pl. This is like
13124 '-newreq', but calls 'openssl req' with the '-nodes' option
13125 so that the resulting key is not encrypted.
13127 *Damien Miller <djm@mindrot.org>*
13129 * New configuration for the GNU Hurd.
13131 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13133 * Initial code to implement OCSP basic response verify. This
13134 is currently incomplete. Currently just finds the signer's
13135 certificate and verifies the signature on the response.
13139 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13140 value of OPENSSLDIR. This is available via the new '-d' option
13141 to 'openssl version', and is also included in 'openssl version -a'.
13145 * Allowing defining memory allocation callbacks that will be given
13146 file name and line number information in additional arguments
13147 (a `const char*` and an int). The basic functionality remains, as
13148 well as the original possibility to just replace malloc(),
13149 realloc() and free() by functions that do not know about these
13150 additional arguments. To register and find out the current
13151 settings for extended allocation functions, the following
13152 functions are provided:
13154 CRYPTO_set_mem_ex_functions
13155 CRYPTO_set_locked_mem_ex_functions
13156 CRYPTO_get_mem_ex_functions
13157 CRYPTO_get_locked_mem_ex_functions
13159 These work the same way as CRYPTO_set_mem_functions and friends.
13160 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13161 extended allocation function is enabled.
13162 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13163 a conventional allocation function is enabled.
13165 *Richard Levitte, Bodo Moeller*
13167 * Finish off removing the remaining LHASH function pointer casts.
13168 There should no longer be any prototype-casting required when using
13169 the LHASH abstraction, and any casts that remain are "bugs". See
13170 the callback types and macros at the head of lhash.h for details
13171 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13175 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13176 If /dev/[u]random devices are not available or do not return enough
13177 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13179 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13180 /etc/entropy will be queried once each in this sequence, querying stops
13181 when enough entropy was collected without querying more sockets.
13185 * Change the Unix RAND_poll() variant to be able to poll several
13186 random devices, as specified by DEVRANDOM, until a sufficient amount
13187 of data has been collected. We spend at most 10 ms on each file
13188 (select timeout) and read in non-blocking mode. DEVRANDOM now
13189 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13190 (previously it was just the string "/dev/urandom"), so on typical
13191 platforms the 10 ms delay will never occur.
13192 Also separate out the Unix variant to its own file, rand_unix.c.
13193 For VMS, there's a currently-empty rand_vms.c.
13197 * Move OCSP client related routines to ocsp_cl.c. These
13198 provide utility functions which an application needing
13199 to issue a request to an OCSP responder and analyse the
13200 response will typically need: as opposed to those which an
13201 OCSP responder itself would need which will be added later.
13203 OCSP_request_sign() signs an OCSP request with an API similar
13204 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13205 response. OCSP_response_get1_basic() extracts basic response
13206 from response. OCSP_resp_find_status(): finds and extracts status
13207 information from an OCSP_CERTID structure (which will be created
13208 when the request structure is built). These are built from lower
13209 level functions which work on OCSP_SINGLERESP structures but
13210 won't normally be used unless the application wishes to examine
13211 extensions in the OCSP response for example.
13213 Replace nonce routines with a pair of functions.
13214 OCSP_request_add1_nonce() adds a nonce value and optionally
13215 generates a random value. OCSP_check_nonce() checks the
13216 validity of the nonce in an OCSP response.
13220 * Change function OCSP_request_add() to OCSP_request_add0_id().
13221 This doesn't copy the supplied OCSP_CERTID and avoids the
13222 need to free up the newly created id. Change return type
13223 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13224 This can then be used to add extensions to the request.
13225 Deleted OCSP_request_new(), since most of its functionality
13226 is now in OCSP_REQUEST_new() (and the case insensitive name
13227 clash) apart from the ability to set the request name which
13228 will be added elsewhere.
13232 * Update OCSP API. Remove obsolete extensions argument from
13233 various functions. Extensions are now handled using the new
13234 OCSP extension code. New simple OCSP HTTP function which
13235 can be used to send requests and parse the response.
13239 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13240 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13241 uses the special reorder version of SET OF to sort the attributes
13242 and reorder them to match the encoded order. This resolves a long
13243 standing problem: a verify on a PKCS7 structure just after signing
13244 it used to fail because the attribute order did not match the
13245 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13246 it uses the received order. This is necessary to tolerate some broken
13247 software that does not order SET OF. This is handled by encoding
13248 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13249 to produce the required SET OF.
13253 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13254 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13255 files to get correct declarations of the ASN.1 item variables.
13259 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13260 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13261 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13262 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13263 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13264 ASN1_ITEM and no wrapper functions.
13268 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13269 replace the old function pointer based I/O routines. Change most of
13270 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13274 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13275 lines, recognize more "algorithms" that can be deselected, and make
13276 it complain about algorithm deselection that isn't recognised.
13280 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13281 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13282 to use new functions. Add NO_ASN1_OLD which can be set to remove
13283 some old style ASN1 functions: this can be used to determine if old
13284 code will still work when these eventually go away.
13288 * New extension functions for OCSP structures, these follow the
13289 same conventions as certificates and CRLs.
13293 * New function X509V3_add1_i2d(). This automatically encodes and
13294 adds an extension. Its behaviour can be customised with various
13295 flags to append, replace or delete. Various wrappers added for
13296 certificates and CRLs.
13300 * Fix to avoid calling the underlying ASN1 print routine when
13301 an extension cannot be parsed. Correct a typo in the
13302 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13306 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13307 entries for variables.
13311 * Add functionality to `apps/openssl.c` for detecting locking
13312 problems: As the program is single-threaded, all we have
13313 to do is register a locking callback using an array for
13314 storing which locks are currently held by the program.
13318 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13319 SSL_get_ex_data_X509_STORE_idx(), which is used in
13320 ssl_verify_cert_chain() and thus can be called at any time
13321 during TLS/SSL handshakes so that thread-safety is essential.
13322 Unfortunately, the ex_data design is not at all suited
13323 for multi-threaded use, so it probably should be abolished.
13327 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13329 *Broadcom, tweaked and integrated by Geoff Thorpe*
13331 * Move common extension printing code to new function
13332 X509V3_print_extensions(). Reorganise OCSP print routines and
13333 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13337 * New function X509_signature_print() to remove duplication in some
13342 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13343 set (this was treated exactly the same as SET OF previously). This
13344 is used to reorder the STACK representing the structure to match the
13345 encoding. This will be used to get round a problem where a PKCS7
13346 structure which was signed could not be verified because the STACK
13347 order did not reflect the encoded order.
13351 * Reimplement the OCSP ASN1 module using the new code.
13355 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13356 for its ASN1 operations. The old style function pointers still exist
13357 for now but they will eventually go away.
13361 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13362 completely replaces the old ASN1 functionality with a table driven
13363 encoder and decoder which interprets an ASN1_ITEM structure describing
13364 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13365 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13366 has also been converted to the new form.
13370 * Change BN_mod_exp_recp so that negative moduli are tolerated
13371 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13372 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13373 for negative moduli.
13377 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13378 of not touching the result's sign bit.
13382 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13387 * Changed the LHASH code to use prototypes for callbacks, and created
13388 macros to declare and implement thin (optionally static) functions
13389 that provide type-safety and avoid function pointer casting for the
13390 type-specific callbacks.
13394 * Added Kerberos Cipher Suites to be used with TLS, as written in
13396 *Veers Staats <staatsvr@asc.hpc.mil>,
13397 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13399 * Reformat the FAQ so the different questions and answers can be divided
13400 in sections depending on the subject.
13404 * Have the zlib compression code load ZLIB.DLL dynamically under
13409 * New function BN_mod_sqrt for computing square roots modulo a prime
13410 (using the probabilistic Tonelli-Shanks algorithm unless
13411 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13412 be handled deterministically).
13414 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13416 * Make BN_mod_inverse faster by explicitly handling small quotients
13417 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13418 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13422 * New function BN_kronecker.
13426 * Fix BN_gcd so that it works on negative inputs; the result is
13427 positive unless both parameters are zero.
13428 Previously something reasonably close to an infinite loop was
13429 possible because numbers could be growing instead of shrinking
13430 in the implementation of Euclid's algorithm.
13434 * Fix BN_is_word() and BN_is_one() macros to take into account the
13435 sign of the number in question.
13437 Fix BN_is_word(a,w) to work correctly for w == 0.
13439 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13440 because its test if the absolute value of 'a' equals 'w'.
13441 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13442 it exists mostly for use in the implementations of BN_is_zero(),
13443 BN_is_one(), and BN_is_word().
13447 * New function BN_swap.
13451 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13452 the exponentiation functions are more likely to produce reasonable
13453 results on negative inputs.
13457 * Change BN_mod_mul so that the result is always non-negative.
13458 Previously, it could be negative if one of the factors was negative;
13459 I don't think anyone really wanted that behaviour.
13463 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13464 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13465 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13466 and add new functions:
13475 BN_mod_lshift1_quick
13477 BN_mod_lshift_quick
13479 These functions always generate non-negative results.
13481 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13482 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13484 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13485 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13486 be reduced modulo `m`.
13488 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13491 The following entry accidentally appeared in the CHANGES file
13492 distributed with OpenSSL 0.9.7. The modifications described in
13493 it do *not* apply to OpenSSL 0.9.7.
13495 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13496 was actually never needed) and in BN_mul(). The removal in BN_mul()
13497 required a small change in bn_mul_part_recursive() and the addition
13498 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13499 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13500 bn_sub_words() and bn_add_words() except they take arrays with
13506 * In 'openssl passwd', verify passwords read from the terminal
13507 unless the '-salt' option is used (which usually means that
13508 verification would just waste user's time since the resulting
13509 hash is going to be compared with some given password hash)
13510 or the new '-noverify' option is used.
13512 This is an incompatible change, but it does not affect
13513 non-interactive use of 'openssl passwd' (passwords on the command
13514 line, '-stdin' option, '-in ...' option) and thus should not
13515 cause any problems.
13519 * Remove all references to RSAref, since there's no more need for it.
13523 * Make DSO load along a path given through an environment variable
13524 (SHLIB_PATH) with shl_load().
13528 * Constify the ENGINE code as a result of BIGNUM constification.
13529 Also constify the RSA code and most things related to it. In a
13530 few places, most notable in the depth of the ASN.1 code, ugly
13531 casts back to non-const were required (to be solved at a later
13536 * Make it so the openssl application has all engines loaded by default.
13540 * Constify the BIGNUM routines a little more.
13544 * Add the following functions:
13546 ENGINE_load_cswift()
13548 ENGINE_load_atalla()
13549 ENGINE_load_nuron()
13550 ENGINE_load_builtin_engines()
13552 That way, an application can itself choose if external engines that
13553 are built-in in OpenSSL shall ever be used or not. The benefit is
13554 that applications won't have to be linked with libdl or other dso
13555 libraries unless it's really needed.
13557 Changed 'openssl engine' to load all engines on demand.
13558 Changed the engine header files to avoid the duplication of some
13559 declarations (they differed!).
13563 * 'openssl engine' can now list capabilities.
13567 * Better error reporting in 'openssl engine'.
13571 * Never call load_dh_param(NULL) in s_server.
13575 * Add engine application. It can currently list engines by name and
13576 identity, and test if they are actually available.
13580 * Improve RPM specification file by forcing symbolic linking and making
13581 sure the installed documentation is also owned by root.root.
13583 *Damien Miller <djm@mindrot.org>*
13585 * Give the OpenSSL applications more possibilities to make use of
13586 keys (public as well as private) handled by engines.
13590 * Add OCSP code that comes from CertCo.
13594 * Add VMS support for the Rijndael code.
13598 * Added untested support for Nuron crypto accelerator.
13602 * Add support for external cryptographic devices. This code was
13603 previously distributed separately as the "engine" branch.
13605 *Geoff Thorpe, Richard Levitte*
13607 * Rework the filename-translation in the DSO code. It is now possible to
13608 have far greater control over how a "name" is turned into a filename
13609 depending on the operating environment and any oddities about the
13610 different shared library filenames on each system.
13614 * Support threads on FreeBSD-elf in Configure.
13618 * Fix for SHA1 assembly problem with MASM: it produces
13619 warnings about corrupt line number information when assembling
13620 with debugging information. This is caused by the overlapping
13623 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13626 NCONF_get_number() has no error checking at all. As a replacement,
13627 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13628 promoted strongly. The old NCONF_get_number is kept around for
13629 binary backward compatibility.
13630 Make it possible for methods to load from something other than a BIO,
13631 by providing a function pointer that is given a name instead of a BIO.
13632 For example, this could be used to load configuration data from an
13637 * Fix for non blocking accept BIOs. Added new I/O special reason
13638 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13639 with non blocking I/O was not possible because no retry code was
13640 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13645 * Added the beginnings of Rijndael support.
13649 * Fix for bug in DirectoryString mask setting. Add support for
13650 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13651 to allow certificate printing to more controllable, additional
13652 'certopt' option to 'x509' to allow new printing options to be
13657 * Clean old EAY MD5 hack from e_os.h.
13661 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13663 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13664 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13666 *Joe Orton, Steve Henson*
13668 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13670 * Fix additional bug revealed by the NISCC test suite:
13672 Stop bug triggering large recursion when presented with
13673 certain ASN.1 tags ([CVE-2003-0851])
13677 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13679 * Fix various bugs revealed by running the NISCC test suite:
13681 Stop out of bounds reads in the ASN1 code when presented with
13682 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13684 If verify callback ignores invalid public key errors don't try to check
13685 certificate signature with the NULL public key.
13689 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13690 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13695 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13696 extra data after the compression methods not only for TLS 1.0
13697 but also for SSL 3.0 (as required by the specification).
13699 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13701 * Change X509_certificate_type() to mark the key as exported/exportable
13702 when it's 512 *bits* long, not 512 bytes.
13706 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13708 * Countermeasure against the Klima-Pokorny-Rosa extension of
13709 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13710 a protocol version number mismatch like a decryption error
13711 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13715 * Turn on RSA blinding by default in the default implementation
13716 to avoid a timing attack. Applications that don't want it can call
13717 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13718 They would be ill-advised to do so in most cases.
13720 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13722 * Change RSA blinding code so that it works when the PRNG is not
13723 seeded (in this case, the secret RSA exponent is abused as
13724 an unpredictable seed -- if it is not unpredictable, there
13725 is no point in blinding anyway). Make RSA blinding thread-safe
13726 by remembering the creator's thread ID in rsa->blinding and
13727 having all other threads use local one-time blinding factors
13728 (this requires more computation than sharing rsa->blinding, but
13729 avoids excessive locking; and if an RSA object is not shared
13730 between threads, blinding will still be very fast).
13734 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13736 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13737 via timing by performing a MAC computation even if incorrect
13738 block cipher padding has been found. This is a countermeasure
13739 against active attacks where the attacker has to distinguish
13740 between bad padding and a MAC verification error. ([CVE-2003-0078])
13742 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13743 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13744 Martin Vuagnoux (EPFL, Ilion)*
13746 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13748 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13749 memory from its contents. This is done with a counter that will
13750 place alternating values in each byte. This can be used to solve
13751 two issues: 1) the removal of calls to memset() by highly optimizing
13752 compilers, and 2) cleansing with other values than 0, since those can
13753 be read through on certain media, for example a swap space on disk.
13757 * Bugfix: client side session caching did not work with external caching,
13758 because the session->cipher setting was not restored when reloading
13759 from the external cache. This problem was masked, when
13760 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13761 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13765 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13766 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13768 *Zeev Lieber <zeev-l@yahoo.com>*
13770 * Undo an undocumented change introduced in 0.9.6e which caused
13771 repeated calls to OpenSSL_add_all_ciphers() and
13772 OpenSSL_add_all_digests() to be ignored, even after calling
13777 * Change the default configuration reader to deal with last line not
13778 being properly terminated.
13782 * Change X509_NAME_cmp() so it applies the special rules on handling
13783 DN values that are of type PrintableString, as well as RDNs of type
13784 emailAddress where the value has the type ia5String.
13786 *stefank@valicert.com via Richard Levitte*
13788 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13789 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13790 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13791 the bitwise-OR of the two for use by the majority of applications
13792 wanting this behaviour, and update the docs. The documented
13793 behaviour and actual behaviour were inconsistent and had been
13794 changing anyway, so this is more a bug-fix than a behavioural
13797 *Geoff Thorpe, diagnosed by Nadav Har'El*
13799 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13800 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13804 * Fix initialization code race conditions in
13805 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13806 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13807 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13808 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13809 ssl2_get_cipher_by_char(),
13810 ssl3_get_cipher_by_char().
13812 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13814 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13815 the cached sessions are flushed, as the remove_cb() might use ex_data
13816 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13817 (see [openssl.org #212]).
13819 *Geoff Thorpe, Lutz Jaenicke*
13821 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13822 length, instead of the encoding length to d2i_ASN1_OBJECT.
13826 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13828 * [In 0.9.6g-engine release:]
13829 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13831 *Lynn Gazis <lgazis@rainbow.com>*
13833 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13835 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13836 and get fix the header length calculation.
13837 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13838 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13840 * Use proper error handling instead of 'assertions' in buffer
13841 overflow checks added in 0.9.6e. This prevents DoS (the
13842 assertions could call abort()).
13844 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13846 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13848 * Add various sanity checks to asn1_get_length() to reject
13849 the ASN1 length bytes if they exceed sizeof(long), will appear
13850 negative or the content length exceeds the length of the
13853 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13855 * Fix cipher selection routines: ciphers without encryption had no flags
13856 for the cipher strength set and where therefore not handled correctly
13857 by the selection routines (PR #130).
13861 * Fix EVP_dsa_sha macro.
13866 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13867 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13868 that was added in OpenSSL 0.9.6d.
