4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
27 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
28 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
29 are to be used by servers requiring ephemeral DH keys. Instead applications
30 should consider using the built-in DH parameters that are available by
31 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
32 necessary then applications can use the alternative functions
33 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
34 replacement for the "callback" functions. The callback was originally useful
35 in order to have different parameters for export and non-export ciphersuites.
36 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
37 functions should be replaced by one of the other methods described above.
41 * The -crypt option to the passwd command line tool has been removed.
45 * The -C option to the x509, dhparam, dsaparam, and ecparam commands
50 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
52 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
53 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
54 The inverse ciphers use AES decryption for wrapping, and
55 AES encryption for unwrapping.
59 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
60 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
61 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
62 OpenSSL 3.0 these are replaced by the more generic functions
63 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
64 The old versions have been converted to deprecated macros that just call the
69 * The security callback, which can be customised by application code, supports
70 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
71 in the "other" parameter. In most places this is what is passed. All these
72 places occur server side. However there was one client side call of this
73 security operation and it passed a DH object instead. This is incorrect
74 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
75 of the other locations. Therefore this client side call has been changed to
76 pass an EVP_PKEY instead.
80 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
81 interface. Their functionality remains unchanged.
85 * Added new option for 'openssl list', '-providers', which will display the
86 list of loaded providers, their names, version and status. It optionally
87 displays their gettable parameters.
91 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
92 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
93 type is internally recognised so the workaround is no longer needed.
95 Functionality is still retained as it is, but will only work with
96 EVP_PKEYs with a legacy internal key.
100 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
101 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
105 * Changed all "STACK" functions to be macros instead of inline functions. Macro
106 parameters are still checked for type safety at compile time via helper
111 * Remove the RAND_DRBG API
113 The RAND_DRBG API did not fit well into the new provider concept as
114 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
115 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
116 and some of its API calls are rather low-level. This holds in particular
117 for the callback mechanism (RAND_DRBG_set_callbacks()).
119 Adding a compatibility layer to continue supporting the RAND_DRBG API as
120 a legacy API for a regular deprecation period turned out to come at the
121 price of complicating the new provider API unnecessarily. Since the
122 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
125 *Paul Dale and Matthias St. Pierre*
127 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
128 as well as actual hostnames.
132 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
133 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
134 conversely, silently ignore DTLS protocol version bounds when configuring
135 TLS-based contexts. The commands can be repeated to set bounds of both
136 types. The same applies with the corresponding "min_protocol" and
137 "max_protocol" command-line switches, in case some application uses both TLS
140 SSL_CTX instances that are created for a fixed protocol version (e.g.
141 TLSv1_server_method()) also silently ignore version bounds. Previously
142 attempts to apply bounds to these protocol versions would result in an
143 error. Now only the "version-flexible" SSL_CTX instances are subject to
144 limits in configuration files in command-line options.
148 * Deprecated the `ENGINE` API. Engines should be replaced with providers
153 * Reworked the recorded ERR codes to make better space for system errors.
154 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
155 given code is a system error (true) or an OpenSSL error (false).
159 * Reworked the test perl framework to better allow parallel testing.
161 *Nicola Tuveri and David von Oheimb*
163 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
164 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
168 * 'Configure' has been changed to figure out the configuration target if
169 none is given on the command line. Consequently, the 'config' script is
170 now only a mere wrapper. All documentation is changed to only mention
173 *Rich Salz and Richard Levitte*
175 * Added a library context `OSSL_LIB_CTX` that applications as well as
176 other libraries can use to form a separate context within which
177 libcrypto operations are performed.
179 There are two ways this can be used:
181 - Directly, by passing a library context to functions that take
182 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
184 - Indirectly, by creating a new library context and then assigning
185 it as the new default, with `OSSL_LIB_CTX_set0_default`.
187 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
188 apart from the functions directly related to `OSSL_LIB_CTX`, accept
189 NULL to indicate that the default library context should be used.
191 Library code that changes the default library context using
192 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
193 second call before returning to the caller.
195 _(Note: the library context was initially called `OPENSSL_CTX` and
196 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
200 * Handshake now fails if Extended Master Secret extension is dropped
205 * Dropped interactive mode from the 'openssl' program. From now on,
206 the `openssl` command without arguments is equivalent to `openssl
211 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
212 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
213 While the old function names have been retained for backward compatibility
214 they should not be used in new developments
215 because their return values are confusing: Unlike other `_cmp()` functions
216 they do not return 0 in case their arguments are equal.
220 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
221 EC_GROUP_get_field_type().
225 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
226 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
227 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
228 Applications should rely on the library automatically assigning a suitable
229 EC_METHOD internally upon EC_GROUP construction.
233 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
234 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
235 assigned internally without application intervention.
236 Users of EC_GROUP_new() should switch to a different suitable constructor.
240 * Add CAdES-BES signature verification support, mostly derived
241 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
243 *Filipe Raimundo da Silva*
245 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
249 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
250 functions are not widely used and now OpenSSL automatically perform this
251 conversion when needed.
255 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
256 EC_KEY_precompute_mult(). These functions are not widely used and
257 applications should instead switch to named curves which OpenSSL has
258 hardcoded lookup tables for.
262 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
263 should instead use the L<EC_POINT_mul(3)> function.
267 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
268 that are not applicable to the new provider model. Applications should
269 instead use EVP_default_properties_is_fips_enabled() and
270 EVP_default_properties_enable_fips().
274 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
275 is set, an unexpected EOF is ignored, it pretends a close notify was received
276 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
280 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
281 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
282 used and applications should instead use the
283 L<EC_POINT_set_affine_coordinates(3)> and
284 L<EC_POINT_get_affine_coordinates(3)> functions.
288 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
289 arrays to be more easily constructed via a series of utility functions.
290 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
291 the various push functions and finally convert to a passable OSSL_PARAM
292 array using OSSL_PARAM_BLD_to_param().
296 * The security strength of SHA1 and MD5 based signatures in TLS has been
297 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
298 working at the default security level of 1 and instead requires security
299 level 0. The security level can be changed either using the cipher string
300 with @SECLEVEL, or calling SSL_CTX_set_security_level().
304 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
305 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
306 internal keys, if they correspond to one of those built in types.
310 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
311 contain a provider side internal key.
315 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
316 They are old functions that we don't use, and that you could disable with
317 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
321 * Project text documents not yet having a proper file name extension
322 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
323 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
327 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
328 have been converted to Markdown with the goal to produce documents
329 which not only look pretty when viewed online in the browser, but
330 remain well readable inside a plain text editor.
332 To achieve this goal, a 'minimalistic' Markdown style has been applied
333 which avoids formatting elements that interfere too much with the
334 reading flow in the text file. For example, it
336 * avoids [ATX headings][] and uses [setext headings][] instead
337 (which works for `<h1>` and `<h2>` headings only).
338 * avoids [inline links][] and uses [reference links][] instead.
339 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
341 [ATX headings]: https://github.github.com/gfm/#atx-headings
342 [setext headings]: https://github.github.com/gfm/#setext-headings
343 [inline links]: https://github.github.com/gfm/#inline-link
344 [reference links]: https://github.github.com/gfm/#reference-link
345 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
346 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
348 *Matthias St. Pierre*
350 * The test suite is changed to preserve results of each test recipe.
351 A new directory test-runs/ with subdirectories named like the
352 test recipes are created in the build tree for this purpose.
356 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
357 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
358 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
360 *David von Oheimb, Martin Peylo*
362 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
363 The legacy OCSP-focused and only partly documented API is retained.
364 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
368 * Added `util/check-format.pl`, a tool for checking adherence to the
369 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
370 The checks performed are incomplete and yield some false positives.
371 Still the tool should be useful for detecting most typical glitches.
375 * BIO_do_connect and BIO_do_handshake have been extended:
376 If domain name resolution yields multiple IP addresses all of them are tried
377 after connect() failures.
381 * All of the low level RSA functions have been deprecated including:
383 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
384 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
385 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
386 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
387 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
388 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
389 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
390 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
391 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
392 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
393 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
394 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
395 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
396 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
397 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
398 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
399 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
400 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
401 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
402 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
403 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
404 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
405 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
406 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
407 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
408 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
409 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
410 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
412 Use of these low level functions has been informally discouraged for a long
413 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
414 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
415 L<EVP_PKEY_decrypt(3)>.
419 * X509 certificates signed using SHA1 are no longer allowed at security
421 In TLS/SSL the default security level is 1. It can be set either
422 using the cipher string with @SECLEVEL, or calling
423 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
424 a call to SSL_CTX_use_certificate() will fail if the security level is not
426 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
427 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
432 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
433 modified to use PKEY APIs. These commands are now in maintenance mode
434 and no new features will be added to them.
438 * The command line utility rsautl has been deprecated.
439 Instead use the pkeyutl program.
443 * The command line utilities genrsa and rsa have been modified to use PKEY
444 APIs. They now write PKCS#8 keys by default. These commands are now in
445 maintenance mode and no new features will be added to them.
449 * All of the low level DH functions have been deprecated including:
451 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
452 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
453 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
454 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
455 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
456 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
457 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
458 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
459 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
460 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
461 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
462 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
463 DH_meth_get_generate_params and DH_meth_set_generate_params.
465 Use of these low level functions has been informally discouraged for a long
466 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
467 and L<EVP_PKEY_derive(3)>.
471 * All of the low level DSA functions have been deprecated including:
473 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
474 DSA_get_default_method, DSA_set_method, DSA_get_method,
475 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
476 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
477 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
478 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
479 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
480 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
481 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
482 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
483 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
484 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
485 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
488 Use of these low level functions has been informally discouraged for a long
489 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
490 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
494 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
495 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
496 This means that applications don't have to look at the curve NID and
497 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
498 However, they still can, that EVP_PKEY_set_alias_type() call acts as
499 a no-op when the EVP_PKEY is already of the given type.
501 Parameter and key generation is also reworked to make it possible
502 to generate EVP_PKEY_SM2 parameters and keys without having to go
503 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
504 However, code that does the latter will still work as before.
508 * Deprecated low level ECDH and ECDSA functions. These include:
510 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
511 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
514 Use of these low level functions has been informally discouraged for a long
515 time. Instead applications should use the EVP_PKEY_derive(3),
516 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
520 * Deprecated the EC_KEY_METHOD functions. These include:
522 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
523 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
524 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
525 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
526 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
527 EC_KEY_METHOD_get_verify.
529 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
533 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
534 and EVP_PKEY_decrypt() instead.
535 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
536 and EVP_PKEY_encrypt() instead.
540 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
541 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
542 a new formulation to include all the things it can be used for,
543 as well as words of caution.
547 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
548 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
552 * All of the low level HMAC functions have been deprecated including:
554 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
555 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
558 Use of these low level functions has been informally discouraged for a long
559 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
560 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
561 and L<EVP_MAC_final(3)>.
565 * Over two thousand fixes were made to the documentation, including:
566 - Common options (such as -rand/-writerand, TLS version control, etc)
567 were refactored and point to newly-enhanced descriptions in openssl.pod.
568 - Added style conformance for all options (with help from Richard Levitte),
569 documented all reported missing options, added a CI build to check
570 that all options are documented and that no unimplemented options
572 - Documented some internals, such as all use of environment variables.
573 - Addressed all internal broken L<> references.
577 * All of the low level CMAC functions have been deprecated including:
579 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
580 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
582 Use of these low level functions has been informally discouraged for a long
583 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
584 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
585 and L<EVP_MAC_final(3)>.
589 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
590 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
593 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
594 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
595 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
596 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
597 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
598 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
599 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
600 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
601 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
602 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
604 Use of these low level functions has been informally discouraged
605 for a long time. Applications should use the EVP_DigestInit_ex(3),
606 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
610 * Corrected the documentation of the return values from the `EVP_DigestSign*`
611 set of functions. The documentation mentioned negative values for some
612 errors, but this was never the case, so the mention of negative values
615 Code that followed the documentation and thereby check with something
616 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
620 * All of the low level cipher functions have been deprecated including:
622 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
623 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
624 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
625 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
626 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
627 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
628 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
629 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
630 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
631 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
632 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
633 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
634 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
635 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
636 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
637 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
638 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
639 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
640 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
641 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
642 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
643 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
644 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
645 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
646 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
647 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
648 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
649 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
650 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
652 Use of these low level functions has been informally discouraged for
653 a long time. Applications should use the high level EVP APIs, e.g.
654 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
655 equivalently named decrypt functions instead.
657 *Matt Caswell and Paul Dale*
659 * Removed include/openssl/opensslconf.h.in and replaced it with
660 include/openssl/configuration.h.in, which differs in not including
661 <openssl/macros.h>. A short header include/openssl/opensslconf.h
662 was added to include both.
664 This allows internal hacks where one might need to modify the set
665 of configured macros, for example this if deprecated symbols are
666 still supposed to be available internally:
668 #include <openssl/configuration.h>
670 #undef OPENSSL_NO_DEPRECATED
671 #define OPENSSL_SUPPRESS_DEPRECATED
673 #include <openssl/macros.h>
675 This should not be used by applications that use the exported
676 symbols, as that will lead to linking errors.
680 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
681 used in exponentiation with 512-bit moduli. No EC algorithms are
682 affected. Analysis suggests that attacks against 2-prime RSA1024,
683 3-prime RSA1536, and DSA1024 as a result of this defect would be very
684 difficult to perform and are not believed likely. Attacks against DH512
685 are considered just feasible. However, for an attack the target would
686 have to re-use the DH512 private key, which is not recommended anyway.
687 Also applications directly using the low level API BN_mod_exp may be
688 affected if they use BN_FLG_CONSTTIME.
693 * Most memory-debug features have been deprecated, and the functionality
694 replaced with no-ops.
698 * Added documentation for the STACK API.
702 * Introduced a new method type and API, OSSL_ENCODER, to
703 represent generic encoders. An implementation is expected to
704 be able to encode an object associated with a given name (such
705 as an algorithm name for an asymmetric key) into forms given by
706 implementation properties.
708 Encoders are primarily used from inside libcrypto, through
709 calls to functions like EVP_PKEY_print_private(),
710 PEM_write_bio_PrivateKey() and similar.
712 Encoders are specified in such a way that they can be made to
713 directly handle the provider side portion of an object, if this
714 provider side part comes from the same provider as the encoder
715 itself, but can also be made to handle objects in parametrized
716 form (as an OSSL_PARAM array of data). This allows a provider to
717 offer generic encoders as a service for any other provider.
721 * Added a .pragma directive to the syntax of configuration files, to
722 allow varying behavior in a supported and predictable manner.
723 Currently added pragma:
727 This allows dollar signs to be a keyword character unless it's
728 followed by a opening brace or parenthesis. This is useful for
729 platforms where dollar signs are commonly used in names, such as
730 volume names and system directory names on VMS.
734 * Added functionality to create an EVP_PKEY from user data. This
735 is effectively the same as creating a RSA, DH or DSA object and
736 then assigning them to an EVP_PKEY, but directly using algorithm
737 agnostic EVP functions. A benefit is that this should be future
738 proof for public key algorithms to come.
742 * Change the interpretation of the '--api' configuration option to
743 mean that this is a desired API compatibility level with no
744 further meaning. The previous interpretation, that this would
745 also mean to remove all deprecated symbols up to and including
746 the given version, no requires that 'no-deprecated' is also used
747 in the configuration.
749 When building applications, the desired API compatibility level
750 can be set with the OPENSSL_API_COMPAT macro like before. For
751 API compatibility version below 3.0, the old style numerical
752 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
753 For version 3.0 and on, the value is expected to be the decimal
754 value calculated from the major and minor version like this:
756 MAJOR * 10000 + MINOR * 100
760 -DOPENSSL_API_COMPAT=30000 For 3.0
761 -DOPENSSL_API_COMPAT=30200 For 3.2
763 To hide declarations that are deprecated up to and including the
764 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
765 given when building the application as well.
769 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
770 access to certificate and CRL stores via URIs and OSSL_STORE
773 This adds the following functions:
775 - X509_LOOKUP_store()
776 - X509_STORE_load_file()
777 - X509_STORE_load_path()
778 - X509_STORE_load_store()
779 - SSL_add_store_cert_subjects_to_stack()
780 - SSL_CTX_set_default_verify_store()
781 - SSL_CTX_load_verify_file()
782 - SSL_CTX_load_verify_dir()
783 - SSL_CTX_load_verify_store()
787 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
788 The presence of this system service is determined at run-time.
792 * Added functionality to create an EVP_PKEY context based on data
793 for methods from providers. This takes an algorithm name and a
794 property query string and simply stores them, with the intent
795 that any operation that uses this context will use those strings
796 to fetch the needed methods implicitly, thereby making the port
797 of application written for pre-3.0 OpenSSL easier.
801 * The undocumented function NCONF_WIN32() has been deprecated; for
802 conversion details see the HISTORY section of doc/man5/config.pod
806 * Introduced the new functions EVP_DigestSignInit_ex() and
807 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
808 EVP_DigestVerifyUpdate() have been converted to functions. See the man
809 pages for further details.
813 * Over two thousand fixes were made to the documentation, including:
814 adding missing command flags, better style conformance, documentation
817 *Rich Salz, Richard Levitte*
819 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
820 X25519, X448, Ed25519 and Ed448.
824 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
829 * Deprecated the public definition of ERR_STATE as well as the function
830 ERR_get_state(). This is done in preparation of making ERR_STATE an
835 * Added ERR functionality to give callers access to the stored function
836 names that have replaced the older function code based functions.
838 New functions are ERR_get_error_func(), ERR_peek_error_func(),
839 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
840 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
841 and ERR_peek_last_error_all().
843 These functions have become deprecated: ERR_get_error_line_data(),
844 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
845 ERR_func_error_string().
849 * Extended testing to be verbose for failing tests only. The make variables
850 VERBOSE_FAILURE or VF can be used to enable this:
852 $ make VF=1 test # Unix
853 $ mms /macro=(VF=1) test ! OpenVMS
854 $ nmake VF=1 test # Windows
858 * Added several checks to X509_verify_cert() according to requirements in
859 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
860 (which may be done by using the CLI option `-x509_strict`):
861 * The basicConstraints of CA certificates must be marked critical.
862 * CA certificates must explicitly include the keyUsage extension.
863 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
864 * The issuer name of any certificate must not be empty.
865 * The subject name of CA certs, certs with keyUsage crlSign,
866 and certs without subjectAlternativeName must not be empty.
867 * If a subjectAlternativeName extension is given it must not be empty.
868 * The signatureAlgorithm field and the cert signature must be consistent.
869 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
870 must not be marked critical.
871 * The authorityKeyIdentifier must be given for X.509v3 certs
872 unless they are self-signed.
873 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
877 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
878 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
882 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
883 used even when parsing explicit parameters, when loading a encoded key
884 or calling `EC_GROUP_new_from_ecpkparameters()`/
885 `EC_GROUP_new_from_ecparameters()`.
886 This prevents bypass of security hardening and performance gains,
887 especially for curves with specialized EC_METHODs.
888 By default, if a key encoded with explicit parameters is loaded and later
889 encoded, the output is still encoded with explicit parameters, even if
890 internally a "named" EC_GROUP is used for computation.
894 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
895 this change, EC_GROUP_set_generator would accept order and/or cofactor as
896 NULL. After this change, only the cofactor parameter can be NULL. It also
897 does some minimal sanity checks on the passed order.
902 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
903 An attack is simple, if the first CMS_recipientInfo is valid but the
904 second CMS_recipientInfo is chosen ciphertext. If the second
905 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
906 encryption key will be replaced by garbage, and the message cannot be
907 decoded, but if the RSA decryption fails, the correct encryption key is
908 used and the recipient will not notice the attack.
909 As a work around for this potential attack the length of the decrypted
910 key must be equal to the cipher default key length, in case the
911 certifiate is not given and all recipientInfo are tried out.
912 The old behaviour can be re-enabled in the CMS code by setting the
913 CMS_DEBUG_DECRYPT flag.
917 * Early start up entropy quality from the DEVRANDOM seed source has been
918 improved for older Linux systems. The RAND subsystem will wait for
919 /dev/random to be producing output before seeding from /dev/urandom.
920 The seeded state is stored for future library initialisations using
921 a system global shared memory segment. The shared memory identifier
922 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
923 the desired value. The default identifier is 114.
927 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
928 when primes for RSA keys are computed.
929 Since we previously always generated primes == 2 (mod 3) for RSA keys,
930 the 2-prime and 3-prime RSA modules were easy to distinguish, since
931 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
932 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
933 This avoids possible fingerprinting of newly generated RSA modules.
937 * Correct the extended master secret constant on EBCDIC systems. Without this
938 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
939 negotiate EMS will fail. Unfortunately this also means that TLS connections
940 between EBCDIC systems with this fix, and EBCDIC systems without this
941 fix will fail if they negotiate EMS.
945 * Changed the library initialisation so that the config file is now loaded
946 by default. This was already the case for libssl. It now occurs for both
947 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
948 OPENSSL_init_crypto() to suppress automatic loading of a config file.
952 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
953 where the former acts as a replacement for ERR_put_error(), and the
954 latter replaces the combination ERR_put_error()+ERR_add_error_data().
955 ERR_raise_data() adds more flexibility by taking a format string and
956 an arbitrary number of arguments following it, to be processed with
961 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
962 to check if a named provider is loaded and available. When called, it
963 will also activate all fallback providers if such are still present.
967 * Enforce a minimum DH modulus size of 512 bits.
971 * Changed DH parameters to generate the order q subgroup instead of 2q.
972 Previously generated DH parameters are still accepted by DH_check
973 but DH_generate_key works around that by clearing bit 0 of the
974 private key for those. This avoids leaking bit 0 of the private key.
978 * Significantly reduce secure memory usage by the randomness pools.
982 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
987 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
988 algorithms. An implementation of a key exchange algorithm can be obtained
989 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
990 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
991 the older EVP_PKEY_derive_init() function. See the man pages for the new
992 functions for further details.
996 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1000 * Removed the function names from error messages and deprecated the
1003 * Removed NextStep support and the macro OPENSSL_UNISTD
1007 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1008 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1009 Also removed "export var as function" capability; we do not export
1010 variables, only functions.
1014 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1015 an error and 1 indicating success. In previous versions of OpenSSL this
1016 was a void type. If a key was set longer than the maximum possible this
1021 * Support SM2 signing and verification schemes with X509 certificate.
1025 * Use SHA256 as the default digest for TS query in the ts app.
1029 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1030 This checks that the salt length is at least 128 bits, the derived key
1031 length is at least 112 bits, and that the iteration count is at least 1000.
1032 For backwards compatibility these checks are disabled by default in the
1033 default provider, but are enabled by default in the fips provider.
1034 To enable or disable these checks use the control
1035 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1039 * Default cipher lists/suites are now available via a function, the
1040 #defines are deprecated.
1044 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1045 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1046 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1050 * Join the directories crypto/x509 and crypto/x509v3
1054 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1055 This changes the size when using the genpkey app when no size is given. It
1056 fixes an omission in earlier changes that changed all RSA, DSA and DH
1057 generation apps to use 2048 bits by default.
1061 * Added command 'openssl kdf' that uses the EVP_KDF API.
1065 * Added command 'openssl mac' that uses the EVP_MAC API.
1069 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1070 as default directories. Also added the command 'openssl info'
1071 for scripting purposes.
1075 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1076 deprecated. These undocumented functions were never integrated into the EVP
1077 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1078 Bi-directional IGE mode. These modes were never formally standardised and
1079 usage of these functions is believed to be very small. In particular
1080 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1081 is ever used. The security implications are believed to be minimal, but
1082 this issue was never fixed for backwards compatibility reasons. New code
1083 should not use these modes.
1087 * Add prediction resistance to the DRBG reseeding process.
1091 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1092 mandated by IEEE Std 1619-2018.
1096 * Added newline escaping functionality to a filename when using openssl dgst.
1097 This output format is to replicate the output format found in the `*sum`
1098 checksum programs. This aims to preserve backward compatibility.
1100 *Matt Eaton, Richard Levitte, and Paul Dale*
1102 * Removed the heartbeat message in DTLS feature, as it has very
1103 little usage and doesn't seem to fulfill a valuable purpose.