13870 As the countermeasure turned out to be incompatible with some
13871 broken SSL implementations, the new option is part of SSL_OP_ALL.
13872 SSL_OP_ALL is usually employed when compatibility with weird SSL
13873 implementations is desired (e.g. '-bugs' option to 's_client' and
13874 's_server'), so the new option is automatically set in many
13879 * Changes in security patch:
13881 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13882 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13883 Air Force Materiel Command, USAF, under agreement number
13886 * Add various sanity checks to asn1_get_length() to reject
13887 the ASN1 length bytes if they exceed sizeof(long), will appear
13888 negative or the content length exceeds the length of the
13889 supplied buffer. ([CVE-2002-0659])
13891 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13893 * Assertions for various potential buffer overflows, not known to
13894 happen in practice.
13896 *Ben Laurie (CHATS)*
13898 * Various temporary buffers to hold ASCII versions of integers were
13899 too small for 64 bit platforms. ([CVE-2002-0655])
13900 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13902 * Remote buffer overflow in SSL3 protocol - an attacker could
13903 supply an oversized session ID to a client. ([CVE-2002-0656])
13905 *Ben Laurie (CHATS)*
13907 * Remote buffer overflow in SSL2 protocol - an attacker could
13908 supply an oversized client master key. ([CVE-2002-0656])
13910 *Ben Laurie (CHATS)*
13912 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13914 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13915 encoded as NULL) with id-dsa-with-sha1.
13917 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13919 * Check various `X509_...()` return values in `apps/req.c`.
13921 *Nils Larsch <nla@trustcenter.de>*
13923 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13924 an end-of-file condition would erroneously be flagged, when the CRLF
13925 was just at the end of a processed block. The bug was discovered when
13926 processing data through a buffering memory BIO handing the data to a
13927 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13928 <ptsekov@syntrex.com> and Nedelcho Stanev.
13932 * Implement a countermeasure against a vulnerability recently found
13933 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13934 before application data chunks to avoid the use of known IVs
13935 with data potentially chosen by the attacker.
13939 * Fix length checks in ssl3_get_client_hello().
13943 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13944 to prevent ssl3_read_internal() from incorrectly assuming that
13945 ssl3_read_bytes() found application data while handshake
13946 processing was enabled when in fact s->s3->in_read_app_data was
13947 merely automatically cleared during the initial handshake.
13949 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13951 * Fix object definitions for Private and Enterprise: they were not
13952 recognized in their shortname (=lowercase) representation. Extend
13953 obj_dat.pl to issue an error when using undefined keywords instead
13954 of silently ignoring the problem (Svenning Sorensen
13955 <sss@sss.dnsalias.net>).
13959 * Fix DH_generate_parameters() so that it works for 'non-standard'
13960 generators, i.e. generators other than 2 and 5. (Previously, the
13961 code did not properly initialise the 'add' and 'rem' values to
13962 BN_generate_prime().)
13964 In the new general case, we do not insist that 'generator' is
13965 actually a primitive root: This requirement is rather pointless;
13966 a generator of the order-q subgroup is just as good, if not
13971 * Map new X509 verification errors to alerts. Discovered and submitted by
13972 Tom Wu <tom@arcot.com>.
13976 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13977 returning non-zero before the data has been completely received
13978 when using non-blocking I/O.
13980 *Bodo Moeller; problem pointed out by John Hughes*
13982 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13984 *Ben Laurie, Lutz Jaenicke*
13986 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13987 Yoram Zahavi <YoramZ@gilian.com>).
13991 * Add information about CygWin 1.3 and on, and preserve proper
13992 configuration for the versions before that.
13994 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13996 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13997 check whether we deal with a copy of a session and do not delete from
13998 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13999 <izhar@checkpoint.com>.
14003 * Do not store session data into the internal session cache, if it
14004 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14005 flag is set). Proposed by Aslam <aslam@funk.com>.
14009 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14014 * [In 0.9.6d-engine release:]
14015 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14017 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14019 * Add the configuration target linux-s390x.
14021 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14023 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14024 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14025 variable as an indication that a ClientHello message has been
14026 received. As the flag value will be lost between multiple
14027 invocations of ssl3_accept when using non-blocking I/O, the
14028 function may not be aware that a handshake has actually taken
14029 place, thus preventing a new session from being added to the
14032 To avoid this problem, we now set s->new_session to 2 instead of
14033 using a local variable.
14035 *Lutz Jaenicke, Bodo Moeller*
14037 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14038 if the SSL_R_LENGTH_MISMATCH error is detected.
14040 *Geoff Thorpe, Bodo Moeller*
14042 * New 'shared_ldflag' column in Configure platform table.
14046 * Fix EVP_CIPHER_mode macro.
14048 *"Dan S. Camper" <dan@bti.net>*
14050 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14051 type, we must throw them away by setting rr->length to 0.
14053 *D P Chang <dpc@qualys.com>*
14055 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14057 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14058 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14059 worked incorrectly for those cases where range = `10..._2` and
14060 `3*range` is two bits longer than range.)
14064 * Only add signing time to PKCS7 structures if it is not already
14069 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14070 OBJ_ld_ce should be OBJ_id_ce.
14071 Also some ip-pda OIDs in crypto/objects/objects.txt were
14072 incorrect (cf. RFC 3039).
14074 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14076 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14077 returns early because it has nothing to do.
14079 *Andy Schneider <andy.schneider@bjss.co.uk>*
14081 * [In 0.9.6c-engine release:]
14082 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14084 *Andy Schneider <andy.schneider@bjss.co.uk>*
14086 * [In 0.9.6c-engine release:]
14087 Add support for Cryptographic Appliance's keyserver technology.
14088 (Use engine 'keyclient')
14090 *Cryptographic Appliances and Geoff Thorpe*
14092 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14093 is called via tools/c89.sh because arguments have to be
14094 rearranged (all '-L' options must appear before the first object
14097 *Richard Shapiro <rshapiro@abinitio.com>*
14099 * [In 0.9.6c-engine release:]
14100 Add support for Broadcom crypto accelerator cards, backported
14103 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14105 * [In 0.9.6c-engine release:]
14106 Add support for SureWare crypto accelerator cards from
14107 Baltimore Technologies. (Use engine 'sureware')
14109 *Baltimore Technologies and Mark Cox*
14111 * [In 0.9.6c-engine release:]
14112 Add support for crypto accelerator cards from Accelerated
14113 Encryption Processing, www.aep.ie. (Use engine 'aep')
14115 *AEP Inc. and Mark Cox*
14117 * Add a configuration entry for gcc on UnixWare.
14119 *Gary Benson <gbenson@redhat.com>*
14121 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14122 messages are stored in a single piece (fixed-length part and
14123 variable-length part combined) and fix various bugs found on the way.
14127 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14128 instead. BIO_gethostbyname() does not know what timeouts are
14129 appropriate, so entries would stay in cache even when they have
14131 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14133 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14134 faced with a pathologically small ClientHello fragment that does
14135 not contain client_version: Instead of aborting with an error,
14136 simply choose the highest available protocol version (i.e.,
14137 TLS 1.0 unless it is disabled). In practice, ClientHello
14138 messages are never sent like this, but this change gives us
14139 strictly correct behaviour at least for TLS.
14143 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14144 never resets s->method to s->ctx->method when called from within
14145 one of the SSL handshake functions.
14147 *Bodo Moeller; problem pointed out by Niko Baric*
14149 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14150 (sent using the client's version number) if client_version is
14151 smaller than the protocol version in use. Also change
14152 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14153 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14154 the client will at least see that alert.
14158 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14163 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14164 client receives HelloRequest while in a handshake.
14166 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14168 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14169 should end in 'break', not 'goto end' which circumvents various
14170 cleanups done in state SSL_ST_OK. But session related stuff
14171 must be disabled for SSL_ST_OK in the case that we just sent a
14174 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14175 before just sending a HelloRequest.
14177 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14179 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14180 reveal whether illegal block cipher padding was found or a MAC
14181 verification error occurred. (Neither SSLerr() codes nor alerts
14182 are directly visible to potential attackers, but the information
14183 may leak via logfiles.)
14185 Similar changes are not required for the SSL 2.0 implementation
14186 because the number of padding bytes is sent in clear for SSL 2.0,
14187 and the extra bytes are just ignored. However ssl/s2_pkt.c
14188 failed to verify that the purported number of padding bytes is in
14193 * Add OpenUNIX-8 support including shared libraries
14194 (Boyd Lynn Gerber <gerberb@zenez.com>).
14198 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14199 'wristwatch attack' using huge encoding parameters (cf.
14200 James H. Manger's CRYPTO 2001 paper). Note that the
14201 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14202 encoding parameters and hence was not vulnerable.
14206 * BN_sqr() bug fix.
14208 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14210 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14211 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14212 followed by modular reduction.
14214 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14216 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14217 equivalent based on BN_pseudo_rand() instead of BN_rand().
14221 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14222 This function was broken, as the check for a new client hello message
14223 to handle SGC did not allow these large messages.
14224 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14228 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14232 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14233 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14237 * Rework the configuration and shared library support for Tru64 Unix.
14238 The configuration part makes use of modern compiler features and
14239 still retains old compiler behavior for those that run older versions
14240 of the OS. The shared library support part includes a variant that
14241 uses the RPATH feature, and is available through the special
14242 configuration target "alpha-cc-rpath", which will never be selected
14245 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14247 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14248 with the same message size as in ssl3_get_certificate_request().
14249 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14250 messages might inadvertently be reject as too long.
14252 *Petr Lampa <lampa@fee.vutbr.cz>*
14254 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14258 * Modified SSL library such that the verify_callback that has been set
14259 specifically for an SSL object with SSL_set_verify() is actually being
14260 used. Before the change, a verify_callback set with this function was
14261 ignored and the verify_callback() set in the SSL_CTX at the time of
14262 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14263 to allow the necessary settings.
14267 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14268 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14269 done automatically (in contradiction to the requirements of the C
14270 standard). This made problems when used from OpenSSH.
14274 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14275 dh->length and always used
14277 BN_rand_range(priv_key, dh->p).
14279 BN_rand_range() is not necessary for Diffie-Hellman, and this
14280 specific range makes Diffie-Hellman unnecessarily inefficient if
14281 dh->length (recommended exponent length) is much smaller than the
14282 length of dh->p. We could use BN_rand_range() if the order of
14283 the subgroup was stored in the DH structure, but we only have
14288 BN_rand(priv_key, l, ...)
14290 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14297 RSA_eay_public_encrypt
14298 RSA_eay_private_decrypt
14299 RSA_eay_private_encrypt (signing)
14300 RSA_eay_public_decrypt (signature verification)
14302 (default implementations for RSA_public_encrypt,
14303 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14304 always reject numbers >= n.
14308 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14309 to synchronize access to 'locking_thread'. This is necessary on
14310 systems where access to 'locking_thread' (an 'unsigned long'
14311 variable) is not atomic.
14315 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14316 *before* setting the 'crypto_lock_rand' flag. The previous code had
14317 a race condition if 0 is a valid thread ID.
14319 *Travis Vitek <vitek@roguewave.com>*
14321 * Add support for shared libraries under Irix.
14323 *Albert Chin-A-Young <china@thewrittenword.com>*
14325 * Add configuration option to build on Linux on both big-endian and
14326 little-endian MIPS.
14328 *Ralf Baechle <ralf@uni-koblenz.de>*
14330 * Add the possibility to create shared libraries on HP-UX.
14334 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14336 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14337 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14338 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14339 PRNG state recovery was possible based on the output of
14340 one PRNG request appropriately sized to gain knowledge on
14341 'md' followed by enough consecutive 1-byte PRNG requests
14342 to traverse all of 'state'.
14344 1. When updating 'md_local' (the current thread's copy of 'md')
14345 during PRNG output generation, hash all of the previous
14346 'md_local' value, not just the half used for PRNG output.
14348 2. Make the number of bytes from 'state' included into the hash
14349 independent from the number of PRNG bytes requested.
14351 The first measure alone would be sufficient to avoid
14352 Markku-Juhani's attack. (Actually it had never occurred
14353 to me that the half of 'md_local' used for chaining was the
14354 half from which PRNG output bytes were taken -- I had always
14355 assumed that the secret half would be used.) The second
14356 measure makes sure that additional data from 'state' is never
14357 mixed into 'md_local' in small portions; this heuristically
14358 further strengthens the PRNG.
14362 * Fix crypto/bn/asm/mips3.s.
14366 * When only the key is given to "enc", the IV is undefined. Print out
14367 an error message in this case.
14371 * Handle special case when X509_NAME is empty in X509 printing routines.
14375 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14376 positive and less than q.
14380 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14381 used: it isn't thread safe and the add_lock_callback should handle
14384 *Paul Rose <Paul.Rose@bridge.com>*
14386 * Verify that incoming data obeys the block size in
14387 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14393 *Ulf Möller, Bodo Möller*
14395 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14396 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14397 when fixing the server behaviour for backwards-compatible 'client
14398 hello' messages. (Note that the attack is impractical against
14399 SSL 3.0 and TLS 1.0 anyway because length and version checking
14400 means that the probability of guessing a valid ciphertext is
14401 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14404 Before 0.9.5, the countermeasure (hide the error by generating a
14405 random 'decryption result') did not work properly because
14406 ERR_clear_error() was missing, meaning that SSL_get_error() would
14407 detect the supposedly ignored error.
14409 Both problems are now fixed.
14413 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14414 (previously it was 1024).
14418 * Fix for compatibility mode trust settings: ignore trust settings
14419 unless some valid trust or reject settings are present.
14423 * Fix for blowfish EVP: its a variable length cipher.
14427 * Fix various bugs related to DSA S/MIME verification. Handle missing
14428 parameters in DSA public key structures and return an error in the
14429 DSA routines if parameters are absent.
14433 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14434 in the current directory if neither $RANDFILE nor $HOME was set.
14435 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14436 caused some confusion to Windows users who haven't defined $HOME.
14437 Thus RAND_file_name() is changed again: e_os.h can define a
14438 DEFAULT_HOME, which will be used if $HOME is not set.
14439 For Windows, we use "C:"; on other platforms, we still require
14440 environment variables.
14442 * Move 'if (!initialized) RAND_poll()' into regions protected by
14443 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14444 having multiple threads call RAND_poll() concurrently.
14448 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14449 combination of a flag and a thread ID variable.
14450 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14451 flag), *other* threads can enter ssleay_add_bytes without obeying
14452 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14453 that they do not hold after the first thread unsets add_do_not_lock).
14457 * Change bctest again: '-x' expressions are not available in all
14458 versions of 'test'.
14462 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14464 * Fix a couple of memory leaks in PKCS7_dataDecode()
14466 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14468 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14469 the default extension for executables, if any. Also, make the perl
14470 scripts that use symlink() to test if it really exists and use "cp"
14471 if it doesn't. All this made OpenSSL compilable and installable in
14476 * Fix for asn1_GetSequence() for indefinite length constructed data.
14477 If SEQUENCE is length is indefinite just set c->slen to the total
14478 amount of data available.
14480 *Steve Henson, reported by shige@FreeBSD.org*
14482 *This change does not apply to 0.9.7.*
14484 * Change bctest to avoid here-documents inside command substitution
14485 (workaround for FreeBSD /bin/sh bug).
14486 For compatibility with Ultrix, avoid shell functions (introduced
14487 in the bctest version that searches along $PATH).
14491 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14492 with des_encrypt() defined on some operating systems, like Solaris
14497 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14498 On the Importance of Eliminating Errors in Cryptographic
14499 Computations, J. Cryptology 14 (2001) 2, 101-119,
14500 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14504 * MIPS assembler BIGNUM division bug fix.
14508 * Disabled incorrect Alpha assembler code.
14512 * Fix PKCS#7 decode routines so they correctly update the length
14513 after reading an EOC for the EXPLICIT tag.
14517 *This change does not apply to 0.9.7.*
14519 * Fix bug in PKCS#12 key generation routines. This was triggered
14520 if a 3DES key was generated with a 0 initial byte. Include
14521 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14522 (but broken) behaviour.
14526 * Enhance bctest to search for a working bc along $PATH and print
14529 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14531 * Fix memory leaks in err.c: free err_data string if necessary;
14532 don't write to the wrong index in ERR_set_error_data.
14536 * Implement ssl23_peek (analogous to ssl23_read), which previously
14541 * Replace rdtsc with `_emit` statements for VC++ version 5.
14543 *Jeremy Cooper <jeremy@baymoo.org>*
14545 * Make it possible to reuse SSLv2 sessions.
14549 * In copy_email() check for >= 0 as a return value for
14550 X509_NAME_get_index_by_NID() since 0 is a valid index.
14552 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14554 * Avoid coredump with unsupported or invalid public keys by checking if
14555 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14556 PKCS7_verify() fails with non detached data.
14560 * Don't use getenv in library functions when run as setuid/setgid.
14561 New function OPENSSL_issetugid().
14565 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14566 due to incorrect handling of multi-threading:
14568 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14570 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14572 3. Count how many times MemCheck_off() has been called so that
14573 nested use can be treated correctly. This also avoids
14574 inband-signalling in the previous code (which relied on the
14575 assumption that thread ID 0 is impossible).
14579 * Add "-rand" option also to s_client and s_server.
14583 * Fix CPU detection on Irix 6.x.
14584 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14585 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14587 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14592 *This change does not apply to 0.9.7.*
14594 * Use the cached encoding of an X509_NAME structure rather than
14595 copying it. This is apparently the reason for the libsafe "errors"
14596 but the code is actually correct.
14600 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14601 Bleichenbacher's DSA attack.
14602 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14603 to be set and top=0 forces the highest bit to be set; top=-1 is new
14604 and leaves the highest bit random.