1104 The configuration option is now deprecated.
1108 * Changed the output of 'openssl {digestname} < file' to display the
1109 digest name in its output.
1113 * Added a new generic trace API which provides support for enabling
1114 instrumentation through trace output. This feature is mainly intended
1115 as an aid for developers and is disabled by default. To utilize it,
1116 OpenSSL needs to be configured with the `enable-trace` option.
1118 If the tracing API is enabled, the application can activate trace output
1119 by registering BIOs as trace channels for a number of tracing and debugging
1122 The 'openssl' application has been expanded to enable any of the types
1123 available via environment variables defined by the user, and serves as
1124 one possible example on how to use this functionality.
1126 *Richard Levitte & Matthias St. Pierre*
1128 * Added build tests for C++. These are generated files that only do one
1129 thing, to include one public OpenSSL head file each. This tests that
1130 the public header files can be usefully included in a C++ application.
1132 This test isn't enabled by default. It can be enabled with the option
1133 'enable-buildtest-c++'.
1137 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1141 * Add KMAC to EVP_MAC.
1145 * Added property based algorithm implementation selection framework to
1150 * Added SCA hardening for modular field inversion in EC_GROUP through
1151 a new dedicated field_inv() pointer in EC_METHOD.
1152 This also addresses a leakage affecting conversions from projective
1153 to affine coordinates.
1155 *Billy Bob Brumley, Nicola Tuveri*
1157 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1158 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1159 those algorithms that were already supported through the EVP_PKEY API
1160 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1161 and scrypt are now wrappers that call EVP_KDF.
1165 * Build devcrypto engine as a dynamic engine.
1167 *Eneas U de Queiroz*
1169 * Add keyed BLAKE2 to EVP_MAC.
1173 * Fix a bug in the computation of the endpoint-pair shared secret used
1174 by DTLS over SCTP. This breaks interoperability with older versions
1175 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1176 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1177 interoperability with such broken implementations. However, enabling
1178 this switch breaks interoperability with correct implementations.
1180 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1181 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1185 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1189 * Change the license to the Apache License v2.0.
1193 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1195 - Major releases (indicated by incrementing the MAJOR release number)
1196 may introduce incompatible API/ABI changes.
1197 - Minor releases (indicated by incrementing the MINOR release number)
1198 may introduce new features but retain API/ABI compatibility.
1199 - Patch releases (indicated by incrementing the PATCH number)
1200 are intended for bug fixes and other improvements of existing
1201 features only (like improving performance or adding documentation)
1202 and retain API/ABI compatibility.
1206 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1210 * Remove the 'dist' target and add a tarball building script. The
1211 'dist' target has fallen out of use, and it shouldn't be
1212 necessary to configure just to create a source distribution.
1216 * Recreate the OS390-Unix config target. It no longer relies on a
1217 special script like it did for OpenSSL pre-1.1.0.
1221 * Instead of having the source directories listed in Configure, add
1222 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1227 * Add GMAC to EVP_MAC.
1231 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1235 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1236 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1237 to facilitate the continued use of MACs through raw private keys in
1238 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1242 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1243 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1247 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1248 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1249 are retained for backwards compatibility.
1253 * AES-XTS mode now enforces that its two keys are different to mitigate
1254 the attacked described in "Efficient Instantiations of Tweakable
1255 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1256 Details of this attack can be obtained from:
1257 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1261 * Rename the object files, i.e. give them other names than in previous
1262 versions. Their names now include the name of the final product, as
1263 well as its type mnemonic (bin, lib, shlib).
1267 * Added new option for 'openssl list', '-objects', which will display the
1268 list of built in objects, i.e. OIDs with names.
1272 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1273 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1278 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1279 improves application performance by removing data copies and providing
1280 applications with zero-copy system calls such as sendfile and splice.
1284 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1285 option is set, openssl cleanses (zeroize) plaintext bytes from
1286 internal buffers after delivering them to the application. Note,
1287 the application is still responsible for cleansing other copies
1288 (e.g.: data received by SSL_read(3)).
1292 * `PKCS12_parse` now maintains the order of the parsed certificates
1293 when outputting them via `*ca` (rather than reversing it).
1297 * Deprecated pthread fork support methods. These were unused so no
1298 replacement is required.
1300 - OPENSSL_fork_prepare()
1301 - OPENSSL_fork_parent()
1302 - OPENSSL_fork_child()
1309 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1313 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1315 * Certificates with explicit curve parameters are now disallowed in
1316 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1320 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1321 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1322 conversely, silently ignore DTLS protocol version bounds when configuring
1323 TLS-based contexts. The commands can be repeated to set bounds of both
1324 types. The same applies with the corresponding "min_protocol" and
1325 "max_protocol" command-line switches, in case some application uses both TLS
1328 SSL_CTX instances that are created for a fixed protocol version (e.g.
1329 TLSv1_server_method()) also silently ignore version bounds. Previously
1330 attempts to apply bounds to these protocol versions would result in an
1331 error. Now only the "version-flexible" SSL_CTX instances are subject to
1332 limits in configuration files in command-line options.
1336 * Handshake now fails if Extended Master Secret extension is dropped
1341 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1343 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1345 * Fixed segmentation fault in SSL_check_chain()
1346 Server or client applications that call the SSL_check_chain() function
1347 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1348 dereference as a result of incorrect handling of the
1349 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1350 or unrecognised signature algorithm is received from the peer. This could
1351 be exploited by a malicious peer in a Denial of Service attack.
1356 * Added AES consttime code for no-asm configurations
1357 an optional constant time support for AES was added
1358 when building openssl for no-asm.
1359 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1360 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1361 At this time this feature is by default disabled.
1362 It will be enabled by default in 3.0.
1366 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1368 * Revert the change of EOF detection while reading in libssl to avoid
1369 regressions in applications depending on the current way of reporting
1370 the EOF. As the existing method is not fully accurate the change to
1371 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1372 branch and will be present in the 3.0 release.
1376 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1377 when primes for RSA keys are computed.
1378 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1379 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1380 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1381 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1382 This avoids possible fingerprinting of newly generated RSA modules.
1386 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1388 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1389 while reading in libssl then we would report an error back to the
1390 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1391 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1392 therefore give a hint as to what went wrong.
1396 * Check that ed25519 and ed448 are allowed by the security level. Previously
1397 signature algorithms not using an MD were not being checked that they were
1398 allowed by the security level.
1402 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1403 was not quite right. The behaviour was not consistent between resumption
1404 and normal handshakes, and also not quite consistent with historical
1405 behaviour. The behaviour in various scenarios has been clarified and
1406 it has been updated to make it match historical behaviour as closely as
1411 * *[VMS only]* The header files that the VMS compilers include automatically,
1412 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1413 that the C++ compiler doesn't understand. This is a shortcoming in the
1414 compiler, but can be worked around with `__cplusplus` guards.
1416 C++ applications that use OpenSSL libraries must be compiled using the
1417 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1418 functions. Otherwise, only functions with symbols of less than 31
1419 characters can be used, as the linker will not be able to successfully
1420 resolve symbols with longer names.
1424 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1425 The presence of this system service is determined at run-time.
1429 * Added newline escaping functionality to a filename when using openssl dgst.
1430 This output format is to replicate the output format found in the `*sum`
1431 checksum programs. This aims to preserve backward compatibility.
1433 *Matt Eaton, Richard Levitte, and Paul Dale*
1435 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1440 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1442 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1443 number generator (RNG). This was intended to include protection in the
1444 event of a fork() system call in order to ensure that the parent and child
1445 processes did not share the same RNG state. However this protection was not
1446 being used in the default case.
1448 A partial mitigation for this issue is that the output from a high
1449 precision timer is mixed into the RNG state so the likelihood of a parent
1450 and child process sharing state is significantly reduced.
1452 If an application already calls OPENSSL_init_crypto() explicitly using
1453 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1456 *Matthias St. Pierre*
1458 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1459 used even when parsing explicit parameters, when loading a encoded key
1460 or calling `EC_GROUP_new_from_ecpkparameters()`/
1461 `EC_GROUP_new_from_ecparameters()`.
1462 This prevents bypass of security hardening and performance gains,
1463 especially for curves with specialized EC_METHODs.
1464 By default, if a key encoded with explicit parameters is loaded and later
1465 encoded, the output is still encoded with explicit parameters, even if
1466 internally a "named" EC_GROUP is used for computation.
1470 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1471 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1472 NULL. After this change, only the cofactor parameter can be NULL. It also
1473 does some minimal sanity checks on the passed order.
1478 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1479 An attack is simple, if the first CMS_recipientInfo is valid but the
1480 second CMS_recipientInfo is chosen ciphertext. If the second
1481 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1482 encryption key will be replaced by garbage, and the message cannot be
1483 decoded, but if the RSA decryption fails, the correct encryption key is
1484 used and the recipient will not notice the attack.
1485 As a work around for this potential attack the length of the decrypted
1486 key must be equal to the cipher default key length, in case the
1487 certifiate is not given and all recipientInfo are tried out.
1488 The old behaviour can be re-enabled in the CMS code by setting the
1489 CMS_DEBUG_DECRYPT flag.
1494 * Early start up entropy quality from the DEVRANDOM seed source has been
1495 improved for older Linux systems. The RAND subsystem will wait for
1496 /dev/random to be producing output before seeding from /dev/urandom.
1497 The seeded state is stored for future library initialisations using
1498 a system global shared memory segment. The shared memory identifier
1499 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1500 the desired value. The default identifier is 114.
1504 * Correct the extended master secret constant on EBCDIC systems. Without this
1505 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1506 negotiate EMS will fail. Unfortunately this also means that TLS connections
1507 between EBCDIC systems with this fix, and EBCDIC systems without this
1508 fix will fail if they negotiate EMS.
1512 * Use Windows installation paths in the mingw builds
1514 Mingw isn't a POSIX environment per se, which means that Windows
1515 paths should be used for installation.
1520 * Changed DH_check to accept parameters with order q and 2q subgroups.
1521 With order 2q subgroups the bit 0 of the private key is not secret
1522 but DH_generate_key works around that by clearing bit 0 of the
1523 private key for those. This avoids leaking bit 0 of the private key.
1527 * Significantly reduce secure memory usage by the randomness pools.
1531 * Revert the DEVRANDOM_WAIT feature for Linux systems
1533 The DEVRANDOM_WAIT feature added a select() call to wait for the
1534 /dev/random device to become readable before reading from the
1535 /dev/urandom device.
1537 It turned out that this change had negative side effects on
1538 performance which were not acceptable. After some discussion it
1539 was decided to revert this feature and leave it up to the OS
1540 resp. the platform maintainer to ensure a proper initialization
1541 during early boot time.
1543 *Matthias St. Pierre*
1545 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1547 * Add build tests for C++. These are generated files that only do one
1548 thing, to include one public OpenSSL head file each. This tests that
1549 the public header files can be usefully included in a C++ application.
1551 This test isn't enabled by default. It can be enabled with the option
1552 'enable-buildtest-c++'.
1556 * Enable SHA3 pre-hashing for ECDSA and DSA.
1560 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1561 This changes the size when using the genpkey app when no size is given. It
1562 fixes an omission in earlier changes that changed all RSA, DSA and DH
1563 generation apps to use 2048 bits by default.
1567 * Reorganize the manual pages to consistently have RETURN VALUES,
1568 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1569 util/fix-doc-nits accordingly.
1571 *Paul Yang, Joshua Lock*
1573 * Add the missing accessor EVP_PKEY_get0_engine()
1577 * Have apps like 's_client' and 's_server' output the signature scheme
1578 along with other cipher suite parameters when debugging.
1582 * Make OPENSSL_config() error agnostic again.
1586 * Do the error handling in RSA decryption constant time.
1590 * Prevent over long nonces in ChaCha20-Poly1305.
1592 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1593 for every encryption operation. RFC 7539 specifies that the nonce value
1594 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1595 and front pads the nonce with 0 bytes if it is less than 12
1596 bytes. However it also incorrectly allows a nonce to be set of up to 16
1597 bytes. In this case only the last 12 bytes are significant and any
1598 additional leading bytes are ignored.
1600 It is a requirement of using this cipher that nonce values are
1601 unique. Messages encrypted using a reused nonce value are susceptible to
1602 serious confidentiality and integrity attacks. If an application changes
1603 the default nonce length to be longer than 12 bytes and then makes a
1604 change to the leading bytes of the nonce expecting the new value to be a
1605 new unique nonce then such an application could inadvertently encrypt
1606 messages with a reused nonce.
1608 Additionally the ignored bytes in a long nonce are not covered by the
1609 integrity guarantee of this cipher. Any application that relies on the
1610 integrity of these ignored leading bytes of a long nonce may be further
1611 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1612 is safe because no such use sets such a long nonce value. However user
1613 applications that use this cipher directly and set a non-default nonce
1614 length to be longer than 12 bytes may be vulnerable.
1616 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1622 * Add DEVRANDOM_WAIT feature for Linux systems
1624 On older Linux systems where the getrandom() system call is not available,
1625 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1626 Contrary to getrandom(), the /dev/urandom device will not block during
1627 early boot when the kernel CSPRNG has not been seeded yet.
1629 To mitigate this known weakness, use select() to wait for /dev/random to
1630 become readable before reading from /dev/urandom.
1632 * Ensure that SM2 only uses SM3 as digest algorithm
1636 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1638 * Change the info callback signals for the start and end of a post-handshake
1639 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1640 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1641 confused by this and assume that a TLSv1.2 renegotiation has started. This
1642 can break KeyUpdate handling. Instead we no longer signal the start and end
1643 of a post handshake message exchange (although the messages themselves are
1644 still signalled). This could break some applications that were expecting
1645 the old signals. However without this KeyUpdate is not usable for many
1650 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1652 * Timing vulnerability in DSA signature generation
1654 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1655 timing side channel attack. An attacker could use variations in the signing
1656 algorithm to recover the private key.
1658 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1663 * Timing vulnerability in ECDSA signature generation
1665 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1666 timing side channel attack. An attacker could use variations in the signing
1667 algorithm to recover the private key.
1669 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1674 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1675 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1676 of two gigabytes and the error handling improved.
1678 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1679 categorized as a normal bug, not a security issue, because the DRBG reseeds
1680 automatically and is fully functional even without additional randomness
1681 provided by the application.
1683 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1685 * Add a new ClientHello callback. Provides a callback interface that gives
1686 the application the ability to adjust the nascent SSL object at the
1687 earliest stage of ClientHello processing, immediately after extensions have
1688 been collected but before they have been processed. In particular, this
1689 callback can adjust the supported TLS versions in response to the contents
1694 * Add SM2 base algorithm support.
1698 * s390x assembly pack: add (improved) hardware-support for the following
1699 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1700 aes-cfb/cfb8, aes-ecb.
1704 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1705 parameter is no longer accepted, as it leads to a corrupt table. NULL
1706 pem_str is reserved for alias entries only.
1710 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1711 step for prime curves. The new implementation is based on formulae from
1712 differential addition-and-doubling in homogeneous projective coordinates
1713 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1714 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1715 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1716 to work in projective coordinates.
1718 *Billy Bob Brumley, Nicola Tuveri*
1720 * Change generating and checking of primes so that the error rate of not
1721 being prime depends on the intended use based on the size of the input.
1722 For larger primes this will result in more rounds of Miller-Rabin.
1723 The maximal error rate for primes with more than 1080 bits is lowered
1726 *Kurt Roeckx, Annie Yousar*
1728 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1732 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1733 moving between systems, and to avoid confusion when a Windows build is
1734 done with mingw vs with MSVC. For POSIX installs, there's still a
1735 symlink or copy named 'tsget' to avoid that confusion as well.
1739 * Revert blinding in ECDSA sign and instead make problematic addition
1740 length-invariant. Switch even to fixed-length Montgomery multiplication.
1744 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1745 step for binary curves. The new implementation is based on formulae from
1746 differential addition-and-doubling in mixed Lopez-Dahab projective
1747 coordinates, modified to independently blind the operands.
1749 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1751 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1752 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1753 EC_METHODs to implement their own specialized "ladder step", to take
1754 advantage of more favorable coordinate systems or more efficient
1755 differential addition-and-doubling algorithms.
1757 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1759 * Modified the random device based seed sources to keep the relevant
1760 file descriptors open rather than reopening them on each access.
1761 This allows such sources to operate in a chroot() jail without
1762 the associated device nodes being available. This behaviour can be
1763 controlled using RAND_keep_random_devices_open().
1767 * Numerous side-channel attack mitigations have been applied. This may have
1768 performance impacts for some algorithms for the benefit of improved
1769 security. Specific changes are noted in this change log by their respective
1774 * AIX shared library support overhaul. Switch to AIX "natural" way of
1775 handling shared libraries, which means collecting shared objects of
1776 different versions and bitnesses in one common archive. This allows to
1777 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1778 doesn't affect the way 3rd party applications are linked, only how
1779 multi-version installation is managed.
1783 * Make ec_group_do_inverse_ord() more robust and available to other
1784 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1785 mitigations are applied to the fallback BN_mod_inverse().
1786 When using this function rather than BN_mod_inverse() directly, new
1787 EC cryptosystem implementations are then safer-by-default.
1791 * Add coordinate blinding for EC_POINT and implement projective
1792 coordinate blinding for generic prime curves as a countermeasure to
1793 chosen point SCA attacks.
1795 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1797 * Add blinding to ECDSA and DSA signatures to protect against side channel
1798 attacks discovered by Keegan Ryan (NCC Group).
1802 * Enforce checking in the pkeyutl command line app to ensure that the input
1803 length does not exceed the maximum supported digest length when performing
1804 a sign, verify or verifyrecover operation.
1808 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1809 I/O in combination with something like select() or poll() will hang. This
1810 can be turned off again using SSL_CTX_clear_mode().
1811 Many applications do not properly handle non-application data records, and
1812 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1813 around the problems in those applications, but can also break some.
1814 It's recommended to read the manpages about SSL_read(), SSL_write(),
1815 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1816 SSL_CTX_set_read_ahead() again.
1820 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1821 now allow empty (zero character) pass phrases.
1825 * Apply blinding to binary field modular inversion and remove patent
1826 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1830 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1831 binary and prime elliptic curves.
1835 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1836 constant time fixed point multiplication.
1840 * Revise elliptic curve scalar multiplication with timing attack
1841 defenses: ec_wNAF_mul redirects to a constant time implementation
1842 when computing fixed point and variable point multiplication (which
1843 in OpenSSL are mostly used with secret scalars in keygen, sign,
1844 ECDH derive operations).
1845 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1848 * Updated CONTRIBUTING
1852 * Updated DRBG / RAND to request nonce and additional low entropy
1853 randomness from the system.
1855 *Matthias St. Pierre*
1857 * Updated 'openssl rehash' to use OpenSSL consistent default.
1861 * Moved the load of the ssl_conf module to libcrypto, which helps
1862 loading engines that libssl uses before libssl is initialised.
1866 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1870 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1872 *Ingo Schwarze, Rich Salz*
1874 * Added output of accepting IP address and port for 'openssl s_server'
1878 * Added a new API for TLSv1.3 ciphersuites:
1879 SSL_CTX_set_ciphersuites()
1880 SSL_set_ciphersuites()
1884 * Memory allocation failures consistently add an error to the error
1889 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1890 in libcrypto when run as setuid/setgid.
1894 * Load any config file by default when libssl is used.
1898 * Added new public header file <openssl/rand_drbg.h> and documentation
1899 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1901 *Matthias St. Pierre*
1903 * QNX support removed (cannot find contributors to get their approval
1904 for the license change).
1908 * TLSv1.3 replay protection for early data has been implemented. See the
1909 SSL_read_early_data() man page for further details.
1913 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1914 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1915 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1916 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1917 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1918 configuration has been separated out. See the ciphers man page or the
1919 SSL_CTX_set_ciphersuites() man page for more information.
1923 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1924 in responder mode now supports the new "-multi" option, which
1925 spawns the specified number of child processes to handle OCSP
1926 requests. The "-timeout" option now also limits the OCSP
1927 responder's patience to wait to receive the full client request
1928 on a newly accepted connection. Child processes are respawned
1929 as needed, and the CA index file is automatically reloaded
1930 when changed. This makes it possible to run the "ocsp" responder
1931 as a long-running service, making the OpenSSL CA somewhat more
1932 feature-complete. In this mode, most diagnostic messages logged
1933 after entering the event loop are logged via syslog(3) rather than
1938 * Added support for X448 and Ed448. Heavily based on original work by
1943 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1944 objects loaded. This adds the functions OSSL_STORE_expect() and
1945 OSSL_STORE_find() as well as needed tools to construct searches and
1946 get the search data out of them.
1950 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1951 version of OpenSSL should review their configuration settings to ensure
1952 that they are still appropriate for TLSv1.3. For further information see:
1953 <https://wiki.openssl.org/index.php/TLS1.3>
1957 * Grand redesign of the OpenSSL random generator
1959 The default RAND method now utilizes an AES-CTR DRBG according to
1960 NIST standard SP 800-90Ar1. The new random generator is essentially
1961 a port of the default random generator from the OpenSSL FIPS 2.0
1962 object module. It is a hybrid deterministic random bit generator
1963 using an AES-CTR bit stream and which seeds and reseeds itself
1964 automatically using trusted system entropy sources.
1966 Some of its new features are:
1967 - Support for multiple DRBG instances with seed chaining.
1968 - The default RAND method makes use of a DRBG.
1969 - There is a public and private DRBG instance.
1970 - The DRBG instances are fork-safe.
1971 - Keep all global DRBG instances on the secure heap if it is enabled.
1972 - The public and private DRBG instance are per thread for lock free
1975 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1977 * Changed Configure so it only says what it does and doesn't dump
1978 so much data. Instead, ./configdata.pm should be used as a script
1979 to display all sorts of configuration data.
1983 * Added processing of "make variables" to Configure.
1987 * Added SHA512/224 and SHA512/256 algorithm support.
1991 * The last traces of Netware support, first removed in 1.1.0, have
1996 * Get rid of Makefile.shared, and in the process, make the processing
1997 of certain files (rc.obj, or the .def/.map/.opt files produced from
1998 the ordinal files) more visible and hopefully easier to trace and
1999 debug (or make silent).
2003 * Make it possible to have environment variable assignments as
2004 arguments to config / Configure.
2008 * Add multi-prime RSA (RFC 8017) support.
2012 * Add SM3 implemented according to GB/T 32905-2016
2013 *Jack Lloyd <jack.lloyd@ribose.com>,*
2014 *Ronald Tse <ronald.tse@ribose.com>,*
2015 *Erick Borsboom <erick.borsboom@ribose.com>*
2017 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2018 as documented in RFC6066.
2019 Based on a patch from Tomasz Moń
2021 *Filipe Raimundo da Silva*
2023 * Add SM4 implemented according to GB/T 32907-2016.
2024 *Jack Lloyd <jack.lloyd@ribose.com>,*
2025 *Ronald Tse <ronald.tse@ribose.com>,*
2026 *Erick Borsboom <erick.borsboom@ribose.com>*
2028 * Reimplement -newreq-nodes and ERR_error_string_n; the
2029 original author does not agree with the license change.
2033 * Add ARIA AEAD TLS support.
2037 * Some macro definitions to support VS6 have been removed. Visual
2038 Studio 6 has not worked since 1.1.0
2042 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2043 without clearing the errors.
2047 * Add "atfork" functions. If building on a system that without
2048 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2049 requirements. The RAND facility now uses/requires this.
2057 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2058 not possible to disable entirely. However, it's still possible to
2059 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2062 To disable, configure with 'no-ui-console'. 'no-ui' is still
2063 possible to use as an alias. Check at compile time with the
2064 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2065 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2069 * Add a STORE module, which implements a uniform and URI based reader of
2070 stores that can contain keys, certificates, CRLs and numerous other
2071 objects. The main API is loosely based on a few stdio functions,
2072 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2073 OSSL_STORE_error and OSSL_STORE_close.
2074 The implementation uses backends called "loaders" to implement arbitrary
2075 URI schemes. There is one built in "loader" for the 'file' scheme.
2079 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2080 then adjusted to work on FreeBSD 8.4 as well.