14606 *Ulf Moeller, Bodo Moeller*
14608 * In the `NCONF_...`-based implementations for `CONF_...` queries
14609 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14610 a temporary CONF structure with the data component set to NULL
14611 (which gives segmentation faults in lh_retrieve).
14612 Instead, use NULL for the CONF pointer in CONF_get_string and
14613 CONF_get_number (which may use environment variables) and directly
14614 return NULL from CONF_get_section.
14618 * Fix potential buffer overrun for EBCDIC.
14622 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14623 keyUsage if basicConstraints absent for a CA.
14627 * Make SMIME_write_PKCS7() write mail header values with a format that
14628 is more generally accepted (no spaces before the semicolon), since
14629 some programs can't parse those values properly otherwise. Also make
14630 sure BIO's that break lines after each write do not create invalid
14635 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14636 macros previously used would not encode an empty SEQUENCE OF
14637 and break the signature.
14641 *This change does not apply to 0.9.7.*
14643 * Zero the premaster secret after deriving the master secret in
14648 * Add some EVP_add_digest_alias registrations (as found in
14649 OpenSSL_add_all_digests()) to SSL_library_init()
14650 aka OpenSSL_add_ssl_algorithms(). This provides improved
14651 compatibility with peers using X.509 certificates
14652 with unconventional AlgorithmIdentifier OIDs.
14656 * Fix for Irix with NO_ASM.
14658 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14660 * ./config script fixes.
14662 *Ulf Moeller, Richard Levitte*
14664 * Fix 'openssl passwd -1'.
14668 * Change PKCS12_key_gen_asc() so it can cope with non null
14669 terminated strings whose length is passed in the passlen
14670 parameter, for example from PEM callbacks. This was done
14671 by adding an extra length parameter to asc2uni().
14673 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14675 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14676 call failed, free the DSA structure.
14680 * Fix to uni2asc() to cope with zero length Unicode strings.
14681 These are present in some PKCS#12 files.
14685 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14686 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14687 when writing a 32767 byte record.
14689 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14691 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14692 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14694 (RSA objects have a reference count access to which is protected
14695 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14696 so they are meant to be shared between threads.)
14697 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14698 "Reddie, Steven" <Steven.Reddie@ca.com>*
14700 * Fix a deadlock in CRYPTO_mem_leaks().
14704 * Use better test patterns in bntest.
14708 * rand_win.c fix for Borland C.
14712 * BN_rshift bugfix for n == 0.
14716 * Add a 'bctest' script that checks for some known 'bc' bugs
14717 so that 'make test' does not abort just because 'bc' is broken.
14721 * Store verify_result within SSL_SESSION also for client side to
14722 avoid potential security hole. (Re-used sessions on the client side
14723 always resulted in verify_result==X509_V_OK, not using the original
14724 result of the server certificate verification.)
14728 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14729 SSL3_RT_APPLICATION_DATA, return 0.
14730 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14735 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14736 releases, have been re-implemented by renaming the previous
14737 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14738 and ssl3_read_internal, respectively, and adding 'peek' parameters
14739 to them. The new ssl[23]_{read,peek} functions are calls to
14740 ssl[23]_read_internal with the 'peek' flag set appropriately.
14741 A 'peek' parameter has also been added to ssl3_read_bytes, which
14742 does the actual work for ssl3_read_internal.
14746 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14747 the method-specific "init()" handler. Also clean up ex_data after
14748 calling the method-specific "finish()" handler. Previously, this was
14749 happening the other way round.
14753 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14754 The previous value, 12, was not always sufficient for BN_mod_exp().
14758 * Make sure that shared libraries get the internal name engine with
14759 the full version number and not just 0. This should mark the
14760 shared libraries as not backward compatible. Of course, this should
14761 be changed again when we can guarantee backward binary compatibility.
14765 * Fix typo in get_cert_by_subject() in by_dir.c
14767 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14769 * Rework the system to generate shared libraries:
14771 - Make note of the expected extension for the shared libraries and
14772 if there is a need for symbolic links from for example libcrypto.so.0
14773 to libcrypto.so.0.9.7. There is extended info in Configure for
14776 - Make as few rebuilds of the shared libraries as possible.
14778 - Still avoid linking the OpenSSL programs with the shared libraries.
14780 - When installing, install the shared libraries separately from the
14785 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14787 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14788 and not in SSL_clear because the latter is also used by the
14789 accept/connect functions; previously, the settings made by
14790 SSL_set_read_ahead would be lost during the handshake.
14792 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14794 * Correct util/mkdef.pl to be selective about disabled algorithms.
14795 Previously, it would create entries for disabled algorithms no
14800 * Added several new manual pages for SSL_* function.
14804 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14806 * In ssl23_get_client_hello, generate an error message when faced
14807 with an initial SSL 3.0/TLS record that is too small to contain the
14808 first two bytes of the ClientHello message, i.e. client_version.
14809 (Note that this is a pathologic case that probably has never happened
14810 in real life.) The previous approach was to use the version number
14811 from the record header as a substitute; but our protocol choice
14812 should not depend on that one because it is not authenticated
14813 by the Finished messages.
14817 * More robust randomness gathering functions for Windows.
14819 *Jeffrey Altman <jaltman@columbia.edu>*
14821 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14822 not set then we don't setup the error code for issuer check errors
14823 to avoid possibly overwriting other errors which the callback does
14824 handle. If an application does set the flag then we assume it knows
14825 what it is doing and can handle the new informational codes
14830 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14831 a general "ANY" type, as such it should be able to decode anything
14832 including tagged types. However it didn't check the class so it would
14833 wrongly interpret tagged types in the same way as their universal
14834 counterpart and unknown types were just rejected. Changed so that the
14835 tagged and unknown types are handled in the same way as a SEQUENCE:
14836 that is the encoding is stored intact. There is also a new type
14837 "V_ASN1_OTHER" which is used when the class is not universal, in this
14838 case we have no idea what the actual type is so we just lump them all
14843 * On VMS, stdout may very well lead to a file that is written to
14844 in a record-oriented fashion. That means that every write() will
14845 write a separate record, which will be read separately by the
14846 programs trying to read from it. This can be very confusing.
14848 The solution is to put a BIO filter in the way that will buffer
14849 text until a linefeed is reached, and then write everything a
14850 line at a time, so every record written will be an actual line,
14851 not chunks of lines and not (usually doesn't happen, but I've
14852 seen it once) several lines in one record. BIO_f_linebuffer() is
14855 Currently, it's a VMS-only method, because that's where it has
14856 been tested well enough.
14860 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14861 it can return incorrect results.
14862 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14863 but it was in 0.9.6-beta[12].)
14867 * Disable the check for content being present when verifying detached
14868 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14869 include zero length content when signing messages.
14873 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14874 BIO_ctrl (for BIO pairs).
14878 * Add DSO method for VMS.
14882 * Bug fix: Montgomery multiplication could produce results with the
14887 * Add RPM specification openssl.spec and modify it to build three
14888 packages. The default package contains applications, application
14889 documentation and run-time libraries. The devel package contains
14890 include files, static libraries and function documentation. The
14891 doc package contains the contents of the doc directory. The original
14892 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14896 * Add a large number of documentation files for many SSL routines.
14898 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14900 * Add a configuration entry for Sony News 4.
14902 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14904 * Don't set the two most significant bits to one when generating a
14905 random number < q in the DSA library.
14909 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14910 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14911 the underlying transport is blocking) if a handshake took place.
14912 (The default behaviour is needed by applications such as s_client
14913 and s_server that use select() to determine when to use SSL_read;
14914 but for applications that know in advance when to expect data, it
14915 just makes things more complicated.)
14919 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14924 * Add a few more EBCDIC conditionals that make `req` and `x509`
14925 work better on such systems.
14927 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14929 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14930 Update PKCS12_parse() so it copies the friendlyName and the
14931 keyid to the certificates aux info.
14935 * Fix bug in PKCS7_verify() which caused an infinite loop
14936 if there was more than one signature.
14938 *Sven Uszpelkat <su@celocom.de>*
14940 * Major change in util/mkdef.pl to include extra information
14941 about each symbol, as well as presenting variables as well
14942 as functions. This change means that there's n more need
14943 to rebuild the .num files when some algorithms are excluded.
14947 * Allow the verify time to be set by an application,
14948 rather than always using the current time.
14952 * Phase 2 verify code reorganisation. The certificate
14953 verify code now looks up an issuer certificate by a
14954 number of criteria: subject name, authority key id
14955 and key usage. It also verifies self signed certificates
14956 by the same criteria. The main comparison function is
14957 X509_check_issued() which performs these checks.
14959 Lot of changes were necessary in order to support this
14960 without completely rewriting the lookup code.
14962 Authority and subject key identifier are now cached.
14964 The LHASH 'certs' is X509_STORE has now been replaced
14965 by a STACK_OF(X509_OBJECT). This is mainly because an
14966 LHASH can't store or retrieve multiple objects with
14967 the same hash value.
14969 As a result various functions (which were all internal
14970 use only) have changed to handle the new X509_STORE
14971 structure. This will break anything that messed round
14972 with X509_STORE internally.
14974 The functions X509_STORE_add_cert() now checks for an
14975 exact match, rather than just subject name.
14977 The X509_STORE API doesn't directly support the retrieval
14978 of multiple certificates matching a given criteria, however
14979 this can be worked round by performing a lookup first
14980 (which will fill the cache with candidate certificates)
14981 and then examining the cache for matches. This is probably
14982 the best we can do without throwing out X509_LOOKUP
14983 entirely (maybe later...).
14985 The X509_VERIFY_CTX structure has been enhanced considerably.
14987 All certificate lookup operations now go via a get_issuer()
14988 callback. Although this currently uses an X509_STORE it
14989 can be replaced by custom lookups. This is a simple way
14990 to bypass the X509_STORE hackery necessary to make this
14991 work and makes it possible to use more efficient techniques
14992 in future. A very simple version which uses a simple
14993 STACK for its trusted certificate store is also provided
14994 using X509_STORE_CTX_trusted_stack().
14996 The verify_cb() and verify() callbacks now have equivalents
14997 in the X509_STORE_CTX structure.
14999 X509_STORE_CTX also has a 'flags' field which can be used
15000 to customise the verify behaviour.
15004 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15005 excludes S/MIME capabilities.
15009 * When a certificate request is read in keep a copy of the
15010 original encoding of the signed data and use it when outputting
15011 again. Signatures then use the original encoding rather than
15012 a decoded, encoded version which may cause problems if the
15013 request is improperly encoded.
15017 * For consistency with other BIO_puts implementations, call
15018 buffer_write(b, ...) directly in buffer_puts instead of calling
15021 In BIO_puts, increment b->num_write as in BIO_write.
15023 *Peter.Sylvester@EdelWeb.fr*
15025 * Fix BN_mul_word for the case where the word is 0. (We have to use
15026 BN_zero, we may not return a BIGNUM with an array consisting of
15027 words set to zero.)
15031 * Avoid calling abort() from within the library when problems are
15032 detected, except if preprocessor symbols have been defined
15033 (such as REF_CHECK, BN_DEBUG etc.).
15037 * New openssl application 'rsautl'. This utility can be
15038 used for low level RSA operations. DER public key
15039 BIO/fp routines also added.
15043 * New Configure entry and patches for compiling on QNX 4.
15045 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15047 * A demo state-machine implementation was sponsored by
15048 Nuron (<http://www.nuron.com/>) and is now available in
15049 demos/state_machine.
15053 * New options added to the 'dgst' utility for signature
15054 generation and verification.
15058 * Unrecognized PKCS#7 content types are now handled via a
15059 catch all ASN1_TYPE structure. This allows unsupported
15060 types to be stored as a "blob" and an application can
15061 encode and decode it manually.
15065 * Fix various signed/unsigned issues to make a_strex.c
15066 compile under VC++.
15068 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15070 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15071 length if passed a buffer. ASN1_INTEGER_to_BN failed
15072 if passed a NULL BN and its argument was negative.
15074 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15076 * Modification to PKCS#7 encoding routines to output definite
15077 length encoding. Since currently the whole structures are in
15078 memory there's not real point in using indefinite length
15079 constructed encoding. However if OpenSSL is compiled with
15080 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15084 * Added BIO_vprintf() and BIO_vsnprintf().
15088 * Added more prefixes to parse for in the strings written
15089 through a logging bio, to cover all the levels that are available
15090 through syslog. The prefixes are now:
15092 PANIC, EMERG, EMR => LOG_EMERG
15093 ALERT, ALR => LOG_ALERT
15094 CRIT, CRI => LOG_CRIT
15095 ERROR, ERR => LOG_ERR
15096 WARNING, WARN, WAR => LOG_WARNING
15097 NOTICE, NOTE, NOT => LOG_NOTICE
15098 INFO, INF => LOG_INFO
15099 DEBUG, DBG => LOG_DEBUG
15101 and as before, if none of those prefixes are present at the
15102 beginning of the string, LOG_ERR is chosen.
15104 On Win32, the `LOG_*` levels are mapped according to this:
15106 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15107 LOG_WARNING => EVENTLOG_WARNING_TYPE
15108 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15112 * Made it possible to reconfigure with just the configuration
15113 argument "reconf" or "reconfigure". The command line arguments
15114 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15115 and are retrieved from there when reconfiguring.
15121 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15123 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15127 * The obj_dat.pl script was messing up the sorting of object
15128 names. The reason was that it compared the quoted version
15129 of strings as a result "OCSP" > "OCSP Signing" because
15130 " > SPACE. Changed script to store unquoted versions of
15131 names and add quotes on output. It was also omitting some
15132 names from the lookup table if they were given a default
15133 value (that is if SN is missing it is given the same
15134 value as LN and vice versa), these are now added on the
15135 grounds that if an object has a name we should be able to
15136 look it up. Finally added warning output when duplicate
15137 short or long names are found.
15141 * Changes needed for Tandem NSK.
15143 *Scott Uroff <scott@xypro.com>*
15145 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15146 RSA_padding_check_SSLv23(), special padding was never detected
15147 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15148 version rollback attacks was not effective.
15150 In s23_clnt.c, don't use special rollback-attack detection padding
15151 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15152 client; similarly, in s23_srvr.c, don't do the rollback check if
15153 SSL 2.0 is the only protocol enabled in the server.
15157 * Make it possible to get hexdumps of unprintable data with 'openssl
15158 asn1parse'. By implication, the functions ASN1_parse_dump() and
15159 BIO_dump_indent() are added.
15163 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15164 these print out strings and name structures based on various
15165 flags including RFC2253 support and proper handling of
15166 multibyte characters. Added options to the 'x509' utility
15167 to allow the various flags to be set.
15171 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15172 Also change the functions X509_cmp_current_time() and
15173 X509_gmtime_adj() work with an ASN1_TIME structure,
15174 this will enable certificates using GeneralizedTime in validity
15175 dates to be checked.
15179 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15180 negative public key encodings) on by default,
15181 NO_NEG_PUBKEY_BUG can be set to disable it.
15185 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15186 content octets. An i2c_ASN1_OBJECT is unnecessary because
15187 the encoding can be trivially obtained from the structure.
15191 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15192 not read locks (`CRYPTO_r_[un]lock`).
15196 * A first attempt at creating official support for shared
15197 libraries through configuration. I've kept it so the
15198 default is static libraries only, and the OpenSSL programs
15199 are always statically linked for now, but there are
15200 preparations for dynamic linking in place.
15201 This has been tested on Linux and Tru64.
15205 * Randomness polling function for Win9x, as described in:
15206 Peter Gutmann, Software Generation of Practically Strong
15211 * Fix so PRNG is seeded in req if using an already existing
15216 * New options to smime application. -inform and -outform
15217 allow alternative formats for the S/MIME message including
15218 PEM and DER. The -content option allows the content to be
15219 specified separately. This should allow things like Netscape
15220 form signing output easier to verify.
15224 * Fix the ASN1 encoding of tags using the 'long form'.
15228 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15229 STRING types. These convert content octets to and from the
15230 underlying type. The actual tag and length octets are
15231 already assumed to have been read in and checked. These
15232 are needed because all other string types have virtually
15233 identical handling apart from the tag. By having versions
15234 of the ASN1 functions that just operate on content octets
15235 IMPLICIT tagging can be handled properly. It also allows
15236 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15237 and ASN1_INTEGER are identical apart from the tag.
15241 * Change the handling of OID objects as follows:
15243 - New object identifiers are inserted in objects.txt, following
15244 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15245 - objects.pl is used to process obj_mac.num and create a new
15247 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15250 This is currently kind of a hack, and the perl code in objects.pl
15251 isn't very elegant, but it works as I intended. The simplest way
15252 to check that it worked correctly is to look in obj_dat.h and
15253 check the array nid_objs and make sure the objects haven't moved
15254 around (this is important!). Additions are OK, as well as
15255 consistent name changes.
15259 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15263 * Addition of the command line parameter '-rand file' to 'openssl req'.
15264 The given file adds to whatever has already been seeded into the
15265 random pool through the RANDFILE configuration file option or
15266 environment variable, or the default random state file.
15270 * mkstack.pl now sorts each macro group into lexical order.
15271 Previously the output order depended on the order the files
15272 appeared in the directory, resulting in needless rewriting
15277 * Patches to make OpenSSL compile under Win32 again. Mostly
15278 work arounds for the VC++ problem that it treats func() as
15279 func(void). Also stripped out the parts of mkdef.pl that
15280 added extra typesafe functions: these no longer exist.
15284 * Reorganisation of the stack code. The macros are now all
15285 collected in safestack.h . Each macro is defined in terms of
15286 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15287 DEBUG_SAFESTACK is now handled in terms of function casts,
15288 this has the advantage of retaining type safety without the
15289 use of additional functions. If DEBUG_SAFESTACK is not defined
15290 then the non typesafe macros are used instead. Also modified the
15291 mkstack.pl script to handle the new form. Needs testing to see
15292 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15293 the default if no major problems. Similar behaviour for ASN1_SET_OF
15294 and PKCS12_STACK_OF.