2081 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2082 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2086 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2087 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2088 error code calls like this:
2090 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2092 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2093 that can be encoded in C. For the foreseeable future, this will only
2096 *Richard Levitte and Tim Hudson*
2098 * Removed BSD cryptodev engine.
2102 * Add a build target 'build_all_generated', to build all generated files
2103 and only that. This can be used to prepare everything that requires
2104 things like perl for a system that lacks perl and then move everything
2105 to that system and do the rest of the build there.
2109 * In the UI interface, make it possible to duplicate the user data. This
2110 can be used by engines that need to retain the data for a longer time
2111 than just the call where this user data is passed.
2115 * Ignore the '-named_curve auto' value for compatibility of applications
2118 *Tomas Mraz <tmraz@fedoraproject.org>*
2120 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2121 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2122 alerts across multiple records (some of which could be empty). In practice
2123 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2124 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2125 support this at all. Supporting it adds significant complexity to the
2126 record layer, and its removal is unlikely to cause interoperability
2131 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2132 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2133 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2138 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2139 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2141 *Richard Levitte, Andy Polyakov*
2143 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2148 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2149 platform rather than 'mingw'.
2153 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2154 success if they are asked to add an object which already exists
2155 in the store. This change cascades to other functions which load
2156 certificates and CRLs.
2160 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2161 facilitate stack unwinding even from assembly subroutines.
2165 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2166 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2170 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2171 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2172 which is the minimum version we support.
2176 * Certificate time validation (X509_cmp_time) enforces stricter
2177 compliance with RFC 5280. Fractional seconds and timezone offsets
2178 are no longer allowed.
2182 * Add support for ARIA
2186 * s_client will now send the Server Name Indication (SNI) extension by
2187 default unless the new "-noservername" option is used. The server name is
2188 based on the host provided to the "-connect" option unless overridden by
2189 using "-servername".
2193 * Add support for SipHash
2197 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2198 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2199 prevent issues where no progress is being made and the peer continually
2200 sends unrecognised record types, using up resources processing them.
2204 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2205 using the algorithm defined in
2206 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2210 * Heartbeat support has been removed; the ABI is changed for now.
2212 *Richard Levitte, Rich Salz*
2214 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2218 * The RSA "null" method, which was partially supported to avoid patent
2219 issues, has been replaced to always returns NULL.
2226 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2228 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2229 used even when parsing explicit parameters, when loading a encoded key
2230 or calling `EC_GROUP_new_from_ecpkparameters()`/
2231 `EC_GROUP_new_from_ecparameters()`.
2232 This prevents bypass of security hardening and performance gains,
2233 especially for curves with specialized EC_METHODs.
2234 By default, if a key encoded with explicit parameters is loaded and later
2235 encoded, the output is still encoded with explicit parameters, even if
2236 internally a "named" EC_GROUP is used for computation.
2240 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2241 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2242 NULL. After this change, only the cofactor parameter can be NULL. It also
2243 does some minimal sanity checks on the passed order.
2248 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2249 An attack is simple, if the first CMS_recipientInfo is valid but the
2250 second CMS_recipientInfo is chosen ciphertext. If the second
2251 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2252 encryption key will be replaced by garbage, and the message cannot be
2253 decoded, but if the RSA decryption fails, the correct encryption key is
2254 used and the recipient will not notice the attack.
2255 As a work around for this potential attack the length of the decrypted
2256 key must be equal to the cipher default key length, in case the
2257 certifiate is not given and all recipientInfo are tried out.
2258 The old behaviour can be re-enabled in the CMS code by setting the
2259 CMS_DEBUG_DECRYPT flag.
2264 * Use Windows installation paths in the mingw builds
2266 Mingw isn't a POSIX environment per se, which means that Windows
2267 paths should be used for installation.
2272 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2274 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2275 This changes the size when using the genpkey app when no size is given. It
2276 fixes an omission in earlier changes that changed all RSA, DSA and DH
2277 generation apps to use 2048 bits by default.
2281 * Prevent over long nonces in ChaCha20-Poly1305.
2283 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2284 for every encryption operation. RFC 7539 specifies that the nonce value
2285 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2286 and front pads the nonce with 0 bytes if it is less than 12
2287 bytes. However it also incorrectly allows a nonce to be set of up to 16
2288 bytes. In this case only the last 12 bytes are significant and any
2289 additional leading bytes are ignored.
2291 It is a requirement of using this cipher that nonce values are
2292 unique. Messages encrypted using a reused nonce value are susceptible to
2293 serious confidentiality and integrity attacks. If an application changes
2294 the default nonce length to be longer than 12 bytes and then makes a
2295 change to the leading bytes of the nonce expecting the new value to be a
2296 new unique nonce then such an application could inadvertently encrypt
2297 messages with a reused nonce.
2299 Additionally the ignored bytes in a long nonce are not covered by the
2300 integrity guarantee of this cipher. Any application that relies on the
2301 integrity of these ignored leading bytes of a long nonce may be further
2302 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2303 is safe because no such use sets such a long nonce value. However user
2304 applications that use this cipher directly and set a non-default nonce
2305 length to be longer than 12 bytes may be vulnerable.
2307 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2313 * Added SCA hardening for modular field inversion in EC_GROUP through
2314 a new dedicated field_inv() pointer in EC_METHOD.
2315 This also addresses a leakage affecting conversions from projective
2316 to affine coordinates.
2318 *Billy Bob Brumley, Nicola Tuveri*
2320 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2321 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2325 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2329 * Remove the 'dist' target and add a tarball building script. The
2330 'dist' target has fallen out of use, and it shouldn't be
2331 necessary to configure just to create a source distribution.
2335 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2337 * Timing vulnerability in DSA signature generation
2339 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2340 timing side channel attack. An attacker could use variations in the signing
2341 algorithm to recover the private key.
2343 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2348 * Timing vulnerability in ECDSA signature generation
2350 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2351 timing side channel attack. An attacker could use variations in the signing
2352 algorithm to recover the private key.
2354 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2359 * Add coordinate blinding for EC_POINT and implement projective
2360 coordinate blinding for generic prime curves as a countermeasure to
2361 chosen point SCA attacks.
2363 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2365 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2367 * Client DoS due to large DH parameter
2369 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2370 malicious server can send a very large prime value to the client. This will
2371 cause the client to spend an unreasonably long period of time generating a
2372 key for this prime resulting in a hang until the client has finished. This
2373 could be exploited in a Denial Of Service attack.
2375 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2380 * Cache timing vulnerability in RSA Key Generation
2382 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2383 a cache timing side channel attack. An attacker with sufficient access to
2384 mount cache timing attacks during the RSA key generation process could
2385 recover the private key.
2387 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2388 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2393 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2394 parameter is no longer accepted, as it leads to a corrupt table. NULL
2395 pem_str is reserved for alias entries only.
2399 * Revert blinding in ECDSA sign and instead make problematic addition
2400 length-invariant. Switch even to fixed-length Montgomery multiplication.
2404 * Change generating and checking of primes so that the error rate of not
2405 being prime depends on the intended use based on the size of the input.
2406 For larger primes this will result in more rounds of Miller-Rabin.
2407 The maximal error rate for primes with more than 1080 bits is lowered
2410 *Kurt Roeckx, Annie Yousar*
2412 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2416 * Add blinding to ECDSA and DSA signatures to protect against side channel
2417 attacks discovered by Keegan Ryan (NCC Group).
2421 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2422 now allow empty (zero character) pass phrases.
2426 * Certificate time validation (X509_cmp_time) enforces stricter
2427 compliance with RFC 5280. Fractional seconds and timezone offsets
2428 are no longer allowed.
2432 * Fixed a text canonicalisation bug in CMS
2434 Where a CMS detached signature is used with text content the text goes
2435 through a canonicalisation process first prior to signing or verifying a
2436 signature. This process strips trailing space at the end of lines, converts
2437 line terminators to CRLF and removes additional trailing line terminators
2438 at the end of a file. A bug in the canonicalisation process meant that
2439 some characters, such as form-feed, were incorrectly treated as whitespace
2440 and removed. This is contrary to the specification (RFC5485). This fix
2441 could mean that detached text data signed with an earlier version of
2442 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2443 signed with a fixed OpenSSL may fail to verify with an earlier version of
2444 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2445 and use the "-binary" flag (for the "cms" command line application) or set
2446 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2450 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2452 * Constructed ASN.1 types with a recursive definition could exceed the stack
2454 Constructed ASN.1 types with a recursive definition (such as can be found
2455 in PKCS7) could eventually exceed the stack given malicious input with
2456 excessive recursion. This could result in a Denial Of Service attack. There
2457 are no such structures used within SSL/TLS that come from untrusted sources
2458 so this is considered safe.
2460 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2466 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2468 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2469 effectively reduced to only comparing the least significant bit of each
2470 byte. This allows an attacker to forge messages that would be considered as
2471 authenticated in an amount of tries lower than that guaranteed by the
2472 security claims of the scheme. The module can only be compiled by the
2473 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2475 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2481 * Add a build target 'build_all_generated', to build all generated files
2482 and only that. This can be used to prepare everything that requires
2483 things like perl for a system that lacks perl and then move everything
2484 to that system and do the rest of the build there.
2488 * Backport SSL_OP_NO_RENGOTIATION
2490 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2491 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2492 changes this is no longer possible in 1.1.0. Therefore the new
2493 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2494 1.1.0 to provide equivalent functionality.
2496 Note that if an application built against 1.1.0h headers (or above) is run
2497 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2498 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2502 * Removed the OS390-Unix config target. It relied on a script that doesn't
2507 * rsaz_1024_mul_avx2 overflow bug on x86_64
2509 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2510 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2511 Analysis suggests that attacks against RSA and DSA as a result of this
2512 defect would be very difficult to perform and are not believed likely.
2513 Attacks against DH1024 are considered just feasible, because most of the
2514 work necessary to deduce information about a private key may be performed
2515 offline. The amount of resources required for such an attack would be
2516 significant. However, for an attack on TLS to be meaningful, the server
2517 would have to share the DH1024 private key among multiple clients, which is
2518 no longer an option since CVE-2016-0701.
2520 This only affects processors that support the AVX2 but not ADX extensions
2521 like Intel Haswell (4th generation).
2523 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2524 was originally found via the OSS-Fuzz project.
2529 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2531 * bn_sqrx8x_internal carry bug on x86_64
2533 There is a carry propagating bug in the x86_64 Montgomery squaring
2534 procedure. No EC algorithms are affected. Analysis suggests that attacks
2535 against RSA and DSA as a result of this defect would be very difficult to
2536 perform and are not believed likely. Attacks against DH are considered just
2537 feasible (although very difficult) because most of the work necessary to
2538 deduce information about a private key may be performed offline. The amount
2539 of resources required for such an attack would be very significant and
2540 likely only accessible to a limited number of attackers. An attacker would
2541 additionally need online access to an unpatched system using the target
2542 private key in a scenario with persistent DH parameters and a private
2543 key that is shared between multiple clients.
2545 This only affects processors that support the BMI1, BMI2 and ADX extensions
2546 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2548 This issue was reported to OpenSSL by the OSS-Fuzz project.
2553 * Malformed X.509 IPAddressFamily could cause OOB read
2555 If an X.509 certificate has a malformed IPAddressFamily extension,
2556 OpenSSL could do a one-byte buffer overread. The most likely result
2557 would be an erroneous display of the certificate in text format.
2559 This issue was reported to OpenSSL by the OSS-Fuzz project.
2564 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2566 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2567 platform rather than 'mingw'.
2571 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2572 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2573 which is the minimum version we support.
2577 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2579 * Encrypt-Then-Mac renegotiation crash
2581 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2582 negotiated where it was not in the original handshake (or vice-versa) then
2583 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2584 and servers are affected.
2586 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2591 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2593 * Truncated packet could crash via OOB read
2595 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2596 cipher is being used, then a truncated packet can cause that host to
2597 perform an out-of-bounds read, usually resulting in a crash.
2599 This issue was reported to OpenSSL by Robert Święcki of Google.
2604 * Bad (EC)DHE parameters cause a client crash
2606 If a malicious server supplies bad parameters for a DHE or ECDHE key
2607 exchange then this can result in the client attempting to dereference a
2608 NULL pointer leading to a client crash. This could be exploited in a Denial
2611 This issue was reported to OpenSSL by Guido Vranken.
2616 * BN_mod_exp may produce incorrect results on x86_64
2618 There is a carry propagating bug in the x86_64 Montgomery squaring
2619 procedure. No EC algorithms are affected. Analysis suggests that attacks
2620 against RSA and DSA as a result of this defect would be very difficult to
2621 perform and are not believed likely. Attacks against DH are considered just
2622 feasible (although very difficult) because most of the work necessary to
2623 deduce information about a private key may be performed offline. The amount
2624 of resources required for such an attack would be very significant and
2625 likely only accessible to a limited number of attackers. An attacker would
2626 additionally need online access to an unpatched system using the target
2627 private key in a scenario with persistent DH parameters and a private
2628 key that is shared between multiple clients. For example this can occur by
2629 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2630 similar to CVE-2015-3193 but must be treated as a separate problem.
2632 This issue was reported to OpenSSL by the OSS-Fuzz project.
2637 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2639 * ChaCha20/Poly1305 heap-buffer-overflow
2641 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2642 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2643 crash. This issue is not considered to be exploitable beyond a DoS.
2645 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2650 * CMS Null dereference
2652 Applications parsing invalid CMS structures can crash with a NULL pointer
2653 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2654 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2655 structure callback if an attempt is made to free certain invalid encodings.
2656 Only CHOICE structures using a callback which do not handle NULL value are
2659 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2664 * Montgomery multiplication may produce incorrect results
2666 There is a carry propagating bug in the Broadwell-specific Montgomery
2667 multiplication procedure that handles input lengths divisible by, but
2668 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2669 and DH private keys are impossible. This is because the subroutine in
2670 question is not used in operations with the private key itself and an input
2671 of the attacker's direct choice. Otherwise the bug can manifest itself as
2672 transient authentication and key negotiation failures or reproducible
2673 erroneous outcome of public-key operations with specially crafted input.
2674 Among EC algorithms only Brainpool P-512 curves are affected and one
2675 presumably can attack ECDH key negotiation. Impact was not analyzed in
2676 detail, because pre-requisites for attack are considered unlikely. Namely
2677 multiple clients have to choose the curve in question and the server has to
2678 share the private key among them, neither of which is default behaviour.
2679 Even then only clients that chose the curve will be affected.
2681 This issue was publicly reported as transient failures and was not
2682 initially recognized as a security issue. Thanks to Richard Morgan for
2683 providing reproducible case.
2688 * Removed automatic addition of RPATH in shared libraries and executables,
2689 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2693 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2695 * Fix Use After Free for large message sizes
2697 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2698 message larger than approx 16k is received then the underlying buffer to
2699 store the incoming message is reallocated and moved. Unfortunately a
2700 dangling pointer to the old location is left which results in an attempt to
2701 write to the previously freed location. This is likely to result in a
2702 crash, however it could potentially lead to execution of arbitrary code.
2704 This issue only affects OpenSSL 1.1.0a.
2706 This issue was reported to OpenSSL by Robert Święcki.
2711 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2713 * OCSP Status Request extension unbounded memory growth
2715 A malicious client can send an excessively large OCSP Status Request
2716 extension. If that client continually requests renegotiation, sending a
2717 large OCSP Status Request extension each time, then there will be unbounded
2718 memory growth on the server. This will eventually lead to a Denial Of
2719 Service attack through memory exhaustion. Servers with a default
2720 configuration are vulnerable even if they do not support OCSP. Builds using
2721 the "no-ocsp" build time option are not affected.
2723 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2728 * SSL_peek() hang on empty record
2730 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2731 sends an empty record. This could be exploited by a malicious peer in a
2732 Denial Of Service attack.
2734 This issue was reported to OpenSSL by Alex Gaynor.
2739 * Excessive allocation of memory in tls_get_message_header() and
2740 dtls1_preprocess_fragment()
2742 A (D)TLS message includes 3 bytes for its length in the header for the
2743 message. This would allow for messages up to 16Mb in length. Messages of
2744 this length are excessive and OpenSSL includes a check to ensure that a
2745 peer is sending reasonably sized messages in order to avoid too much memory
2746 being consumed to service a connection. A flaw in the logic of version
2747 1.1.0 means that memory for the message is allocated too early, prior to
2748 the excessive message length check. Due to way memory is allocated in
2749 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2750 to service a connection. This could lead to a Denial of Service through
2751 memory exhaustion. However, the excessive message length check still takes
2752 place, and this would cause the connection to immediately fail. Assuming
2753 that the application calls SSL_free() on the failed connection in a timely
2754 manner then the 21Mb of allocated memory will then be immediately freed
2755 again. Therefore the excessive memory allocation will be transitory in
2756 nature. This then means that there is only a security impact if:
2758 1) The application does not call SSL_free() in a timely manner in the event
2759 that the connection fails
2761 2) The application is working in a constrained environment where there is
2762 very little free memory
2764 3) The attacker initiates multiple connection attempts such that there are
2765 multiple connections in a state where memory has been allocated for the
2766 connection; SSL_free() has not yet been called; and there is insufficient
2767 memory to service the multiple requests.
2769 Except in the instance of (1) above any Denial Of Service is likely to be
2770 transitory because as soon as the connection fails the memory is
2771 subsequently freed again in the SSL_free() call. However there is an
2772 increased risk during this period of application crashes due to the lack of
2773 memory - which would then mean a more serious Denial of Service.
2775 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2776 (CVE-2016-6307 and CVE-2016-6308)
2780 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2781 had to be removed. Primary reason is that vendor assembler can't
2782 assemble our modules with -KPIC flag. As result it, assembly
2783 support, was not even available as option. But its lack means
2784 lack of side-channel resistant code, which is incompatible with
2785 security by todays standards. Fortunately gcc is readily available
2786 prepackaged option, which we firmly point at...
2790 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2792 * Windows command-line tool supports UTF-8 opt-in option for arguments
2793 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2794 (to any value) allows Windows user to access PKCS#12 file generated
2795 with Windows CryptoAPI and protected with non-ASCII password, as well
2796 as files generated under UTF-8 locale on Linux also protected with
2801 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2802 have been disabled by default and removed from DEFAULT, just like RC4.
2803 See the RC4 item below to re-enable both.
2807 * The method for finding the storage location for the Windows RAND seed file
2808 has changed. First we check %RANDFILE%. If that is not set then we check
2809 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2810 all else fails we fall back to C:\.
2814 * The EVP_EncryptUpdate() function has had its return type changed from void
2815 to int. A return of 0 indicates and error while a return of 1 indicates
2820 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2821 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2822 off the constant time implementation for RSA, DSA and DH have been made
2823 no-ops and deprecated.
2827 * Windows RAND implementation was simplified to only get entropy by
2828 calling CryptGenRandom(). Various other RAND-related tickets
2831 *Joseph Wylie Yandle, Rich Salz*
2833 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2834 and `OPENSSL_LH_`, respectively. The old names are available
2835 with API compatibility. They new names are now completely documented.
2839 * Unify TYPE_up_ref(obj) methods signature.
2840 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2841 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2842 int (instead of void) like all others TYPE_up_ref() methods.
2843 So now these methods also check the return value of CRYPTO_atomic_add(),
2844 and the validity of object reference counter.
2846 *fdasilvayy@gmail.com*
2848 * With Windows Visual Studio builds, the .pdb files are installed
2849 alongside the installed libraries and executables. For a static
2850 library installation, ossl_static.pdb is the associate compiler
2851 generated .pdb file to be used when linking programs.
2855 * Remove openssl.spec. Packaging files belong with the packagers.
2859 * Automatic Darwin/OSX configuration has had a refresh, it will now
2860 recognise x86_64 architectures automatically. You can still decide
2861 to build for a different bitness with the environment variable
2862 KERNEL_BITS (can be 32 or 64), for example:
2864 KERNEL_BITS=32 ./config
2868 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2869 256 bit AES and HMAC with SHA256.
2873 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2877 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2881 * To enable users to have their own config files and build file templates,
2882 Configure looks in the directory indicated by the environment variable
2883 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2884 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2885 name and is used as is.
2889 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2890 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2891 X509_CERT_FILE_CTX was removed.
2895 * "shared" builds are now the default. To create only static libraries use
2896 the "no-shared" Configure option.
2900 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2901 All of these option have not worked for some while and are fundamental
2906 * Make various cleanup routines no-ops and mark them as deprecated. Most
2907 global cleanup functions are no longer required because they are handled
2908 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2909 Explicitly de-initing can cause problems (e.g. where a library that uses
2910 OpenSSL de-inits, but an application is still using it). The affected
2911 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2912 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2913 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2914 COMP_zlib_cleanup().
2918 * --strict-warnings no longer enables runtime debugging options
2919 such as REF_DEBUG. Instead, debug options are automatically
2920 enabled with '--debug' builds.
2922 *Andy Polyakov, Emilia Käsper*
2924 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2925 have been moved out of the public header files. New functions for managing
2926 these have been added.
2930 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2931 objects have been moved out of the public header files. New
2932 functions for managing these have been added.
2936 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2937 have been moved out of the public header files. New functions for managing
2938 these have been added.
2942 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2943 moved out of the public header files. New functions for managing these
2948 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2952 * Removed the mk1mf build scripts.
2956 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2957 it is always safe to #include a header now.
2961 * Removed the aged BC-32 config and all its supporting scripts
2965 * Removed support for Ultrix, Netware, and OS/2.
2969 * Add support for HKDF.
2971 *Alessandro Ghedini*
2973 * Add support for blake2b and blake2s
2977 * Added support for "pipelining". Ciphers that have the
2978 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2979 encryptions/decryptions simultaneously. There are currently no built-in
2980 ciphers with this property but the expectation is that engines will be able
2981 to offer it to significantly improve throughput. Support has been extended
2982 into libssl so that multiple records for a single connection can be
2983 processed in one go (for >=TLS 1.1).
2987 * Added the AFALG engine. This is an async capable engine which is able to
2988 offload work to the Linux kernel. In this initial version it only supports
2989 AES128-CBC. The kernel must be version 4.1.0 or greater.
2993 * OpenSSL now uses a new threading API. It is no longer necessary to
2994 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2995 are two supported threading models: pthreads and windows threads. It is
2996 also possible to configure OpenSSL at compile time for "no-threads". The
2997 old threading API should no longer be used. The functions have been
2998 replaced with "no-op" compatibility macros.
3000 *Alessandro Ghedini, Matt Caswell*
3002 * Modify behavior of ALPN to invoke callback after SNI/servername
3003 callback, such that updates to the SSL_CTX affect ALPN.
3007 * Add SSL_CIPHER queries for authentication and key-exchange.
3011 * Changes to the DEFAULT cipherlist:
3012 - Prefer (EC)DHE handshakes over plain RSA.
3013 - Prefer AEAD ciphers over legacy ciphers.
3014 - Prefer ECDSA over RSA when both certificates are available.
3015 - Prefer TLSv1.2 ciphers/PRF.
3016 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3021 * Change the ECC default curve list to be this, in order: x25519,
3022 secp256r1, secp521r1, secp384r1.
3026 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3027 disabled by default. They can be re-enabled using the
3028 enable-weak-ssl-ciphers option to Configure.
3032 * If the server has ALPN configured, but supports no protocols that the
3033 client advertises, send a fatal "no_application_protocol" alert.
3034 This behaviour is SHALL in RFC 7301, though it isn't universally
3035 implemented by other servers.
3039 * Add X25519 support.
3040 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3041 for public and private key encoding using the format documented in
3042 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3043 key generation and key derivation.
3045 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3050 * Deprecate SRP_VBASE_get_by_user.
3051 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3052 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3053 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3054 seed, even if the seed is configured.
3056 Users should use SRP_VBASE_get1_by_user instead. Note that in
3057 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3058 also that even though configuring the SRP seed attempts to hide
3059 invalid usernames by continuing the handshake with fake
3060 credentials, this behaviour is not constant time and no strong
3061 guarantees are made that the handshake is indistinguishable from
3062 that of a valid user.
3066 * Configuration change; it's now possible to build dynamic engines
3067 without having to build shared libraries and vice versa. This
3068 only applies to the engines in engines/, those in crypto/engine/
3069 will always be built into libcrypto (i.e. "static").