15298 * When some versions of IIS use the 'NET' form of private key the
15299 key derivation algorithm is different. Normally MD5(password) is
15300 used as a 128 bit RC4 key. In the modified case
15301 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15302 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15303 as the old Netscape_RSA functions except they have an additional
15304 'sgckey' parameter which uses the modified algorithm. Also added
15305 an -sgckey command line option to the rsa utility. Thanks to
15306 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15307 algorithm to openssl-dev.
15311 * The evp_local.h macros were using 'c.##kname' which resulted in
15312 invalid expansion on some systems (SCO 5.0.5 for example).
15313 Corrected to 'c.kname'.
15315 *Phillip Porch <root@theporch.com>*
15317 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15318 a STACK of email addresses from a certificate or request, these look
15319 in the subject name and the subject alternative name extensions and
15320 omit any duplicate addresses.
15324 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15325 This makes DSA verification about 2 % faster.
15329 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15330 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15331 plus overhead for 1024 bit moduli).
15332 This makes exponentiations about 0.5 % faster for 1024 bit
15333 exponents (as measured by "openssl speed rsa2048").
15337 * Rename memory handling macros to avoid conflicts with other
15339 Malloc => OPENSSL_malloc
15340 Malloc_locked => OPENSSL_malloc_locked
15341 Realloc => OPENSSL_realloc
15342 Free => OPENSSL_free
15346 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15347 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15351 * CygWin32 support.
15353 *John Jarvie <jjarvie@newsguy.com>*
15355 * The type-safe stack code has been rejigged. It is now only compiled
15356 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15357 by default all type-specific stack functions are "#define"d back to
15358 standard stack functions. This results in more streamlined output
15359 but retains the type-safety checking possibilities of the original
15364 * The STACK code has been cleaned up, and certain type declarations
15365 that didn't make a lot of sense have been brought in line. This has
15366 also involved a cleanup of sorts in safestack.h to more correctly
15367 map type-safe stack functions onto their plain stack counterparts.
15368 This work has also resulted in a variety of "const"ifications of
15369 lots of the code, especially `_cmp` operations which should normally
15370 be prototyped with "const" parameters anyway.
15374 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15375 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15376 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15377 where all of 'md' is used each time the PRNG is used, but 'state'
15378 is used only indexed by a cyclic counter. As entropy may not be
15379 well distributed from the beginning, 'md' is important as a
15380 chaining variable. However, the output function chains only half
15381 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15382 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15383 in all of 'state' being rewritten, with the new values depending
15384 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15388 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15389 the handshake is continued after ssl_verify_cert_chain();
15390 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15391 can lead to 'unexplainable' connection aborts later.
15393 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15395 * Major EVP API cipher revision.
15396 Add hooks for extra EVP features. This allows various cipher
15397 parameters to be set in the EVP interface. Support added for variable
15398 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15399 setting of RC2 and RC5 parameters.
15401 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15404 Remove lots of duplicated code from the EVP library. For example *every*
15405 cipher init() function handles the 'iv' in the same way according to the
15406 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15407 for CFB and OFB modes they zero ctx->num.
15409 New functionality allows removal of S/MIME code RC2 hack.
15411 Most of the routines have the same form and so can be declared in terms
15414 By shifting this to the top level EVP_CipherInit() it can be removed from
15415 all individual ciphers. If the cipher wants to handle IVs or keys
15416 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15419 Change lots of functions like EVP_EncryptUpdate() to now return a
15420 value: although software versions of the algorithms cannot fail
15421 any installed hardware versions can.
15425 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15426 this option is set, tolerate broken clients that send the negotiated
15427 protocol version number instead of the requested protocol version
15432 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15433 i.e. non-zero for export ciphersuites, zero otherwise.
15434 Previous versions had this flag inverted, inconsistent with
15435 rsa_tmp_cb (..._TMP_RSA_CB).
15437 *Bodo Moeller; problem reported by Amit Chopra*
15439 * Add missing DSA library text string. Work around for some IIS
15440 key files with invalid SEQUENCE encoding.
15444 * Add a document (doc/standards.txt) that list all kinds of standards
15445 and so on that are implemented in OpenSSL.
15449 * Enhance c_rehash script. Old version would mishandle certificates
15450 with the same subject name hash and wouldn't handle CRLs at all.
15451 Added -fingerprint option to crl utility, to support new c_rehash
15456 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15460 * Fix for SSL server purpose checking. Server checking was
15461 rejecting certificates which had extended key usage present
15462 but no ssl client purpose.
15464 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15466 * Make PKCS#12 code work with no password. The PKCS#12 spec
15467 is a little unclear about how a blank password is handled.
15468 Since the password in encoded as a BMPString with terminating
15469 double NULL a zero length password would end up as just the
15470 double NULL. However no password at all is different and is
15471 handled differently in the PKCS#12 key generation code. NS
15472 treats a blank password as zero length. MSIE treats it as no
15473 password on export: but it will try both on import. We now do
15474 the same: PKCS12_parse() tries zero length and no password if
15475 the password is set to "" or NULL (NULL is now a valid password:
15476 it wasn't before) as does the pkcs12 application.
15480 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15481 perror when PEM_read_bio_X509_REQ fails, the error message must
15482 be obtained from the error queue.
15486 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15487 it in ERR_remove_state if appropriate, and change ERR_get_state
15488 accordingly to avoid race conditions (this is necessary because
15489 thread_hash is no longer constant once set).
15493 * Bugfix for linux-elf makefile.one.
15497 * RSA_get_default_method() will now cause a default
15498 RSA_METHOD to be chosen if one doesn't exist already.
15499 Previously this was only set during a call to RSA_new()
15500 or RSA_new_method(NULL) meaning it was possible for
15501 RSA_get_default_method() to return NULL.
15505 * Added native name translation to the existing DSO code
15506 that will convert (if the flag to do so is set) filenames
15507 that are sufficiently small and have no path information
15508 into a canonical native form. Eg. "blah" converted to
15509 "libblah.so" or "blah.dll" etc.
15513 * New function ERR_error_string_n(e, buf, len) which is like
15514 ERR_error_string(e, buf), but writes at most 'len' bytes
15515 including the 0 terminator. For ERR_error_string_n, 'buf'
15518 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15520 * CONF library reworked to become more general. A new CONF
15521 configuration file reader "class" is implemented as well as a
15522 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15523 old `CONF_*` functions are still there, but are reimplemented to
15524 work in terms of the new functions. Also, a set of functions
15525 to handle the internal storage of the configuration data is
15526 provided to make it easier to write new configuration file
15527 reader "classes" (I can definitely see something reading a
15528 configuration file in XML format, for example), called `_CONF_*`,
15529 or "the configuration storage API"...
15531 The new configuration file reading functions are:
15533 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15534 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15536 NCONF_default, NCONF_WIN32
15538 NCONF_dump_fp, NCONF_dump_bio
15540 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15541 NCONF_new creates a new CONF object. This works in the same way
15542 as other interfaces in OpenSSL, like the BIO interface.
15543 `NCONF_dump_*` dump the internal storage of the configuration file,
15544 which is useful for debugging. All other functions take the same
15545 arguments as the old `CONF_*` functions with the exception of the
15546 first that must be a `CONF *` instead of a `LHASH *`.
15548 To make it easier to use the new classes with the old `CONF_*` functions,
15549 the function CONF_set_default_method is provided.
15553 * Add '-tls1' option to 'openssl ciphers', which was already
15554 mentioned in the documentation but had not been implemented.
15555 (This option is not yet really useful because even the additional
15556 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15560 * Initial DSO code added into libcrypto for letting OpenSSL (and
15561 OpenSSL-based applications) load shared libraries and bind to
15562 them in a portable way.
15564 *Geoff Thorpe, with contributions from Richard Levitte*
15566 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15568 * Make sure _lrotl and _lrotr are only used with MSVC.
15570 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15571 (the default implementation of RAND_status).
15573 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15574 to '-clrext' (= clear extensions), as intended and documented.
15575 *Bodo Moeller; inconsistency pointed out by Michael Attili
15576 <attili@amaxo.com>*
15578 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15579 was larger than the MD block size.
15581 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15583 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15584 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15585 using the passed key: if the passed key was a private key the result
15586 of X509_print(), for example, would be to print out all the private key
15591 * des_quad_cksum() byte order bug fix.
15592 *Ulf Möller, using the problem description in krb4-0.9.7, where
15593 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15595 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15598 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15600 * For easily testing in shell scripts whether some command
15601 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15602 returns with exit code 0 iff no command of the given name is available.
15603 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15604 the output goes to stdout and nothing is printed to stderr.
15605 Additional arguments are always ignored.
15607 Since for each cipher there is a command of the same name,
15608 the 'no-cipher' compilation switches can be tested this way.
15610 ('openssl no-XXX' is not able to detect pseudo-commands such
15611 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15615 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15619 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15620 is set; it will be thrown away anyway because each handshake creates
15622 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15623 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15624 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15625 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15629 * New s_client option -ign_eof: EOF at stdin is ignored, and
15630 'Q' and 'R' lose their special meanings (quit/renegotiate).
15631 This is part of what -quiet does; unlike -quiet, -ign_eof
15632 does not suppress any output.
15636 * Add compatibility options to the purpose and trust code. The
15637 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15638 accepts a certificate or CA, this was the previous behaviour,
15639 with all the associated security issues.
15641 X509_TRUST_COMPAT is the old trust behaviour: only and
15642 automatically trust self signed roots in certificate store. A
15643 new trust setting X509_TRUST_DEFAULT is used to specify that
15644 a purpose has no associated trust setting and it should instead
15645 use the value in the default purpose.
15649 * Fix the PKCS#8 DSA private key code so it decodes keys again
15650 and fix a memory leak.
15654 * In util/mkerr.pl (which implements 'make errors'), preserve
15655 reason strings from the previous version of the .c file, as
15656 the default to have only downcase letters (and digits) in
15657 automatically generated reasons codes is not always appropriate.
15661 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15662 using strerror. Previously, ERR_reason_error_string() returned
15663 library names as reason strings for SYSerr; but SYSerr is a special
15664 case where small numbers are errno values, not library numbers.
15668 * Add '-dsaparam' option to 'openssl dhparam' application. This
15669 converts DSA parameters into DH parameters. (When creating parameters,
15670 DSA_generate_parameters is used.)
15674 * Include 'length' (recommended exponent length) in C code generated
15675 by 'openssl dhparam -C'.
15679 * The second argument to set_label in perlasm was already being used
15680 so couldn't be used as a "file scope" flag. Moved to third argument
15685 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15686 instead of RAND_bytes for encryption IVs and salts.
15690 * Include RAND_status() into RAND_METHOD instead of implementing
15691 it only for md_rand.c Otherwise replacing the PRNG by calling
15692 RAND_set_rand_method would be impossible.
15696 * Don't let DSA_generate_key() enter an infinite loop if the random
15697 number generation fails.
15701 * New 'rand' application for creating pseudo-random output.
15705 * Added configuration support for Linux/IA64
15707 *Rolf Haberrecker <rolf@suse.de>*
15709 * Assembler module support for Mingw32.
15713 * Shared library support for HPUX (in shlib/).
15715 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15717 * Shared library support for Solaris gcc.
15719 *Lutz Behnke <behnke@trustcenter.de>*
15721 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15723 * PKCS7_encrypt() was adding text MIME headers twice because they
15724 were added manually and by SMIME_crlf_copy().
15728 * In bntest.c don't call BN_rand with zero bits argument.
15730 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15732 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15733 case was implemented. This caused BN_div_recp() to fail occasionally.
15737 * Add an optional second argument to the set_label() in the perl
15738 assembly language builder. If this argument exists and is set
15739 to 1 it signals that the assembler should use a symbol whose
15740 scope is the entire file, not just the current function. This
15741 is needed with MASM which uses the format label:: for this scope.
15743 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15745 * Change the ASN1 types so they are typedefs by default. Before
15746 almost all types were #define'd to ASN1_STRING which was causing
15747 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15752 * Change names of new functions to the new get1/get0 naming
15753 convention: After 'get1', the caller owns a reference count
15754 and has to call `..._free`; 'get0' returns a pointer to some
15755 data structure without incrementing reference counters.
15756 (Some of the existing 'get' functions increment a reference
15757 counter, some don't.)
15758 Similarly, 'set1' and 'add1' functions increase reference
15759 counters or duplicate objects.
15763 * Allow for the possibility of temp RSA key generation failure:
15764 the code used to assume it always worked and crashed on failure.
15768 * Fix potential buffer overrun problem in BIO_printf().
15769 *Ulf Möller, using public domain code by Patrick Powell; problem
15770 pointed out by David Sacerdote <das33@cornell.edu>*
15772 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15773 RAND_egd() and RAND_status(). In the command line application,
15774 the EGD socket can be specified like a seed file using RANDFILE
15779 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15780 Some CAs (e.g. Verisign) distribute certificates in this form.
15784 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15785 list to exclude them. This means that no special compilation option
15786 is needed to use anonymous DH: it just needs to be included in the
15791 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15792 EVP_MD_type. The old functionality is available in a new macro called
15793 EVP_MD_md(). Change code that uses it and update docs.
15797 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15798 where the `void *` argument is replaced by a function pointer argument.
15799 Previously `void *` was abused to point to functions, which works on
15800 many platforms, but is not correct. As these functions are usually
15801 called by macros defined in OpenSSL header files, most source code
15802 should work without changes.
15806 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15807 sections with information on -D... compiler switches used for
15808 compiling the library so that applications can see them. To enable
15809 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15810 must be defined. E.g.,
15811 #define OPENSSL_ALGORITHM_DEFINES
15812 #include <openssl/opensslconf.h>
15813 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15815 *Richard Levitte, Ulf and Bodo Möller*
15817 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15822 * Change the 'other' type in certificate aux info to a STACK_OF
15823 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15824 the required ASN1 format: arbitrary types determined by an OID.
15828 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15829 argument to 'req'. This is not because the function is newer or
15830 better than others it just uses the work 'NEW' in the certificate
15831 request header lines. Some software needs this.
15835 * Reorganise password command line arguments: now passwords can be
15836 obtained from various sources. Delete the PEM_cb function and make
15837 it the default behaviour: i.e. if the callback is NULL and the
15838 usrdata argument is not NULL interpret it as a null terminated pass
15839 phrase. If usrdata and the callback are NULL then the pass phrase
15840 is prompted for as usual.
15844 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15845 the support is automatically enabled. The resulting binaries will
15846 autodetect the card and use it if present.
15848 *Ben Laurie and Compaq Inc.*
15850 * Work around for Netscape hang bug. This sends certificate request
15851 and server done in one record. Since this is perfectly legal in the
15852 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15853 the bugs/SSLv3 entry for more info.
15857 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15861 * Add -rand argument to smime and pkcs12 applications and read/write
15866 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15870 * Add command line password options to the remaining applications.
15874 * Bug fix for BN_div_recp() for numerators with an even number of
15879 * More tests in bntest.c, and changed test_bn output.
15883 * ./config recognizes MacOS X now.
15887 * Bug fix for BN_div() when the first words of num and divisor are
15888 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15892 * Add support for various broken PKCS#8 formats, and command line
15893 options to produce them.
15897 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15898 get temporary BIGNUMs from a BN_CTX.
15902 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15907 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15908 include a #define from the old name to the new. The original intent
15909 was that statically linked binaries could for example just call
15910 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15911 link with digests. This never worked because SSLeay_add_all_digests()
15912 and SSLeay_add_all_ciphers() were in the same source file so calling
15913 one would link with the other. They are now in separate source files.
15917 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15921 * Use a less unusual form of the Miller-Rabin primality test (it used
15922 a binary algorithm for exponentiation integrated into the Miller-Rabin
15923 loop, our standard modexp algorithms are faster).
15927 * Support for the EBCDIC character set completed.
15929 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15931 * Source code cleanups: use const where appropriate, eliminate casts,
15932 use `void *` instead of `char *` in lhash.
15936 * Bugfix: ssl3_send_server_key_exchange was not restartable
15937 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15938 this the server could overwrite ephemeral keys that the client
15943 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15944 using 50 iterations of the Rabin-Miller test.
15946 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15947 iterations of the Rabin-Miller test as required by the appendix
15948 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15949 As BN_is_prime_fasttest includes trial division, DSA parameter
15950 generation becomes much faster.
15952 This implies a change for the callback functions in DSA_is_prime
15953 and DSA_generate_parameters: The callback function is called once
15954 for each positive witness in the Rabin-Miller test, not just
15955 occasionally in the inner loop; and the parameters to the
15956 callback function now provide an iteration count for the outer
15957 loop rather than for the current invocation of the inner loop.
15958 DSA_generate_parameters additionally can call the callback
15959 function with an 'iteration count' of -1, meaning that a
15960 candidate has passed the trial division test (when q is generated
15961 from an application-provided seed, trial division is skipped).
15965 * New function BN_is_prime_fasttest that optionally does trial
15966 division before starting the Rabin-Miller test and has
15967 an additional BN_CTX * argument (whereas BN_is_prime always
15968 has to allocate at least one BN_CTX).
15969 'callback(1, -1, cb_arg)' is called when a number has passed the
15970 trial division stage.
15974 * Fix for bug in CRL encoding. The validity dates weren't being handled
15979 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15983 * New function BN_pseudo_rand().
15987 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15988 bignum version of BN_from_montgomery() with the working code from
15989 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15994 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15995 made it impossible to use the same SSL_SESSION data structure in
15996 SSL2 clients in multiple threads.
16000 * The return value of RAND_load_file() no longer counts bytes obtained
16001 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16002 to seed the PRNG (previously an explicit byte count was required).