3071 Building dynamic engines is enabled by default; to disable, use
3072 the configuration option "disable-dynamic-engine".
3074 The only requirements for building dynamic engines are the
3075 presence of the DSO module and building with position independent
3076 code, so they will also automatically be disabled if configuring
3077 with "disable-dso" or "disable-pic".
3079 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3080 are also taken away from openssl/opensslconf.h, as they are
3085 * Configuration change; if there is a known flag to compile
3086 position independent code, it will always be applied on the
3087 libcrypto and libssl object files, and never on the application
3088 object files. This means other libraries that use routines from
3089 libcrypto / libssl can be made into shared libraries regardless
3090 of how OpenSSL was configured.
3092 If this isn't desirable, the configuration options "disable-pic"
3093 or "no-pic" can be used to disable the use of PIC. This will
3094 also disable building shared libraries and dynamic engines.
3098 * Removed JPAKE code. It was experimental and has no wide use.
3102 * The INSTALL_PREFIX Makefile variable has been renamed to
3103 DESTDIR. That makes for less confusion on what this variable
3104 is for. Also, the configuration option --install_prefix is
3109 * Heartbeat for TLS has been removed and is disabled by default
3110 for DTLS; configure with enable-heartbeats. Code that uses the
3111 old #define's might need to be updated.
3113 *Emilia Käsper, Rich Salz*
3115 * Rename REF_CHECK to REF_DEBUG.
3119 * New "unified" build system
3121 The "unified" build system is aimed to be a common system for all
3122 platforms we support. With it comes new support for VMS.
3124 This system builds supports building in a different directory tree
3125 than the source tree. It produces one Makefile (for unix family
3126 or lookalikes), or one descrip.mms (for VMS).
3128 The source of information to make the Makefile / descrip.mms is
3129 small files called 'build.info', holding the necessary
3130 information for each directory with source to compile, and a
3131 template in Configurations, like unix-Makefile.tmpl or
3134 With this change, the library names were also renamed on Windows
3135 and on VMS. They now have names that are closer to the standard
3136 on Unix, and include the major version number, and in certain
3137 cases, the architecture they are built for. See "Notes on shared
3138 libraries" in INSTALL.
3140 We rely heavily on the perl module Text::Template.
3144 * Added support for auto-initialisation and de-initialisation of the library.
3145 OpenSSL no longer requires explicit init or deinit routines to be called,
3146 except in certain circumstances. See the OPENSSL_init_crypto() and
3147 OPENSSL_init_ssl() man pages for further information.
3151 * The arguments to the DTLSv1_listen function have changed. Specifically the
3152 "peer" argument is now expected to be a BIO_ADDR object.
3154 * Rewrite of BIO networking library. The BIO library lacked consistent
3155 support of IPv6, and adding it required some more extensive
3156 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3157 which hold all types of addresses and chains of address information.
3158 It also introduces a new API, with functions like BIO_socket,
3159 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3160 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3161 have been adapted accordingly.
3165 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3170 * CRIME protection: disable compression by default, even if OpenSSL is
3171 compiled with zlib enabled. Applications can still enable compression
3172 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3173 using the SSL_CONF library to configure compression.
3177 * The signature of the session callback configured with
3178 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3179 was explicitly marked as `const unsigned char*` instead of
3184 * Always DPURIFY. Remove the use of uninitialized memory in the
3185 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3189 * Removed many obsolete configuration items, including
3190 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3191 MD2_CHAR, MD2_INT, MD2_LONG
3193 IDEA_SHORT, IDEA_LONG
3194 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3196 *Rich Salz, with advice from Andy Polyakov*
3198 * Many BN internals have been moved to an internal header file.
3200 *Rich Salz with help from Andy Polyakov*
3202 * Configuration and writing out the results from it has changed.
3203 Files such as Makefile include/openssl/opensslconf.h and are now
3204 produced through general templates, such as Makefile.in and
3205 crypto/opensslconf.h.in and some help from the perl module
3208 Also, the center of configuration information is no longer
3209 Makefile. Instead, Configure produces a perl module in
3210 configdata.pm which holds most of the config data (in the hash
3211 table %config), the target data that comes from the target
3212 configuration in one of the `Configurations/*.conf` files (in
3217 * To clarify their intended purposes, the Configure options
3218 --prefix and --openssldir change their semantics, and become more
3219 straightforward and less interdependent.
3221 --prefix shall be used exclusively to give the location INSTALLTOP
3222 where programs, scripts, libraries, include files and manuals are
3223 going to be installed. The default is now /usr/local.
3225 --openssldir shall be used exclusively to give the default
3226 location OPENSSLDIR where certificates, private keys, CRLs are
3227 managed. This is also where the default openssl.cnf gets
3229 If the directory given with this option is a relative path, the
3230 values of both the --prefix value and the --openssldir value will
3231 be combined to become OPENSSLDIR.
3232 The default for --openssldir is INSTALLTOP/ssl.
3234 Anyone who uses --openssldir to specify where OpenSSL is to be
3235 installed MUST change to use --prefix instead.
3239 * The GOST engine was out of date and therefore it has been removed. An up
3240 to date GOST engine is now being maintained in an external repository.
3241 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3242 support for GOST ciphersuites (these are only activated if a GOST engine
3247 * EGD is no longer supported by default; use enable-egd when
3250 *Ben Kaduk and Rich Salz*
3252 * The distribution now has Makefile.in files, which are used to
3253 create Makefile's when Configure is run. *Configure must be run
3254 before trying to build now.*
3258 * The return value for SSL_CIPHER_description() for error conditions
3263 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3265 Obtaining and performing DNSSEC validation of TLSA records is
3266 the application's responsibility. The application provides
3267 the TLSA records of its choice to OpenSSL, and these are then
3268 used to authenticate the peer.
3270 The TLSA records need not even come from DNS. They can, for
3271 example, be used to implement local end-entity certificate or
3272 trust-anchor "pinning", where the "pin" data takes the form
3273 of TLSA records, which can augment or replace verification
3274 based on the usual WebPKI public certification authorities.
3278 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3279 continues to support deprecated interfaces in default builds.
3280 However, applications are strongly advised to compile their
3281 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3282 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3283 or the 1.1.0 releases.
3285 In environments in which all applications have been ported to
3286 not use any deprecated interfaces OpenSSL's Configure script
3287 should be used with the --api=1.1.0 option to entirely remove
3288 support for the deprecated features from the library and
3289 unconditionally disable them in the installed headers.
3290 Essentially the same effect can be achieved with the "no-deprecated"
3291 argument to Configure, except that this will always restrict
3292 the build to just the latest API, rather than a fixed API
3295 As applications are ported to future revisions of the API,
3296 they should update their compile-time OPENSSL_API_COMPAT define
3297 accordingly, but in most cases should be able to continue to
3298 compile with later releases.
3300 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3301 0x10000000L and 0x00908000L, respectively. However those
3302 versions did not support the OPENSSL_API_COMPAT feature, and
3303 so applications are not typically tested for explicit support
3304 of just the undeprecated features of either release.
3308 * Add support for setting the minimum and maximum supported protocol.
3309 It can bet set via the SSL_set_min_proto_version() and
3310 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3311 MaxProtocol. It's recommended to use the new APIs to disable
3312 protocols instead of disabling individual protocols using
3313 SSL_set_options() or SSL_CONF's Protocol. This change also
3314 removes support for disabling TLS 1.2 in the OpenSSL TLS
3315 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3319 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3323 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3324 and integrates ECDSA and ECDH functionality into EC. Implementations can
3325 now redirect key generation and no longer need to convert to or from
3328 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3329 include the ec.h header file instead.
3333 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3334 ciphers who are no longer supported and drops support the ephemeral RSA key
3335 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3339 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3340 opaque. For HMAC_CTX, the following constructors and destructors
3343 HMAC_CTX *HMAC_CTX_new(void);
3344 void HMAC_CTX_free(HMAC_CTX *ctx);
3346 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3347 destroy such methods has been added. See EVP_MD_meth_new(3) and
3348 EVP_CIPHER_meth_new(3) for documentation.
3351 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3352 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3353 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3354 an already created structure.
3355 2) For consistency with the majority of our object creators and
3356 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3357 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3358 for deprecated builds.
3362 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3363 cryptographic operations to be performed asynchronously as long as an
3364 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3365 further details. Libssl has also had this capability integrated with the
3366 introduction of the new mode SSL_MODE_ASYNC and associated error
3367 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3368 pages. This work was developed in partnership with Intel Corp.
3372 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3373 always enabled now. If you want to disable the support you should
3374 exclude it using the list of supported ciphers. This also means that the
3375 "-no_ecdhe" option has been removed from s_server.
3379 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3380 SSL_{CTX_}set1_curves() which can set a list.
3384 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3385 curve you want to support using SSL_{CTX_}set1_curves().
3389 * State machine rewrite. The state machine code has been significantly
3390 refactored in order to remove much duplication of code and solve issues
3391 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3392 further details). This change does have some associated API changes.
3393 Notably the SSL_state() function has been removed and replaced by
3394 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3395 SSL_set_state() has been removed altogether. The previous handshake states
3396 defined in ssl.h and ssl3.h have also been removed.
3400 * All instances of the string "ssleay" in the public API were replaced
3401 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3402 Some error codes related to internal RSA_eay API's were renamed.
3406 * The demo files in crypto/threads were moved to demo/threads.
3410 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3413 *Matt Caswell, Rich Salz*
3415 * New ASN.1 embed macro.
3417 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3418 structure is not allocated: it is part of the parent. That is instead of
3426 This reduces memory fragmentation and make it impossible to accidentally
3427 set a mandatory field to NULL.
3429 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3430 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3431 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3436 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3440 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3441 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3442 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3443 DES and RC4 ciphersuites.
3447 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3448 This changes the decoding behaviour for some invalid messages,
3449 though the change is mostly in the more lenient direction, and
3450 legacy behaviour is preserved as much as possible.
3454 * Fix no-stdio build.
3455 *David Woodhouse <David.Woodhouse@intel.com> and also*
3456 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3458 * New testing framework
3459 The testing framework has been largely rewritten and is now using
3460 perl and the perl modules Test::Harness and an extended variant of
3461 Test::More called OpenSSL::Test to do its work. All test scripts in
3462 test/ have been rewritten into test recipes, and all direct calls to
3463 executables in test/Makefile have become individual recipes using the
3464 simplified testing OpenSSL::Test::Simple.
3466 For documentation on our testing modules, do:
3468 perldoc test/testlib/OpenSSL/Test/Simple.pm
3469 perldoc test/testlib/OpenSSL/Test.pm
3473 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3474 are used; the latter aborts on memory leaks (usually checked on exit).
3475 Some undocumented "set malloc, etc., hooks" functions were removed
3476 and others were changed. All are now documented.
3480 * In DSA_generate_parameters_ex, if the provided seed is too short,
3483 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3485 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3486 from RFC4279, RFC4785, RFC5487, RFC5489.
3488 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3489 original RSA_PSK patch.
3493 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3494 era flag was never set throughout the codebase (only read). Also removed
3495 SSL3_FLAGS_POP_BUFFER which was only used if
3496 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3500 * Changed the default name options in the "ca", "crl", "req" and "x509"
3501 to be "oneline" instead of "compat".
3505 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3506 not aware of clients that still exhibit this bug, and the workaround
3507 hasn't been working properly for a while.
3511 * The return type of BIO_number_read() and BIO_number_written() as well as
3512 the corresponding num_read and num_write members in the BIO structure has
3513 changed from unsigned long to uint64_t. On platforms where an unsigned
3514 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3519 * Given the pervasive nature of TLS extensions it is inadvisable to run
3520 OpenSSL without support for them. It also means that maintaining
3521 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3522 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3526 * Removed support for the two export grade static DH ciphersuites
3527 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3528 were newly added (along with a number of other static DH ciphersuites) to
3529 1.0.2. However the two export ones have *never* worked since they were
3530 introduced. It seems strange in any case to be adding new export
3531 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3535 * Version negotiation has been rewritten. In particular SSLv23_method(),
3536 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3537 and turned into macros which simply call the new preferred function names
3538 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3539 should use the new names instead. Also as part of this change the ssl23.h
3540 header file has been removed.