16004 *Ulf Möller, Bodo Möller*
16006 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16007 used `char *` instead of `void *` and had casts all over the place.
16011 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16015 * Retain source code compatibility for BN_prime_checks macro:
16016 BN_is_prime(..., BN_prime_checks, ...) now uses
16017 BN_prime_checks_for_size to determine the appropriate number of
16018 Rabin-Miller iterations.
16022 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16023 DH_CHECK_P_NOT_SAFE_PRIME.
16024 (Check if this is true? OpenPGP calls them "strong".)
16028 * Merge the functionality of "dh" and "gendh" programs into a new program
16029 "dhparam". The old programs are retained for now but will handle DH keys
16030 (instead of parameters) in future.
16034 * Make the ciphers, s_server and s_client programs check the return values
16035 when a new cipher list is set.
16039 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16040 ciphers. Before when the 56bit ciphers were enabled the sorting was
16043 The syntax for the cipher sorting has been extended to support sorting by
16044 cipher-strength (using the strength_bits hard coded in the tables).
16045 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16047 Fix a bug in the cipher-command parser: when supplying a cipher command
16048 string with an "undefined" symbol (neither command nor alphanumeric
16049 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16050 an error is flagged.
16052 Due to the strength-sorting extension, the code of the
16053 ssl_create_cipher_list() function was completely rearranged. I hope that
16054 the readability was also increased :-)
16056 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16058 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16059 for the first serial number and places 2 in the serial number file. This
16060 avoids problems when the root CA is created with serial number zero and
16061 the first user certificate has the same issuer name and serial number
16066 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16067 the new code. Add documentation for this stuff.
16071 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16072 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16073 structures and behave in an analogous way to the X509v3 functions:
16074 they shouldn't be called directly but wrapper functions should be used
16077 So we also now have some wrapper functions that call the X509at functions
16078 when passed certificate requests. (TO DO: similar things can be done with
16079 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16080 things. Some of these need some d2i or i2d and print functionality
16081 because they handle more complex structures.)
16085 * Add missing #ifndefs that caused missing symbols when building libssl
16086 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16087 NO_RSA in `ssl/s2*.c`.
16089 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16091 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16092 has a return value which indicates the quality of the random data
16093 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16094 error queue. New function RAND_pseudo_bytes() generates output that is
16095 guaranteed to be unique but not unpredictable. RAND_add is like
16096 RAND_seed, but takes an extra argument for an entropy estimate
16097 (RAND_seed always assumes full entropy).
16101 * Do more iterations of Rabin-Miller probable prime test (specifically,
16102 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16103 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16104 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16105 false-positive rate of at most 2^-80 for random input.
16109 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16113 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16114 in the 0.9.5 release), this returns the chain
16115 from an X509_CTX structure with a dup of the stack and all
16116 the X509 reference counts upped: so the stack will exist
16117 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16120 Also make SSL_SESSION_print() print out the verify return
16125 * Add manpage for the pkcs12 command. Also change the default
16126 behaviour so MAC iteration counts are used unless the new
16127 -nomaciter option is used. This improves file security and
16128 only older versions of MSIE (4.0 for example) need it.
16132 * Honor the no-xxx Configure options when creating .DEF files.
16136 * Add PKCS#10 attributes to field table: challengePassword,
16137 unstructuredName and unstructuredAddress. These are taken from
16138 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16139 international characters are used.
16141 More changes to X509_ATTRIBUTE code: allow the setting of types
16142 based on strings. Remove the 'loc' parameter when adding
16143 attributes because these will be a SET OF encoding which is sorted
16148 * Initial changes to the 'req' utility to allow request generation
16149 automation. This will allow an application to just generate a template
16150 file containing all the field values and have req construct the
16153 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16154 used all over the place including certificate requests and PKCS#7
16155 structures. They are currently handled manually where necessary with
16156 some primitive wrappers for PKCS#7. The new functions behave in a
16157 manner analogous to the X509 extension functions: they allow
16158 attributes to be looked up by NID and added.
16160 Later something similar to the X509V3 code would be desirable to
16161 automatically handle the encoding, decoding and printing of the
16162 more complex types. The string types like challengePassword can
16163 be handled by the string table functions.
16165 Also modified the multi byte string table handling. Now there is
16166 a 'global mask' which masks out certain types. The table itself
16167 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16168 is useful when for example there is only one permissible type
16169 (as in countryName) and using the mask might result in no valid
16174 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16175 SSL_get_peer_finished to allow applications to obtain the latest
16176 Finished messages sent to the peer or expected from the peer,
16177 respectively. (SSL_get_peer_finished is usually the Finished message
16178 actually received from the peer, otherwise the protocol will be aborted.)
16180 As the Finished message are message digests of the complete handshake
16181 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16182 be used for external authentication procedures when the authentication
16183 provided by SSL/TLS is not desired or is not enough.
16187 * Enhanced support for Alpha Linux is added. Now ./config checks if
16188 the host supports BWX extension and if Compaq C is present on the
16189 $PATH. Just exploiting of the BWX extension results in 20-30%
16190 performance kick for some algorithms, e.g. DES and RC4 to mention
16191 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16196 * Add support for MS "fast SGC". This is arguably a violation of the
16197 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16198 weak crypto and after checking the certificate is SGC a second one
16199 with strong crypto. MS SGC stops the first handshake after receiving
16200 the server certificate message and sends a second client hello. Since
16201 a server will typically do all the time consuming operations before
16202 expecting any further messages from the client (server key exchange
16203 is the most expensive) there is little difference between the two.
16205 To get OpenSSL to support MS SGC we have to permit a second client
16206 hello message after we have sent server done. In addition we have to
16207 reset the MAC if we do get this second client hello.
16211 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16212 if a DER encoded private key is RSA or DSA traditional format. Changed
16213 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16214 format DER encoded private key. Newer code should use PKCS#8 format which
16215 has the key type encoded in the ASN1 structure. Added DER private key
16216 support to pkcs8 application.
16220 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16221 ciphersuites has been selected (as required by the SSL 3/TLS 1
16222 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16223 is set, we interpret this as a request to violate the specification
16224 (the worst that can happen is a handshake failure, and 'correct'
16225 behaviour would result in a handshake failure anyway).
16229 * In SSL_CTX_add_session, take into account that there might be multiple
16230 SSL_SESSION structures with the same session ID (e.g. when two threads
16231 concurrently obtain them from an external cache).
16232 The internal cache can handle only one SSL_SESSION with a given ID,
16233 so if there's a conflict, we now throw out the old one to achieve
16238 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16239 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16240 some routines that use cipher OIDs: some ciphers do not have OIDs
16241 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16246 * Simplify the trust setting structure and code. Now we just have
16247 two sequences of OIDs for trusted and rejected settings. These will
16248 typically have values the same as the extended key usage extension
16249 and any application specific purposes.
16251 The trust checking code now has a default behaviour: it will just
16252 check for an object with the same NID as the passed id. Functions can
16253 be provided to override either the default behaviour or the behaviour
16254 for a given id. SSL client, server and email already have functions
16255 in place for compatibility: they check the NID and also return "trusted"
16256 if the certificate is self signed.
16260 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16261 traditional format into an EVP_PKEY structure.
16265 * Add a password callback function PEM_cb() which either prompts for
16266 a password if usr_data is NULL or otherwise assumes it is a null
16267 terminated password. Allow passwords to be passed on command line
16268 environment or config files in a few more utilities.
16272 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16273 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16274 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16275 Update documentation.
16279 * Support for ASN1 "NULL" type. This could be handled before by using
16280 ASN1_TYPE but there wasn't any function that would try to read a NULL
16281 and produce an error if it couldn't. For compatibility we also have
16282 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16283 don't allocate anything because they don't need to.
16287 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16290 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16292 * Rebuild of the memory allocation routines used by OpenSSL code and
16293 possibly others as well. The purpose is to make an interface that
16294 provide hooks so anyone can build a separate set of allocation and
16295 deallocation routines to be used by OpenSSL, for example memory
16296 pool implementations, or something else, which was previously hard
16297 since Malloc(), Realloc() and Free() were defined as macros having
16298 the values malloc, realloc and free, respectively (except for Win32
16299 compilations). The same is provided for memory debugging code.
16300 OpenSSL already comes with functionality to find memory leaks, but
16301 this gives people a chance to debug other memory problems.
16303 With these changes, a new set of functions and macros have appeared:
16305 CRYPTO_set_mem_debug_functions() [F]
16306 CRYPTO_get_mem_debug_functions() [F]
16307 CRYPTO_dbg_set_options() [F]
16308 CRYPTO_dbg_get_options() [F]
16309 CRYPTO_malloc_debug_init() [M]
16311 The memory debug functions are NULL by default, unless the library
16312 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16313 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16314 gives the standard debugging functions that come with OpenSSL) or
16315 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16316 provided by the library user) must be used. When the standard
16317 debugging functions are used, CRYPTO_dbg_set_options can be used to
16318 request additional information:
16319 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16320 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16322 Also, things like CRYPTO_set_mem_functions will always give the
16323 expected result (the new set of functions is used for allocation
16324 and deallocation) at all times, regardless of platform and compiler
16327 To finish it up, some functions that were never use in any other
16328 way than through macros have a new API and new semantic:
16330 CRYPTO_dbg_malloc()
16331 CRYPTO_dbg_realloc()
16334 All macros of value have retained their old syntax.
16336 *Richard Levitte and Bodo Moeller*
16338 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16339 ordering of SMIMECapabilities wasn't in "strength order" and there
16340 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16345 * Some ASN1 types with illegal zero length encoding (INTEGER,
16346 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16348 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16350 * Merge in my S/MIME library for OpenSSL. This provides a simple
16351 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16352 functionality to handle multipart/signed properly) and a utility
16353 called 'smime' to call all this stuff. This is based on code I
16354 originally wrote for Celo who have kindly allowed it to be
16355 included in OpenSSL.
16359 * Add variants des_set_key_checked and des_set_key_unchecked of
16360 des_set_key (aka des_key_sched). Global variable des_check_key
16361 decides which of these is called by des_set_key; this way
16362 des_check_key behaves as it always did, but applications and
16363 the library itself, which was buggy for des_check_key == 1,
16364 have a cleaner way to pick the version they need.
16368 * New function PKCS12_newpass() which changes the password of a
16373 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16374 dynamic mix. In both cases the ids can be used as an index into the
16375 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16376 functions so they accept a list of the field values and the
16377 application doesn't need to directly manipulate the X509_TRUST
16382 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16387 * Modify the way the V3 extension code looks up extensions. This now
16388 works in a similar way to the object code: we have some "standard"
16389 extensions in a static table which is searched with OBJ_bsearch()
16390 and the application can add dynamic ones if needed. The file
16391 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16392 updated whenever a new extension is added to the core code and kept
16393 in ext_nid order. There is a simple program 'tabtest.c' which checks
16394 this. New extensions are not added too often so this file can readily
16395 be maintained manually.
16397 There are two big advantages in doing things this way. The extensions
16398 can be looked up immediately and no longer need to be "added" using
16399 X509V3_add_standard_extensions(): this function now does nothing.
16400 Side note: I get *lots* of email saying the extension code doesn't
16401 work because people forget to call this function.
16402 Also no dynamic allocation is done unless new extensions are added:
16403 so if we don't add custom extensions there is no need to call
16404 X509V3_EXT_cleanup().
16408 * Modify enc utility's salting as follows: make salting the default. Add a
16409 magic header, so unsalted files fail gracefully instead of just decrypting
16410 to garbage. This is because not salting is a big security hole, so people
16411 should be discouraged from doing it.
16415 * Fixes and enhancements to the 'x509' utility. It allowed a message
16416 digest to be passed on the command line but it only used this
16417 parameter when signing a certificate. Modified so all relevant
16418 operations are affected by the digest parameter including the
16419 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16420 DSA key was used because it didn't fix the digest.
16424 * Initial certificate chain verify code. Currently tests the untrusted
16425 certificates for consistency with the verify purpose (which is set
16426 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16428 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16429 this is because it will reject chains with invalid extensions whereas
16430 every previous version of OpenSSL and SSLeay made no checks at all.
16432 Trust code: checks the root CA for the relevant trust settings. Trust
16433 settings have an initial value consistent with the verify purpose: e.g.
16434 if the verify purpose is for SSL client use it expects the CA to be
16435 trusted for SSL client use. However the default value can be changed to
16436 permit custom trust settings: one example of this would be to only trust
16437 certificates from a specific "secure" set of CAs.
16439 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16440 which should be used for version portability: especially since the
16441 verify structure is likely to change more often now.
16443 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16444 to set them. If not set then assume SSL clients will verify SSL servers
16447 Two new options to the verify program: -untrusted allows a set of
16448 untrusted certificates to be passed in and -purpose which sets the
16449 intended purpose of the certificate. If a purpose is set then the
16450 new chain verify code is used to check extension consistency.
16454 * Support for the authority information access extension.
16458 * Modify RSA and DSA PEM read routines to transparently handle
16459 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16460 public keys in a format compatible with certificate
16461 SubjectPublicKeyInfo structures. Unfortunately there were already
16462 functions called *_PublicKey_* which used various odd formats so
16463 these are retained for compatibility: however the DSA variants were
16464 never in a public release so they have been deleted. Changed dsa/rsa
16465 utilities to handle the new format: note no releases ever handled public
16466 keys so we should be OK.
16468 The primary motivation for this change is to avoid the same fiasco
16469 that dogs private keys: there are several incompatible private key
16470 formats some of which are standard and some OpenSSL specific and
16471 require various evil hacks to allow partial transparent handling and
16472 even then it doesn't work with DER formats. Given the option anything
16473 other than PKCS#8 should be dumped: but the other formats have to
16474 stay in the name of compatibility.
16476 With public keys and the benefit of hindsight one standard format
16477 is used which works with EVP_PKEY, RSA or DSA structures: though
16478 it clearly returns an error if you try to read the wrong kind of key.
16480 Added a -pubkey option to the 'x509' utility to output the public key.
16481 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16482 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16483 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16484 that do the same as the `EVP_PKEY_assign_*()` except they up the
16485 reference count of the added key (they don't "swallow" the
16490 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16491 CRLs would fail if the file contained no certificates or no CRLs:
16492 added a new function to read in both types and return the number
16493 read: this means that if none are read it will be an error. The
16494 DER versions of the certificate and CRL reader would always fail
16495 because it isn't possible to mix certificates and CRLs in DER format
16496 without choking one or the other routine. Changed this to just read
16497 a certificate: this is the best we can do. Also modified the code
16498 in `apps/verify.c` to take notice of return codes: it was previously
16499 attempting to read in certificates from NULL pointers and ignoring
16500 any errors: this is one reason why the cert and CRL reader seemed
16501 to work. It doesn't check return codes from the default certificate
16502 routines: these may well fail if the certificates aren't installed.
16506 * Code to support otherName option in GeneralName.
16510 * First update to verify code. Change the verify utility
16511 so it warns if it is passed a self signed certificate:
16512 for consistency with the normal behaviour. X509_verify
16513 has been modified to it will now verify a self signed
16514 certificate if *exactly* the same certificate appears
16515 in the store: it was previously impossible to trust a
16516 single self signed certificate. This means that:
16517 openssl verify ss.pem
16518 now gives a warning about a self signed certificate but
16519 openssl verify -CAfile ss.pem ss.pem
16524 * For servers, store verify_result in SSL_SESSION data structure
16525 (and add it to external session representation).
16526 This is needed when client certificate verifications fails,
16527 but an application-provided verification callback (set by
16528 SSL_CTX_set_cert_verify_callback) allows accepting the session
16529 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16530 but returns 1): When the session is reused, we have to set
16531 ssl->verify_result to the appropriate error code to avoid
16534 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16536 * Fix a bug in the new PKCS#7 code: it didn't consider the
16537 case in PKCS7_dataInit() where the signed PKCS7 structure
16538 didn't contain any existing data because it was being created.
16540 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16542 * Add a salt to the key derivation routines in enc.c. This
16543 forms the first 8 bytes of the encrypted file. Also add a
16544 -S option to allow a salt to be input on the command line.
16548 * New function X509_cmp(). Oddly enough there wasn't a function
16549 to compare two certificates. We do this by working out the SHA1
16550 hash and comparing that. X509_cmp() will be needed by the trust
16555 * SSL_get1_session() is like SSL_get_session(), but increments
16556 the reference count in the SSL_SESSION returned.
16558 *Geoff Thorpe <geoff@eu.c2.net>*
16560 * Fix for 'req': it was adding a null to request attributes.
16561 Also change the X509_LOOKUP and X509_INFO code to handle
16562 certificate auxiliary information.
16566 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16571 * Add the possibility to add extra information to the memory leak
16572 detecting output, to form tracebacks, showing from where each
16573 allocation was originated: CRYPTO_push_info("constant string") adds
16574 the string plus current file name and line number to a per-thread
16575 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16576 is like calling CYRPTO_pop_info() until the stack is empty.
16577 Also updated memory leak detection code to be multi-thread-safe.
16581 * Add options -text and -noout to pkcs7 utility and delete the
16582 encryption options which never did anything. Update docs.
16586 * Add options to some of the utilities to allow the pass phrase
16587 to be included on either the command line (not recommended on
16588 OSes like Unix) or read from the environment. Update the
16589 manpages and fix a few bugs.
16593 * Add a few manpages for some of the openssl commands.
16597 * Fix the -revoke option in ca. It was freeing up memory twice,
16598 leaking and not finding already revoked certificates.
16602 * Extensive changes to support certificate auxiliary information.
16603 This involves the use of X509_CERT_AUX structure and X509_AUX
16604 functions. An X509_AUX function such as PEM_read_X509_AUX()
16605 can still read in a certificate file in the usual way but it
16606 will also read in any additional "auxiliary information". By
16607 doing things this way a fair degree of compatibility can be
16608 retained: existing certificates can have this information added
16609 using the new 'x509' options.
16611 Current auxiliary information includes an "alias" and some trust
16612 settings. The trust settings will ultimately be used in enhanced
16613 certificate chain verification routines: currently a certificate
16614 can only be trusted if it is self signed and then it is trusted
16619 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16620 The problem was that one of the replacement routines had not been working
16621 since SSLeay releases. For now the offending routine has been replaced
16622 with non-optimised assembler. Even so, this now gives around 95%
16623 performance improvement for 1024 bit RSA signs.
16627 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16628 handling. Most clients have the effective key size in bits equal to
16629 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16630 A few however don't do this and instead use the size of the decrypted key
16631 to determine the RC2 key length and the AlgorithmIdentifier to determine
16632 the effective key length. In this case the effective key length can still
16633 be 40 bits but the key length can be 168 bits for example. This is fixed
16634 by manually forcing an RC2 key into the EVP_PKEY structure because the
16635 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16636 the key length and effective key length are equal.
16640 * Add a bunch of functions that should simplify the creation of
16641 X509_NAME structures. Now you should be able to do:
16642 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16643 and have it automatically work out the correct field type and fill in
16644 the structures. The more adventurous can try:
16645 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16646 and it will (hopefully) work out the correct multibyte encoding.
16650 * Change the 'req' utility to use the new field handling and multibyte
16651 copy routines. Before the DN field creation was handled in an ad hoc
16652 way in req, ca, and x509 which was rather broken and didn't support
16653 BMPStrings or UTF8Strings. Since some software doesn't implement
16654 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16655 using the dirstring_type option. See the new comment in the default
16656 openssl.cnf for more info.
16660 * Make crypto/rand/md_rand.c more robust:
16661 - Assure unique random numbers after fork().
16662 - Make sure that concurrent threads access the global counter and
16663 md serializably so that we never lose entropy in them
16664 or use exactly the same state in multiple threads.
16665 Access to the large state is not always serializable because
16666 the additional locking could be a performance killer, and
16667 md should be large enough anyway.
16671 * New file `apps/app_rand.c` with commonly needed functionality
16672 for handling the random seed file.
16674 Use the random seed file in some applications that previously did not:
16676 dsaparam -genkey (which also ignored its '-rand' option),
16679 x509 (when signing).
16680 Except on systems with /dev/urandom, it is crucial to have a random
16681 seed file at least for key creation, DSA signing, and for DH exchanges;
16682 for RSA signatures we could do without one.
16684 gendh and gendsa (unlike genrsa) used to read only the first byte
16685 of each file listed in the '-rand' option. The function as previously
16686 found in genrsa is now in app_rand.c and is used by all programs
16687 that support '-rand'.
16691 * In RAND_write_file, use mode 0600 for creating files;
16692 don't just chmod when it may be too late.
16696 * Report an error from X509_STORE_load_locations
16697 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16701 * New function ASN1_mbstring_copy() this copies a string in either
16702 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16703 into an ASN1_STRING type. A mask of permissible types is passed
16704 and it chooses the "minimal" type to use or an error if not type
16709 * Add function equivalents to the various macros in asn1.h. The old
16710 macros are retained with an `M_` prefix. Code inside the library can
16711 use the `M_` macros. External code (including the openssl utility)
16712 should *NOT* in order to be "shared library friendly".
16716 * Add various functions that can check a certificate's extensions
16717 to see if it usable for various purposes such as SSL client,
16718 server or S/MIME and CAs of these types. This is currently
16719 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16720 verification. Also added a -purpose flag to x509 utility to
16721 print out all the purposes.
16725 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16730 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16731 for, obtain and decode and extension and obtain its critical flag.
16732 This allows all the necessary extension code to be handled in a
16733 single function call.
16737 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16738 platforms. See crypto/rc4/rc4_enc.c for further details.
16742 * New -noout option to asn1parse. This causes no output to be produced
16743 its main use is when combined with -strparse and -out to extract data
16744 from a file (which may not be in ASN.1 format).
16748 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16749 when producing the local key id.
16751 *Richard Levitte <levitte@stacken.kth.se>*
16753 * New option -dhparam in s_server. This allows a DH parameter file to be
16754 stated explicitly. If it is not stated then it tries the first server
16755 certificate file. The previous behaviour hard coded the filename
16760 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16761 a public key to be input or output. For example:
16762 openssl rsa -in key.pem -pubout -out pubkey.pem
16763 Also added necessary DSA public key functions to handle this.
16767 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16768 in the message. This was handled by allowing
16769 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16771 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16773 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16774 to the end of the strings whereas this didn't. This would cause problems
16775 if strings read with d2i_ASN1_bytes() were later modified.
16777 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16779 * Fix for base64 decode bug. When a base64 bio reads only one line of
16780 data and it contains EOF it will end up returning an error. This is
16781 caused by input 46 bytes long. The cause is due to the way base64
16782 BIOs find the start of base64 encoded data. They do this by trying a
16783 trial decode on each line until they find one that works. When they
16784 do a flag is set and it starts again knowing it can pass all the
16785 data directly through the decoder. Unfortunately it doesn't reset
16786 the context it uses. This means that if EOF is reached an attempt
16787 is made to pass two EOFs through the context and this causes the
16788 resulting error. This can also cause other problems as well. As is
16789 usual with these problems it takes *ages* to find and the fix is
16790 trivial: move one line.
16792 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16794 * Ugly workaround to get s_client and s_server working under Windows. The
16795 old code wouldn't work because it needed to select() on sockets and the
16796 tty (for keypresses and to see if data could be written). Win32 only
16797 supports select() on sockets so we select() with a 1s timeout on the
16798 sockets and then see if any characters are waiting to be read, if none
16799 are present then we retry, we also assume we can always write data to
16800 the tty. This isn't nice because the code then blocks until we've
16801 received a complete line of data and it is effectively polling the
16802 keyboard at 1s intervals: however it's quite a bit better than not
16803 working at all :-) A dedicated Windows application might handle this
16804 with an event loop for example.
16808 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16809 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16810 will be called when RSA_sign() and RSA_verify() are used. This is useful
16811 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16812 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16813 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16814 This necessitated the support of an extra signature type NID_md5_sha1
16815 for SSL signatures and modifications to the SSL library to use it instead
16816 of calling RSA_public_decrypt() and RSA_private_encrypt().
16820 * Add new -verify -CAfile and -CApath options to the crl program, these
16821 will lookup a CRL issuers certificate and verify the signature in a
16822 similar way to the verify program. Tidy up the crl program so it
16823 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16824 less strict. It will now permit CRL extensions even if it is not
16825 a V2 CRL: this will allow it to tolerate some broken CRLs.
16829 * Initialize all non-automatic variables each time one of the openssl
16830 sub-programs is started (this is necessary as they may be started
16831 multiple times from the "OpenSSL>" prompt).
16833 *Lennart Bang, Bodo Moeller*
16835 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16836 removing all other RSA functionality (this is what NO_RSA does). This
16837 is so (for example) those in the US can disable those operations covered
16838 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16843 * Non-copying interface to BIO pairs.
16844 (still largely untested)
16848 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16849 ASCII string. This was handled independently in various places before.
16853 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16854 UTF8 strings a character at a time.
16858 * Use client_version from client hello to select the protocol
16859 (s23_srvr.c) and for RSA client key exchange verification
16860 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16864 * Add various utility functions to handle SPKACs, these were previously
16865 handled by poking round in the structure internals. Added new function
16866 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16867 print, verify and generate SPKACs. Based on an original idea from
16868 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16872 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16876 * Allow the config file extension section to be overwritten on the
16877 command line. Based on an original idea from Massimiliano Pala
16878 <madwolf@comune.modena.it>. The new option is called -extensions
16879 and can be applied to ca, req and x509. Also -reqexts to override
16880 the request extensions in req and -crlexts to override the crl extensions
16885 * Add new feature to the SPKAC handling in ca. Now you can include
16886 the same field multiple times by preceding it by "XXXX." for example:
16889 this is the same syntax as used in the req config file.
16893 * Allow certificate extensions to be added to certificate requests. These
16894 are specified in a 'req_extensions' option of the req section of the
16895 config file. They can be printed out with the -text option to req but
16896 are otherwise ignored at present.
16900 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16901 data read consists of only the final block it would not decrypted because
16902 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16903 A misplaced 'break' also meant the decrypted final block might not be
16904 copied until the next read.
16908 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16909 a few extra parameters to the DH structure: these will be useful if
16910 for example we want the value of 'q' or implement X9.42 DH.
16914 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16915 provides hooks that allow the default DSA functions or functions on a
16916 "per key" basis to be replaced. This allows hardware acceleration and
16917 hardware key storage to be handled without major modification to the
16918 library. Also added low level modexp hooks and CRYPTO_EX structure and
16919 associated functions.
16923 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16924 as "read only": it can't be written to and the buffer it points to will
16925 not be freed. Reading from a read only BIO is much more efficient than
16926 a normal memory BIO. This was added because there are several times when
16927 an area of memory needs to be read from a BIO. The previous method was
16928 to create a memory BIO and write the data to it, this results in two
16929 copies of the data and an O(n^2) reading algorithm. There is a new
16930 function BIO_new_mem_buf() which creates a read only memory BIO from
16931 an area of memory. Also modified the PKCS#7 routines to use read only
16936 * Bugfix: ssl23_get_client_hello did not work properly when called in
16937 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16938 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16939 but a retry condition occurred while trying to read the rest.
16943 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16944 NID_pkcs7_encrypted by default: this was wrong since this should almost
16945 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16946 the encrypted data type: this is a more sensible place to put it and it
16947 allows the PKCS#12 code to be tidied up that duplicated this
16952 * Changed obj_dat.pl script so it takes its input and output files on
16953 the command line. This should avoid shell escape redirection problems
16958 * Initial support for certificate extension requests, these are included
16959 in things like Xenroll certificate requests. Included functions to allow
16960 extensions to be obtained and added.
16964 * -crlf option to s_client and s_server for sending newlines as
16965 CRLF (as required by many protocols).
16969 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16971 * Install libRSAglue.a when OpenSSL is built with RSAref.
16973 *Ralf S. Engelschall*
16975 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16977 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16979 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16984 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16985 DH parameters/keys (q is lost during that conversion, but the resulting
16986 DH parameters contain its length).
16988 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16989 much faster than DH_generate_parameters (which creates parameters
16990 where `p = 2*q + 1`), and also the smaller q makes DH computations
16991 much more efficient (160-bit exponentiation instead of 1024-bit
16992 exponentiation); so this provides a convenient way to support DHE
16993 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16994 utter importance to use
16995 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16997 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16998 when such DH parameters are used, because otherwise small subgroup
16999 attacks may become possible!
17003 * Avoid memory leak in i2d_DHparams.
17007 * Allow the -k option to be used more than once in the enc program:
17008 this allows the same encrypted message to be read by multiple recipients.
17012 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17013 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17014 it will always use the numerical form of the OID, even if it has a short
17019 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17020 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17021 otherwise bn_mod_exp was called. In the case of hardware keys for example
17022 no private key components need be present and it might store extra data
17023 in the RSA structure, which cannot be accessed from bn_mod_exp.
17024 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17025 private key operations.
17029 * Added support for SPARC Linux.
17033 * pem_password_cb function type incompatibly changed from
17034 typedef int pem_password_cb(char *buf, int size, int rwflag);
17036 ....(char *buf, int size, int rwflag, void *userdata);
17037 so that applications can pass data to their callbacks:
17038 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17039 additional void * argument, which is just handed through whenever
17040 the password callback is called.
17042 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17044 New function SSL_CTX_set_default_passwd_cb_userdata.
17046 Compatibility note: As many C implementations push function arguments
17047 onto the stack in reverse order, the new library version is likely to
17048 interoperate with programs that have been compiled with the old
17049 pem_password_cb definition (PEM_whatever takes some data that
17050 happens to be on the stack as its last argument, and the callback
17051 just ignores this garbage); but there is no guarantee whatsoever that
17054 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17055 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17056 problems not only on Windows, but also on some Unix platforms.
17057 To avoid problematic command lines, these definitions are now in an
17058 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17059 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17063 * MIPS III/IV assembler module is reimplemented.
17067 * More DES library cleanups: remove references to srand/rand and
17068 delete an unused file.
17072 * Add support for the free Netwide assembler (NASM) under Win32,
17073 since not many people have MASM (ml) and it can be hard to obtain.
17074 This is currently experimental but it seems to work OK and pass all
17075 the tests. Check out INSTALL.W32 for info.
17079 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17080 without temporary keys kept an extra copy of the server key,
17081 and connections with temporary keys did not free everything in case
17086 * New function RSA_check_key and new openssl rsa option -check
17087 for verifying the consistency of RSA keys.
17089 *Ulf Moeller, Bodo Moeller*
17091 * Various changes to make Win32 compile work:
17092 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17093 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17094 comparison" warnings.
17095 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17099 * Add a debugging option to PKCS#5 v2 key generation function: when
17100 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17101 derived keys are printed to stderr.
17105 * Copy the flags in ASN1_STRING_dup().
17107 *Roman E. Pavlov <pre@mo.msk.ru>*
17109 * The x509 application mishandled signing requests containing DSA
17110 keys when the signing key was also DSA and the parameters didn't match.
17112 It was supposed to omit the parameters when they matched the signing key:
17113 the verifying software was then supposed to automatically use the CA's
17114 parameters if they were absent from the end user certificate.
17116 Omitting parameters is no longer recommended. The test was also
17117 the wrong way round! This was probably due to unusual behaviour in
17118 EVP_cmp_parameters() which returns 1 if the parameters match.
17119 This meant that parameters were omitted when they *didn't* match and
17120 the certificate was useless. Certificates signed with 'ca' didn't have
17123 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17125 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17126 The interface is as follows:
17127 Applications can use
17128 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17129 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17130 "off" is now the default.
17131 The library internally uses
17132 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17133 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17134 to disable memory-checking temporarily.
17136 Some inconsistent states that previously were possible (and were
17137 even the default) are now avoided.
17139 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17140 with each memory chunk allocated; this is occasionally more helpful
17141 than just having a counter.
17143 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17145 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17150 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17151 which largely parallels "options", but is for changing API behaviour,
17152 whereas "options" are about protocol behaviour.
17153 Initial "mode" flags are:
17155 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17156 a single record has been written.
17157 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17158 retries use the same buffer location.
17159 (But all of the contents must be
17164 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17167 * Fix problems with no-hmac etc.
17169 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17171 * New functions RSA_get_default_method(), RSA_set_method() and
17172 RSA_get_method(). These allows replacement of RSA_METHODs without having
17173 to mess around with the internals of an RSA structure.
17177 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17178 Also really enable memory leak checks in openssl.c and in some
17181 *Chad C. Mulligan, Bodo Moeller*
17183 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17184 up the length of negative integers. This has now been simplified to just
17185 store the length when it is first determined and use it later, rather
17186 than trying to keep track of where data is copied and updating it to
17188 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17190 * Add a new function PKCS7_signatureVerify. This allows the verification
17191 of a PKCS#7 signature but with the signing certificate passed to the
17192 function itself. This contrasts with PKCS7_dataVerify which assumes the
17193 certificate is present in the PKCS#7 structure. This isn't always the
17194 case: certificates can be omitted from a PKCS#7 structure and be
17195 distributed by "out of band" means (such as a certificate database).
17199 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17200 function prototypes in pem.h, also change util/mkdef.pl to add the
17201 necessary function names.
17205 * mk1mf.pl (used by Windows builds) did not properly read the
17206 options set by Configure in the top level Makefile, and Configure
17207 was not even able to write more than one option correctly.
17208 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17212 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17213 file to be loaded from a BIO or FILE pointer. The BIO version will
17214 for example allow memory BIOs to contain config info.
17218 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17219 Whoever hopes to achieve shared-library compatibility across versions
17220 must use this, not the compile-time macro.
17221 (Exercise 0.9.4: Which is the minimum library version required by
17223 Note: All this applies only to multi-threaded programs, others don't
17228 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17229 through a BIO pair triggered the default case, i.e.
17230 SSLerr(...,SSL_R_UNKNOWN_STATE).
17234 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17235 can use the SSL library even if none of the specific BIOs is
17240 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17241 for the encoded length.
17243 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17245 * Add initial documentation of the X509V3 functions.
17249 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17250 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17251 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17252 secure PKCS#8 private key format with a high iteration count.
17256 * Fix determination of Perl interpreter: A perl or perl5
17257 *directory* in $PATH was also accepted as the interpreter.
17259 *Ralf S. Engelschall*
17261 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17262 wrong with it but it was very old and did things like calling
17263 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17264 unusual formatting.
17268 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17269 to use the new extension code.
17273 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17274 with macros. This should make it easier to change their form, add extra
17275 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17280 * Add to configuration table a new entry that can specify an alternative
17281 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17282 according to Mark Crispin <MRC@Panda.COM>.
17286 * DES CBC did not update the IV. Weird.
17290 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17291 Changing the behaviour of the former might break existing programs --
17292 where IV updating is needed, des_ncbc_encrypt can be used.
17295 * When bntest is run from "make test" it drives bc to check its
17296 calculations, as well as internally checking them. If an internal check
17297 fails, it needs to cause bc to give a non-zero result or make test carries
17298 on without noticing the failure. Fixed.
17302 * DES library cleanups.
17306 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17307 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17308 ciphers. NOTE: although the key derivation function has been verified
17309 against some published test vectors it has not been extensively tested
17310 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17315 * Instead of "mkdir -p", which is not fully portable, use new
17316 Perl script "util/mkdir-p.pl".
17320 * Rewrite the way password based encryption (PBE) is handled. It used to
17321 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17322 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17323 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17324 the 'parameter' field of the AlgorithmIdentifier is passed to the
17325 underlying key generation function so it must do its own ASN1 parsing.
17326 This has also changed the EVP_PBE_CipherInit() function which now has a
17327 'parameter' argument instead of literal salt and iteration count values
17328 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17332 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17333 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17334 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17335 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17336 value was just used as a "magic string" and not used directly its
17337 value doesn't matter.
17341 * Introduce some semblance of const correctness to BN. Shame C doesn't
17346 * "linux-sparc64" configuration (ultrapenguin).
17348 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17349 "linux-sparc" configuration.
17351 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17353 * config now generates no-xxx options for missing ciphers.
17357 * Support the EBCDIC character set (work in progress).
17358 File ebcdic.c not yet included because it has a different license.
17360 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17362 * Support BS2000/OSD-POSIX.
17364 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17366 * Make callbacks for key generation use `void *` instead of `char *`.
17370 * Make S/MIME samples compile (not yet tested).
17374 * Additional typesafe stacks.
17378 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17382 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17384 * New configuration variant "sco5-gcc".
17386 * Updated some demos.
17388 *Sean O Riordain, Wade Scholine*
17390 * Add missing BIO_free at exit of pkcs12 application.
17394 * Fix memory leak in conf.c.
17398 * Updates for Win32 to assembler version of MD5.
17402 * Set #! path to perl in `apps/der_chop` to where we found it
17403 instead of using a fixed path.
17407 * SHA library changes for irix64-mips4-cc.
17411 * Improvements for VMS support.
17415 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17417 * Bignum library bug fix. IRIX 6 passes "make test" now!
17418 This also avoids the problems with SC4.2 and unpatched SC5.
17420 *Andy Polyakov <appro@fy.chalmers.se>*
17422 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17423 These are required because of the typesafe stack would otherwise break
17424 existing code. If old code used a structure member which used to be STACK
17425 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17426 sk_num or sk_value it would produce an error because the num, data members
17427 are not present in STACK_OF. Now it just produces a warning. sk_set
17428 replaces the old method of assigning a value to sk_value
17429 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17430 that does this will no longer work (and should use sk_set instead) but
17431 this could be regarded as a "questionable" behaviour anyway.
17435 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17436 correctly handle encrypted S/MIME data.
17440 * Change type of various DES function arguments from des_cblock
17441 (which means, in function argument declarations, pointer to char)
17442 to des_cblock * (meaning pointer to array with 8 char elements),
17443 which allows the compiler to do more typechecking; it was like
17444 that back in SSLeay, but with lots of ugly casts.
17446 Introduce new type const_des_cblock.
17450 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17451 problems: find RecipientInfo structure that matches recipient certificate
17452 and initialise the ASN1 structures properly based on passed cipher.
17456 * Belatedly make the BN tests actually check the results.
17460 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17461 to and from BNs: it was completely broken. New compilation option
17462 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17463 key elements as negative integers.
17467 * Reorganize and speed up MD5.
17469 *Andy Polyakov <appro@fy.chalmers.se>*
17473 *Richard Levitte <richard@levitte.org>*
17475 * New option -out to asn1parse to allow the parsed structure to be
17476 output to a file. This is most useful when combined with the -strparse
17477 option to examine the output of things like OCTET STRINGS.
17481 * Make SSL library a little more fool-proof by not requiring any longer
17482 that `SSL_set_{accept,connect}_state` be called before
17483 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17484 in many applications because usually everything *appeared* to work as
17485 intended anyway -- now it really works as intended).
17489 * Move openssl.cnf out of lib/.
17493 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17494 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17495 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17497 *Ralf S. Engelschall*
17499 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17500 handle PKCS#7 enveloped data properly.
17502 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17504 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17505 copying pointers. The cert_st handling is changed by this in
17506 various ways (and thus what used to be known as ctx->default_cert
17507 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17508 any longer when s->cert does not give us what we need).
17509 ssl_cert_instantiate becomes obsolete by this change.
17510 As soon as we've got the new code right (possibly it already is?),
17511 we have solved a couple of bugs of the earlier code where s->cert
17512 was used as if it could not have been shared with other SSL structures.
17514 Note that using the SSL API in certain dirty ways now will result
17515 in different behaviour than observed with earlier library versions:
17516 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17517 does not influence s as it used to.
17519 In order to clean up things more thoroughly, inside SSL_SESSION
17520 we don't use CERT any longer, but a new structure SESS_CERT
17521 that holds per-session data (if available); currently, this is
17522 the peer's certificate chain and, for clients, the server's certificate
17523 and temporary key. CERT holds only those values that can have
17524 meaningful defaults in an SSL_CTX.
17528 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17529 from the internal representation. Various PKCS#7 fixes: remove some
17530 evil casts and set the enc_dig_alg field properly based on the signing
17535 * Allow PKCS#12 password to be set from the command line or the
17536 environment. Let 'ca' get its config file name from the environment
17537 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17542 * Allow certificate policies extension to use an IA5STRING for the
17543 organization field. This is contrary to the PKIX definition but
17544 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17549 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17550 without disallowing inline assembler and the like for non-pedantic builds.
17554 * Support Borland C++ builder.
17556 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17562 * SHA-1 cleanups and performance enhancements.
17564 *Andy Polyakov <appro@fy.chalmers.se>*
17566 * Sparc v8plus assembler for the bignum library.
17568 *Andy Polyakov <appro@fy.chalmers.se>*
17570 * Accept any -xxx and +xxx compiler options in Configure.
17574 * Update HPUX configuration.
17578 * Add missing `sk_<type>_unshift()` function to safestack.h
17580 *Ralf S. Engelschall*
17582 * New function SSL_CTX_use_certificate_chain_file that sets the
17583 "extra_cert"s in addition to the certificate. (This makes sense
17584 only for "PEM" format files, as chains as a whole are not
17589 * Support verify_depth from the SSL API.
17590 x509_vfy.c had what can be considered an off-by-one-error:
17591 Its depth (which was not part of the external interface)
17592 was actually counting the number of certificates in a chain;
17593 now it really counts the depth.
17597 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17598 instead of X509err, which often resulted in confusing error
17599 messages since the error codes are not globally unique
17600 (e.g. an alleged error in ssl3_accept when a certificate
17601 didn't match the private key).
17603 * New function SSL_CTX_set_session_id_context that allows to set a default
17604 value (so that you don't need SSL_set_session_id_context for each
17605 connection using the SSL_CTX).
17609 * OAEP decoding bug fix.
17613 * Support INSTALL_PREFIX for package builders, as proposed by
17618 * New Configure options "threads" and "no-threads". For systems
17619 where the proper compiler options are known (currently Solaris
17620 and Linux), "threads" is the default.
17624 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17628 * Install various scripts to $(OPENSSLDIR)/misc, not to
17629 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17630 such as /usr/local/bin.
17634 * "make linux-shared" to build shared libraries.
17636 *Niels Poppe <niels@netbox.org>*
17638 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17642 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17643 extension adding in x509 utility.
17647 * Remove NOPROTO sections and error code comments.
17651 * Partial rewrite of the DEF file generator to now parse the ANSI
17656 * New Configure options --prefix=DIR and --openssldir=DIR.
17660 * Complete rewrite of the error code script(s). It is all now handled
17661 by one script at the top level which handles error code gathering,
17662 header rewriting and C source file generation. It should be much better
17663 than the old method: it now uses a modified version of Ulf's parser to
17664 read the ANSI prototypes in all header files (thus the old K&R definitions
17665 aren't needed for error creation any more) and do a better job of
17666 translating function codes into names. The old 'ASN1 error code embedded
17667 in a comment' is no longer necessary and it doesn't use .err files which
17668 have now been deleted. Also the error code call doesn't have to appear all
17669 on one line (which resulted in some large lines...).
17673 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17677 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17678 0 (which usually indicates a closed connection), but continue reading.
17682 * Fix some race conditions.
17686 * Add support for CRL distribution points extension. Add Certificate
17687 Policies and CRL distribution points documentation.
17691 * Move the autogenerated header file parts to crypto/opensslconf.h.
17695 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17696 8 of keying material. Merlin has also confirmed interop with this fix
17697 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17699 *Merlin Hughes <merlin@baltimore.ie>*
17701 * Fix lots of warnings.
17703 *Richard Levitte <levitte@stacken.kth.se>*
17705 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17706 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17708 *Richard Levitte <levitte@stacken.kth.se>*
17710 * Fix problems with sizeof(long) == 8.
17712 *Andy Polyakov <appro@fy.chalmers.se>*
17714 * Change functions to ANSI C.
17718 * Fix typos in error codes.
17720 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17722 * Remove defunct assembler files from Configure.
17726 * SPARC v8 assembler BIGNUM implementation.
17728 *Andy Polyakov <appro@fy.chalmers.se>*
17730 * Support for Certificate Policies extension: both print and set.
17731 Various additions to support the r2i method this uses.
17735 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17736 return a const string when you are expecting an allocated buffer.
17740 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17741 types DirectoryString and DisplayText.
17745 * Add code to allow r2i extensions to access the configuration database,
17746 add an LHASH database driver and add several ctx helper functions.
17750 * Fix an evil bug in bn_expand2() which caused various BN functions to
17751 fail when they extended the size of a BIGNUM.
17755 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17756 support typesafe stack.
17760 * Fix typo in SSL_[gs]et_options().
17762 *Nils Frostberg <nils@medcom.se>*
17764 * Delete various functions and files that belonged to the (now obsolete)
17765 old X509V3 handling code.
17769 * New Configure option "rsaref".
17773 * Don't auto-generate pem.h.
17777 * Introduce type-safe ASN.1 SETs.
17781 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17783 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17785 * Introduce type-safe STACKs. This will almost certainly break lots of code
17786 that links with OpenSSL (well at least cause lots of warnings), but fear
17787 not: the conversion is trivial, and it eliminates loads of evil casts. A
17788 few STACKed things have been converted already. Feel free to convert more.
17789 In the fullness of time, I'll do away with the STACK type altogether.
17793 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17794 specified in `<certfile>` by updating the entry in the index.txt file.
17795 This way one no longer has to edit the index.txt file manually for
17796 revoking a certificate. The -revoke option does the gory details now.
17798 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17800 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17801 `-text` option at all and this way the `-noout -text` combination was
17802 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17804 *Ralf S. Engelschall*
17806 * Make sure a corresponding plain text error message exists for the
17807 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17808 verify callback function determined that a certificate was revoked.
17810 *Ralf S. Engelschall*
17812 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17813 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17814 all available ciphers including rc5, which was forgotten until now.
17815 In order to let the testing shell script know which algorithms
17816 are available, a new (up to now undocumented) command
17817 `openssl list-cipher-commands` is used.
17821 * Bugfix: s_client occasionally would sleep in select() when
17822 it should have checked SSL_pending() first.
17826 * New functions DSA_do_sign and DSA_do_verify to provide access to
17827 the raw DSA values prior to ASN.1 encoding.
17831 * Tweaks to Configure
17833 *Niels Poppe <niels@netbox.org>*
17835 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17840 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17844 * New config option to avoid instructions that are illegal on the 80386.
17845 The default code is faster, but requires at least a 486.
17849 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17850 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17851 same as SSL2_VERSION anyway.
17855 * New "-showcerts" option for s_client.
17859 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17860 application. Various cleanups and fixes.
17864 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17865 modify error routines to work internally. Add error codes and PBE init
17866 to library startup routines.
17870 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17871 packing functions to asn1 and evp. Changed function names and error
17872 codes along the way.
17876 * PKCS12 integration: and so it begins... First of several patches to
17877 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17878 objects to objects.h
17882 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17883 and display support for Thawte strong extranet extension.
17887 * Add LinuxPPC support.
17889 *Jeff Dubrule <igor@pobox.org>*
17891 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17892 bn_div_words in alpha.s.
17894 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17896 * Make sure the RSA OAEP test is skipped under -DRSAref because
17897 OAEP isn't supported when OpenSSL is built with RSAref.
17899 *Ulf Moeller <ulf@fitug.de>*
17901 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17902 so they no longer are missing under -DNOPROTO.
17904 *Soren S. Jorvang <soren@t.dk>*
17906 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17908 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17909 doesn't work when the session is reused. Coming soon!
17913 * Fix a security hole, that allows sessions to be reused in the wrong
17914 context thus bypassing client cert protection! All software that uses
17915 client certs and session caches in multiple contexts NEEDS PATCHING to
17916 allow session reuse! A fuller solution is in the works.
17918 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17920 * Some more source tree cleanups (removed obsolete files
17921 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17922 permission on "config" script to be executable) and a fix for the INSTALL
17925 *Ulf Moeller <ulf@fitug.de>*
17927 * Remove some legacy and erroneous uses of malloc, free instead of
17930 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17932 * Make rsa_oaep_test return non-zero on error.
17934 *Ulf Moeller <ulf@fitug.de>*
17936 * Add support for native Solaris shared libraries. Configure
17937 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17938 if someone would make that last step automatic.
17940 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17942 * ctx_size was not built with the right compiler during "make links". Fixed.
17946 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17947 except NULL ciphers". This means the default cipher list will no longer
17948 enable NULL ciphers. They need to be specifically enabled e.g. with
17949 the string "DEFAULT:eNULL".
17953 * Fix to RSA private encryption routines: if p < q then it would
17954 occasionally produce an invalid result. This will only happen with
17955 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17959 * Be less restrictive and allow also `perl util/perlpath.pl
17960 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17961 because this way one can also use an interpreter named `perl5` (which is
17962 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17963 installed as `perl`).
17965 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17967 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17969 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17971 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17972 advapi32.lib to Win32 build and change the pem test comparison
17973 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17974 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17975 and crypto/des/ede_cbcm_enc.c.
17979 * DES quad checksum was broken on big-endian architectures. Fixed.
17983 * Comment out two functions in bio.h that aren't implemented. Fix up the
17984 Win32 test batch file so it (might) work again. The Win32 test batch file
17985 is horrible: I feel ill....
17989 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17990 in e_os.h. Audit of header files to check ANSI and non ANSI
17991 sections: 10 functions were absent from non ANSI section and not exported
17992 from Windows DLLs. Fixed up libeay.num for new functions.
17996 * Make `openssl version` output lines consistent.
17998 *Ralf S. Engelschall*
18000 * Fix Win32 symbol export lists for BIO functions: Added
18001 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18002 to ms/libeay{16,32}.def.
18004 *Ralf S. Engelschall*
18006 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18007 fine under Unix and passes some trivial tests I've now added. But the
18008 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18009 added to make sure no one expects that this stuff really works in the
18010 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18011 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18014 *Ralf S. Engelschall*
18016 * Fix the generation of two part addresses in perl.
18018 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18020 * Add config entry for Linux on MIPS.
18022 *John Tobey <jtobey@channel1.com>*
18024 * Make links whenever Configure is run, unless we are on Windoze.
18028 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18029 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18034 * Add a useful kludge to allow package maintainers to specify compiler and
18035 other platforms details on the command line without having to patch the
18036 Configure script every time: One now can use
18037 `perl Configure <id>:<details>`,
18038 i.e. platform ids are allowed to have details appended
18039 to them (separated by colons). This is treated as there would be a static
18040 pre-configured entry in Configure's %table under key `<id>` with value
18041 `<details>` and `perl Configure <id>` is called. So, when you want to
18042 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18043 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18044 now, which overrides the FreeBSD-elf entry on-the-fly.
18046 *Ralf S. Engelschall*
18048 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18052 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18053 on the `perl Configure ...` command line. This way one can compile
18054 OpenSSL libraries with Position Independent Code (PIC) which is needed
18055 for linking it into DSOs.
18057 *Ralf S. Engelschall*
18059 * Remarkably, export ciphers were totally broken and no-one had noticed!
18064 * Cleaned up the LICENSE document: The official contact for any license
18065 questions now is the OpenSSL core team under openssl-core@openssl.org.
18066 And add a paragraph about the dual-license situation to make sure people
18067 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18068 to the OpenSSL toolkit.
18070 *Ralf S. Engelschall*
18072 * General source tree makefile cleanups: Made `making xxx in yyy...`
18073 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18074 Additionally cleaned up the `make links` target: Remove unnecessary
18075 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18076 to speed processing and no longer clutter the display with confusing
18077 stuff. Instead only the actually done links are displayed.
18079 *Ralf S. Engelschall*
18081 * Permit null encryption ciphersuites, used for authentication only. It used
18082 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18083 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18088 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18089 signed attributes when verifying signatures (this would break them),
18090 the detached data encoding was wrong and public keys obtained using
18091 X509_get_pubkey() weren't freed.
18095 * Add text documentation for the BUFFER functions. Also added a work around
18096 to a Win95 console bug. This was triggered by the password read stuff: the
18097 last character typed gets carried over to the next fread(). If you were
18098 generating a new cert request using 'req' for example then the last
18099 character of the passphrase would be CR which would then enter the first
18104 * Added the new 'Includes OpenSSL Cryptography Software' button as
18105 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18106 button and can be used by applications based on OpenSSL to show the
18107 relationship to the OpenSSL project.
18109 *Ralf S. Engelschall*
18111 * Remove confusing variables in function signatures in files
18112 ssl/ssl_lib.c and ssl/ssl.h.
18114 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18116 * Don't install bss_file.c under PREFIX/include/
18118 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18120 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18121 functions that return function pointers and has support for NT specific
18122 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18123 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18124 unsigned to signed types: this was killing the Win32 compile.
18128 * Add new certificate file to stack functions,
18129 SSL_add_dir_cert_subjects_to_stack() and
18130 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18131 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18132 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18133 This means that Apache-SSL and similar packages don't have to mess around
18134 to add as many CAs as they want to the preferred list.
18138 * Experiment with doxygen documentation. Currently only partially applied to
18140 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18141 openssl.doxy as the configuration file.
18145 * Get rid of remaining C++-style comments which strict C compilers hate.
18147 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18149 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18150 compiled in by default: it has problems with large keys.
18154 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18155 DH private keys and/or callback functions which directly correspond to
18156 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18157 is needed for applications which have to configure certificates on a
18158 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18160 For the RSA certificate situation is makes no difference, but
18161 for the DSA certificate situation this fixes the "no shared cipher"
18162 problem where the OpenSSL cipher selection procedure failed because the
18163 temporary keys were not overtaken from the context and the API provided
18164 no way to reconfigure them.
18165 The new functions now let applications reconfigure the stuff and they
18166 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18167 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18168 non-public-API function ssl_cert_instantiate() is used as a helper
18169 function and also to reduce code redundancy inside ssl_rsa.c.
18171 *Ralf S. Engelschall*
18173 * Move s_server -dcert and -dkey options out of the undocumented feature
18174 area because they are useful for the DSA situation and should be
18175 recognized by the users.
18177 *Ralf S. Engelschall*
18179 * Fix the cipher decision scheme for export ciphers: the export bits are
18180 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18181 SSL_EXP_MASK. So, the original variable has to be used instead of the
18182 already masked variable.
18184 *Richard Levitte <levitte@stacken.kth.se>*
18186 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18188 *Richard Levitte <levitte@stacken.kth.se>*
18190 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18191 from `int` to `unsigned int` because it is a length and initialized by
18192 EVP_DigestFinal() which expects an `unsigned int *`.
18194 *Richard Levitte <levitte@stacken.kth.se>*
18196 * Don't hard-code path to Perl interpreter on shebang line of Configure
18197 script. Instead use the usual Shell->Perl transition trick.
18199 *Ralf S. Engelschall*
18201 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18202 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18203 -noout -modulus` as it's already the case for `openssl rsa -noout
18204 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18205 currently the public key is printed (a decision which was already done by
18206 `openssl dsa -modulus` in the past) which serves a similar purpose.
18207 Additionally the NO_RSA no longer completely removes the whole -modulus
18208 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18211 *Ralf S. Engelschall*
18213 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18214 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18216 *Arne Ansper <arne@ats.cyber.ee>*
18218 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18219 to be added. Now both 'req' and 'ca' can use new objects defined in the
18224 * Add cool BIO that does syslog (or event log on NT).
18226 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18228 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18229 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18230 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18231 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18235 * Add preliminary config info for new extension code.
18239 * Make RSA_NO_PADDING really use no padding.
18241 *Ulf Moeller <ulf@fitug.de>*
18243 * Generate errors when private/public key check is done.
18247 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18248 for some CRL extensions and new objects added.
18252 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18253 key usage extension and fuller support for authority key id.
18257 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18258 padding method for RSA, which is recommended for new applications in PKCS
18259 #1 v2.0 (RFC 2437, October 1998).
18260 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18261 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18262 against Bleichbacher's attack on RSA.
18263 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18266 * Updates to the new SSL compression code
18268 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18270 * Fix so that the version number in the master secret, when passed
18271 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18272 (because the server will not accept higher), that the version number
18273 is 0x03,0x01, not 0x03,0x00
18275 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18277 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18278 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18279 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18283 * Support for RAW extensions where an arbitrary extension can be
18284 created by including its DER encoding. See `apps/openssl.cnf` for
18289 * Make sure latest Perl versions don't interpret some generated C array
18290 code as Perl array code in the crypto/err/err_genc.pl script.
18292 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18294 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18295 not many people have the assembler. Various Win32 compilation fixes and
18296 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18297 build instructions.
18301 * Modify configure script 'Configure' to automatically create crypto/date.h
18302 file under Win32 and also build pem.h from pem.org. New script
18303 util/mkfiles.pl to create the MINFO file on environments that can't do a
18304 'make files': perl util/mkfiles.pl >MINFO should work.
18308 * Major rework of DES function declarations, in the pursuit of correctness
18309 and purity. As a result, many evil casts evaporated, and some weirdness,
18310 too. You may find this causes warnings in your code. Zapping your evil
18311 casts will probably fix them. Mostly.
18315 * Fix for a typo in asn1.h. Bug fix to object creation script
18316 obj_dat.pl. It considered a zero in an object definition to mean
18317 "end of object": none of the objects in objects.h have any zeros
18318 so it wasn't spotted.
18320 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18322 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18323 Masking (CBCM). In the absence of test vectors, the best I have been able
18324 to do is check that the decrypt undoes the encrypt, so far. Send me test
18325 vectors if you have them.
18329 * Correct calculation of key length for export ciphers (too much space was
18330 allocated for null ciphers). This has not been tested!
18334 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18335 message is now correct (it understands "crypto" and "ssl" on its
18336 command line). There is also now an "update" option. This will update
18337 the util/ssleay.num and util/libeay.num files with any new functions.
18339 perl util/mkdef.pl crypto ssl update
18340 it will update them.
18344 * Overhauled the Perl interface:
18345 - ported BN stuff to OpenSSL's different BN library
18346 - made the perl/ source tree CVS-aware
18347 - renamed the package from SSLeay to OpenSSL (the files still contain
18348 their history because I've copied them in the repository)
18349 - removed obsolete files (the test scripts will be replaced
18350 by better Test::Harness variants in the future)
18352 *Ralf S. Engelschall*
18354 * First cut for a very conservative source tree cleanup:
18355 1. merge various obsolete readme texts into doc/ssleay.txt
18356 where we collect the old documents and readme texts.
18357 2. remove the first part of files where I'm already sure that we no
18358 longer need them because of three reasons: either they are just temporary
18359 files which were left by Eric or they are preserved original files where
18360 I've verified that the diff is also available in the CVS via "cvs diff
18361 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18362 the crypto/md/ stuff).
18364 *Ralf S. Engelschall*
18366 * More extension code. Incomplete support for subject and issuer alt
18367 name, issuer and authority key id. Change the i2v function parameters
18368 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18369 what that's for :-) Fix to ASN1 macro which messed up
18370 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18374 * Preliminary support for ENUMERATED type. This is largely copied from the
18379 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18381 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18383 * Make sure `make rehash` target really finds the `openssl` program.
18385 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18387 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18388 like to hear about it if this slows down other processors.
18392 * Add CygWin32 platform information to Configure script.
18394 *Alan Batie <batie@aahz.jf.intel.com>*
18396 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18398 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18400 * New program nseq to manipulate netscape certificate sequences
18404 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18409 * Fixes to BN code. Previously the default was to define BN_RECURSION
18410 but the BN code had some problems that would cause failures when
18411 doing certificate verification and some other functions.
18413 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18415 * Add ASN1 and PEM code to support netscape certificate sequences.
18419 * Add ASN1 and PEM code to support netscape certificate sequences.
18423 * Add several PKIX and private extended key usage OIDs.
18427 * Modify the 'ca' program to handle the new extension code. Modify
18428 openssl.cnf for new extension format, add comments.
18432 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18433 and add a sample to openssl.cnf so req -x509 now adds appropriate
18438 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18439 error code, add initial support to X509_print() and x509 application.
18443 * Takes a deep breath and start adding X509 V3 extension support code. Add
18444 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18445 stuff is currently isolated and isn't even compiled yet.
18449 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18450 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18451 Removed the versions check from X509 routines when loading extensions:
18452 this allows certain broken certificates that don't set the version
18453 properly to be processed.
18457 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18458 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18459 can still be regenerated with "make depend".
18463 * Spelling mistake in C version of CAST-128.
18465 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18467 * Changes to the error generation code. The perl script err-code.pl
18468 now reads in the old error codes and retains the old numbers, only
18469 adding new ones if necessary. It also only changes the .err files if new
18470 codes are added. The makefiles have been modified to only insert errors
18471 when needed (to avoid needlessly modifying header files). This is done
18472 by only inserting errors if the .err file is newer than the auto generated
18473 C file. To rebuild all the error codes from scratch (the old behaviour)
18474 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18475 or delete all the .err files.
18479 * CAST-128 was incorrectly implemented for short keys. The C version has
18480 been fixed, but is untested. The assembler versions are also fixed, but
18481 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18482 to regenerate it if needed.
18483 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18484 Hagino <itojun@kame.net>*
18486 * File was opened incorrectly in randfile.c.
18488 *Ulf Möller <ulf@fitug.de>*
18490 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18491 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18492 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18493 al: it's just almost always a UTCTime. Note this patch adds new error
18494 codes so do a "make errors" if there are problems.
18498 * Correct Linux 1 recognition in config.
18500 *Ulf Möller <ulf@fitug.de>*
18502 * Remove pointless MD5 hash when using DSA keys in ca.
18504 *Anonymous <nobody@replay.com>*
18506 * Generate an error if given an empty string as a cert directory. Also
18507 generate an error if handed NULL (previously returned 0 to indicate an
18508 error, but didn't set one).
18510 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18512 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18516 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18517 parameters. This was causing a warning which killed off the Win32 compile.
18521 * Remove C++ style comments from crypto/bn/bn_local.h.
18523 *Neil Costigan <neil.costigan@celocom.com>*
18525 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18526 based on a text string, looking up short and long names and finally
18527 "dot" format. The "dot" format stuff didn't work. Added new function
18528 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18529 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18530 OID is not part of the table.
18534 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18535 X509_LOOKUP_by_alias().
18539 * Sort openssl functions by name.
18543 * Get the `gendsa` command working and add it to the `list` command. Remove
18544 encryption from sample DSA keys (in case anyone is interested the password
18549 * Make *all* `*_free` functions accept a NULL pointer.
18551 *Frans Heymans <fheymans@isaserver.be>*
18553 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18556 *Anonymous <nobody@replay.com>*
18558 * s_server should send the CAfile as acceptable CAs, not its own cert.
18560 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18562 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18564 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18566 * Temp key "for export" tests were wrong in s3_srvr.c.
18568 *Anonymous <nobody@replay.com>*
18570 * Add prototype for temp key callback functions
18571 SSL_CTX_set_tmp_{rsa,dh}_callback().
18575 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18576 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18580 * X509_name_add_entry() freed the wrong thing after an error.
18582 *Arne Ansper <arne@ats.cyber.ee>*
18584 * rsa_eay.c would attempt to free a NULL context.
18586 *Arne Ansper <arne@ats.cyber.ee>*
18588 * BIO_s_socket() had a broken should_retry() on Windoze.
18590 *Arne Ansper <arne@ats.cyber.ee>*
18592 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18594 *Arne Ansper <arne@ats.cyber.ee>*
18596 * Make sure the already existing X509_STORE->depth variable is initialized
18597 in X509_STORE_new(), but document the fact that this variable is still
18598 unused in the certificate verification process.
18600 *Ralf S. Engelschall*
18602 * Fix the various library and `apps/` files to free up pkeys obtained from
18603 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18607 * Fix reference counting in X509_PUBKEY_get(). This makes
18608 demos/maurice/example2.c work, amongst others, probably.
18610 *Steve Henson and Ben Laurie*
18612 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18613 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18614 are no longer created. This way we have a single and consistent command
18615 line interface `openssl <command>`, similar to `cvs <command>`.
18617 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18619 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18620 BIT STRING wrapper always have zero unused bits.
18624 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18628 * Make the top-level INSTALL documentation easier to understand.
18632 * Makefiles updated to exit if an error occurs in a sub-directory
18633 make (including if user presses ^C) [Paul Sutton]
18635 * Make Montgomery context stuff explicit in RSA data structure.
18639 * Fix build order of pem and err to allow for generated pem.h.
18643 * Fix renumbering bug in X509_NAME_delete_entry().
18647 * Enhanced the err-ins.pl script so it makes the error library number
18648 global and can add a library name. This is needed for external ASN1 and
18649 other error libraries.
18653 * Fixed sk_insert which never worked properly.
18657 * Fix ASN1 macros so they can handle indefinite length constructed
18658 EXPLICIT tags. Some non standard certificates use these: they can now
18663 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18664 into a single doc/ssleay.txt bundle. This way the information is still
18665 preserved but no longer messes up this directory. Now it's new room for
18666 the new set of documentation files.
18668 *Ralf S. Engelschall*
18670 * SETs were incorrectly DER encoded. This was a major pain, because they
18671 shared code with SEQUENCEs, which aren't coded the same. This means that
18672 almost everything to do with SETs or SEQUENCEs has either changed name or
18673 number of arguments.
18675 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18677 * Fix test data to work with the above.
18681 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18682 was already fixed by Eric for 0.9.1 it seems.
18684 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18686 * Autodetect FreeBSD3.
18690 * Fix various bugs in Configure. This affects the following platforms:
18694 unixware-2.0-pentium
18699 * Eliminate generated files from CVS. Reorder tests to regenerate files
18700 before they are needed.
18704 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18708 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18710 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18711 changed SSLeay to OpenSSL in version strings.
18713 *Ralf S. Engelschall*
18715 * Some fixups to the top-level documents.
18719 * Fixed the nasty bug where rsaref.h was not found under compile-time
18720 because the symlink to include/ was missing.
18722 *Ralf S. Engelschall*
18724 * Incorporated the popular no-RSA/DSA-only patches
18725 which allow to compile a RSA-free SSLeay.
18727 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18729 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18730 when "ssleay" is still not found.
18732 *Ralf S. Engelschall*
18734 * Added more platforms to Configure: Cray T3E, HPUX 11,
18736 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18738 * Updated the README file.
18740 *Ralf S. Engelschall*
18742 * Added various .cvsignore files in the CVS repository subdirs
18743 to make a "cvs update" really silent.
18745 *Ralf S. Engelschall*
18747 * Recompiled the error-definition header files and added
18748 missing symbols to the Win32 linker tables.
18750 *Ralf S. Engelschall*
18752 * Cleaned up the top-level documents;
18753 o new files: CHANGES and LICENSE
18754 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18755 o merged COPYRIGHT into LICENSE
18756 o removed obsolete TODO file
18757 o renamed MICROSOFT to INSTALL.W32
18759 *Ralf S. Engelschall*
18761 * Removed dummy files from the 0.9.1b source tree:
18763 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18764 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18765 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18766 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18767 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18770 *Ralf S. Engelschall*
18772 * Added various platform portability fixes.
18776 * The Genesis of the OpenSSL rpject:
18777 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18778 Young and Tim J. Hudson created while they were working for C2Net until
18781 *The OpenSSL Project*
18783 ### Changes between 0.9.0b and 0.9.1b [not released]
18785 * Updated a few CA certificates under certs/
18789 * Changed some BIGNUM api stuff.
18793 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18794 DGUX x86, Linux Alpha, etc.
18798 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18799 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18804 * Add -strparse option to asn1pars program which parses nested
18807 *Dr Stephen Henson <shenson@bigfoot.com>*
18809 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18813 * DSA fix for "ca" program.
18817 * Added "-genkey" option to "dsaparam" program.
18821 * Added RIPE MD160 (rmd160) message digest.
18825 * Added -a (all) option to "ssleay version" command.
18829 * Added PLATFORM define which is the id given to Configure.
18833 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18837 * Extended the ASN.1 parser routines.
18841 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18845 * Added a BN_CTX to the BN library.
18849 * Fixed the weak key values in DES library
18853 * Changed API in EVP library for cipher aliases.
18857 * Added support for RC2/64bit cipher.
18861 * Converted the lhash library to the crypto/mem.c functions.
18865 * Added more recognized ASN.1 object ids.
18869 * Added more RSA padding checks for SSL/TLS.
18873 * Added BIO proxy/filter functionality.
18877 * Added extra_certs to SSL_CTX which can be used
18878 send extra CA certificates to the client in the CA cert chain sending
18879 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18883 * Now Fortezza is denied in the authentication phase because
18884 this is key exchange mechanism is not supported by SSLeay at all.
18888 * Additional PKCS1 checks.
18892 * Support the string "TLSv1" for all TLS v1 ciphers.
18896 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18897 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18901 * Fixed a few memory leaks.
18905 * Fixed various code and comment typos.
18909 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18910 bytes sent in the client random.
18912 *Edward Bishop <ebishop@spyglass.com>*
18916 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18917 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18918 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18919 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18920 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18921 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18922 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18923 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18924 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18925 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18926 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18927 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18928 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18929 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18930 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18931 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18932 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18933 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18934 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18935 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18936 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18937 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18938 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18939 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18940 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18941 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18942 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18943 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18944 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18945 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18946 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18947 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18948 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18949 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18950 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18951 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18952 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18953 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18954 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18955 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18956 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18957 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18958 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18959 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18960 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18961 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18962 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18963 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18964 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18965 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18966 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18967 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18968 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18969 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18970 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18971 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18972 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18973 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18974 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18975 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18976 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18977 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18978 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18979 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18980 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18981 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18982 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18983 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18984 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18985 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18986 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18987 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18988 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18989 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18990 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18991 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18992 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18993 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18994 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18995 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18996 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18997 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18998 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18999 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19000 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19001 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19002 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19003 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19004 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19005 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19006 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19007 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19008 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19009 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19010 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19011 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19012 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19013 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19014 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19015 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19016 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19017 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19018 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19019 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19020 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19021 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19022 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19023 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19024 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19025 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19026 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19027 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19028 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19029 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19030 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19031 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19032 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19033 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19034 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19035 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19036 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19037 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19038 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19039 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19040 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19041 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19042 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19043 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19044 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19045 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19046 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19047 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19048 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19049 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19050 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19051 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19052 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19053 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19054 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19055 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19056 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19057 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19058 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19059 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19060 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19061 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19062 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19063 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19064 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19065 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19066 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19067 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19068 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19069 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19070 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19071 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19072 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19073 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19074 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19075 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19076 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19077 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655