3544 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3545 code and the associated standard is no longer considered fit-for-purpose.
3549 * RT2547 was closed. When generating a private key, try to make the
3550 output file readable only by the owner. This behavior change might
3551 be noticeable when interacting with other software.
3553 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3558 * Added HTTP GET support to the ocsp command.
3562 * Changed default digest for the dgst and enc commands from MD5 to
3567 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3571 * Added support for TLS extended master secret from
3572 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3573 initial patch which was a great help during development.
3577 * All libssl internal structures have been removed from the public header
3578 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3579 now redundant). Users should not attempt to access internal structures
3580 directly. Instead they should use the provided API functions.
3584 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3585 Access to deprecated functions can be re-enabled by running config with
3586 "enable-deprecated". In addition applications wishing to use deprecated
3587 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3588 will, by default, disable some transitive includes that previously existed
3589 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3593 * Added support for OCB mode. OpenSSL has been granted a patent license
3594 compatible with the OpenSSL license for use of OCB. Details are available
3595 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3596 for OCB can be removed by calling config with no-ocb.
3600 * SSLv2 support has been removed. It still supports receiving a SSLv2
3601 compatible client hello.
3605 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3606 done while fixing the error code for the key-too-small case.
3608 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3610 * CA.sh has been removed; use CA.pl instead.
3614 * Removed old DES API.
3618 * Remove various unsupported platforms:
3624 Sinix/ReliantUNIX RM400
3629 16-bit platforms such as WIN16
3633 * Clean up OPENSSL_NO_xxx #define's
3634 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3635 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3636 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3637 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3638 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3639 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3640 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3641 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3642 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3643 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3647 * Cleaned up dead code
3648 Remove all but one '#ifdef undef' which is to be looked at.
3652 * Clean up calling of xxx_free routines.
3653 Just like free(), fix most of the xxx_free routines to accept
3654 NULL. Remove the non-null checks from callers. Save much code.
3658 * Add secure heap for storage of private keys (when possible).
3659 Add BIO_s_secmem(), CBIGNUM, etc.
3660 Contributed by Akamai Technologies under our Corporate CLA.
3664 * Experimental support for a new, fast, unbiased prime candidate generator,
3665 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3667 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3669 * New output format NSS in the sess_id command line tool. This allows
3670 exporting the session id and the master key in NSS keylog format.
3672 *Martin Kaiser <martin@kaiser.cx>*
3674 * Harmonize version and its documentation. -f flag is used to display
3677 *mancha <mancha1@zoho.com>*
3679 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3680 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3682 *mancha <mancha1@zoho.com>*
3684 * Fix some double frees. These are not thought to be exploitable.
3686 *mancha <mancha1@zoho.com>*
3688 * A missing bounds check in the handling of the TLS heartbeat extension
3689 can be used to reveal up to 64k of memory to a connected client or
3692 Thanks for Neel Mehta of Google Security for discovering this bug and to
3693 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3694 preparing the fix ([CVE-2014-0160])
3696 *Adam Langley, Bodo Moeller*
3698 * Fix for the attack described in the paper "Recovering OpenSSL
3699 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3700 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3701 <http://eprint.iacr.org/2014/140>
3703 Thanks to Yuval Yarom and Naomi Benger for discovering this
3704 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3706 *Yuval Yarom and Naomi Benger*
3708 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3709 this fixes a limitation in previous versions of OpenSSL.
3713 * Experimental encrypt-then-mac support.
3715 Experimental support for encrypt then mac from
3716 draft-gutmann-tls-encrypt-then-mac-02.txt
3718 To enable it set the appropriate extension number (0x42 for the test
3719 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3721 For non-compliant peers (i.e. just about everything) this should have no
3724 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3728 * Add EVP support for key wrapping algorithms, to avoid problems with
3729 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3730 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3731 algorithms and include tests cases.
3735 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3740 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3741 MGF1 digest and OAEP label.
3745 * Make openssl verify return errors.
3747 *Chris Palmer <palmer@google.com> and Ben Laurie*
3749 * New function ASN1_TIME_diff to calculate the difference between two
3750 ASN1_TIME structures or one structure and the current time.
3754 * Update fips_test_suite to support multiple command line options. New
3755 test to induce all self test errors in sequence and check expected
3760 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3761 sign or verify all in one operation.
3765 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3766 test programs and fips_test_suite. Includes functionality to parse
3767 the minimal script output of fipsalgest.pl directly.
3771 * Add authorisation parameter to FIPS_module_mode_set().
3775 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3779 * Use separate DRBG fields for internal and external flags. New function
3780 FIPS_drbg_health_check() to perform on demand health checking. Add
3781 generation tests to fips_test_suite with reduced health check interval to
3782 demonstrate periodic health checking. Add "nodh" option to
3783 fips_test_suite to skip very slow DH test.
3787 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3792 * More extensive health check for DRBG checking many more failure modes.
3793 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3794 combination: call this in fips_test_suite.
3798 * Add support for canonical generation of DSA parameter 'g'. See
3801 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3802 POST to handle HMAC cases.
3806 * Add functions FIPS_module_version() and FIPS_module_version_text()
3807 to return numerical and string versions of the FIPS module number.
3811 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3812 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3813 outside the validated module in the FIPS capable OpenSSL.
3817 * Minor change to DRBG entropy callback semantics. In some cases
3818 there is no multiple of the block length between min_len and
3819 max_len. Allow the callback to return more than max_len bytes
3820 of entropy but discard any extra: it is the callback's responsibility
3821 to ensure that the extra data discarded does not impact the
3822 requested amount of entropy.
3826 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3827 information in FIPS186-3, SP800-57 and SP800-131A.
3831 * CCM support via EVP. Interface is very similar to GCM case except we
3832 must supply all data in one chunk (i.e. no update, final) and the
3833 message length must be supplied if AAD is used. Add algorithm test
3838 * Initial version of POST overhaul. Add POST callback to allow the status
3839 of POST to be monitored and/or failures induced. Modify fips_test_suite
3840 to use callback. Always run all selftests even if one fails.
3844 * XTS support including algorithm test driver in the fips_gcmtest program.
3845 Note: this does increase the maximum key length from 32 to 64 bytes but
3846 there should be no binary compatibility issues as existing applications
3847 will never use XTS mode.
3851 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3852 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3853 performs algorithm blocking for unapproved PRNG types. Also do not
3854 set PRNG type in FIPS_mode_set(): leave this to the application.
3855 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3856 the standard OpenSSL PRNG: set additional data to a date time vector.
3860 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3861 This shouldn't present any incompatibility problems because applications
3862 shouldn't be using these directly and any that are will need to rethink
3863 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3867 * Extensive self tests and health checking required by SP800-90 DRBG.
3868 Remove strength parameter from FIPS_drbg_instantiate and always
3869 instantiate at maximum supported strength.
3873 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3877 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3881 * New function DH_compute_key_padded() to compute a DH key and pad with
3882 leading zeroes if needed: this complies with SP800-56A et al.
3886 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3887 anything, incomplete, subject to change and largely untested at present.
3891 * Modify fipscanisteronly build option to only build the necessary object
3892 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3896 * Add experimental option FIPSSYMS to give all symbols in
3897 fipscanister.o and FIPS or fips prefix. This will avoid
3898 conflicts with future versions of OpenSSL. Add perl script
3899 util/fipsas.pl to preprocess assembly language source files
3900 and rename any affected symbols.
3904 * Add selftest checks and algorithm block of non-fips algorithms in
3905 FIPS mode. Remove DES2 from selftests.
3909 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3910 return internal method without any ENGINE dependencies. Add new
3911 tiny fips sign and verify functions.
3915 * New build option no-ec2m to disable characteristic 2 code.
3919 * New build option "fipscanisteronly". This only builds fipscanister.o
3920 and (currently) associated fips utilities. Uses the file Makefile.fips
3921 instead of Makefile.org as the prototype.
3925 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3926 Update fips_gcmtest to use IV generator.
3930 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3931 setting output buffer to NULL. The `*Final` function must be
3932 called although it will not retrieve any additional data. The tag
3933 can be set or retrieved with a ctrl. The IV length is by default 12
3934 bytes (96 bits) but can be set to an alternative value. If the IV
3935 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3940 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3941 underlying do_cipher function handles all cipher semantics itself
3942 including padding and finalisation. This is useful if (for example)
3943 an ENGINE cipher handles block padding itself. The behaviour of
3944 do_cipher is subtly changed if this flag is set: the return value
3945 is the number of characters written to the output buffer (zero is
3946 no longer an error code) or a negative error code. Also if the
3947 input buffer is NULL and length 0 finalisation should be performed.
3951 * If a candidate issuer certificate is already part of the constructed
3952 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3956 * Improve forward-security support: add functions
3958 void SSL_CTX_set_not_resumable_session_callback(
3959 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3960 void SSL_set_not_resumable_session_callback(
3961 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3963 for use by SSL/TLS servers; the callback function will be called whenever a
3964 new session is created, and gets to decide whether the session may be
3965 cached to make it resumable (return 0) or not (return 1). (As by the
3966 SSL/TLS protocol specifications, the session_id sent by the server will be
3967 empty to indicate that the session is not resumable; also, the server will
3968 not generate RFC 4507 (RFC 5077) session tickets.)
3970 A simple reasonable callback implementation is to return is_forward_secure.
3971 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3972 by the SSL/TLS server library, indicating whether it can provide forward
3975 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3977 * New -verify_name option in command line utilities to set verification
3982 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3983 Add CMAC pkey methods.
3987 * Experimental renegotiation in s_server -www mode. If the client
3988 browses /reneg connection is renegotiated. If /renegcert it is
3989 renegotiated requesting a certificate.
3993 * Add an "external" session cache for debugging purposes to s_server. This
3994 should help trace issues which normally are only apparent in deployed
3995 multi-process servers.
3999 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4000 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4001 BIO_set_cipher() and some obscure PEM functions were changed so they
4002 can now return an error. The RAND changes required a change to the
4003 RAND_METHOD structure.
4007 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4008 a gcc attribute to warn if the result of a function is ignored. This
4009 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4010 whose return value is often ignored.
4014 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4015 These allow SCTs (signed certificate timestamps) to be requested and
4016 validated when establishing a connection.
4018 *Rob Percival <robpercival@google.com>*
4023 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4025 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4026 used even when parsing explicit parameters, when loading a encoded key
4027 or calling `EC_GROUP_new_from_ecpkparameters()`/
4028 `EC_GROUP_new_from_ecparameters()`.
4029 This prevents bypass of security hardening and performance gains,
4030 especially for curves with specialized EC_METHODs.
4031 By default, if a key encoded with explicit parameters is loaded and later
4032 encoded, the output is still encoded with explicit parameters, even if
4033 internally a "named" EC_GROUP is used for computation.
4037 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4038 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4039 NULL. After this change, only the cofactor parameter can be NULL. It also
4040 does some minimal sanity checks on the passed order.
4045 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4046 An attack is simple, if the first CMS_recipientInfo is valid but the
4047 second CMS_recipientInfo is chosen ciphertext. If the second
4048 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4049 encryption key will be replaced by garbage, and the message cannot be
4050 decoded, but if the RSA decryption fails, the correct encryption key is
4051 used and the recipient will not notice the attack.
4052 As a work around for this potential attack the length of the decrypted
4053 key must be equal to the cipher default key length, in case the
4054 certifiate is not given and all recipientInfo are tried out.
4055 The old behaviour can be re-enabled in the CMS code by setting the
4056 CMS_DEBUG_DECRYPT flag.
4061 * Document issue with installation paths in diverse Windows builds
4063 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4064 binaries and run-time config file.
4069 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4071 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4072 This changes the size when using the genpkey app when no size is given. It
4073 fixes an omission in earlier changes that changed all RSA, DSA and DH
4074 generation apps to use 2048 bits by default.
4078 * Add FIPS support for Android Arm 64-bit
4080 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4081 Module in Version 2.0.10. For some reason, the corresponding target
4082 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4083 built with FIPS support on Android Arm 64-bit. This omission has been
4086 *Matthias St. Pierre*
4088 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4090 * 0-byte record padding oracle
4092 If an application encounters a fatal protocol error and then calls
4093 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4094 then OpenSSL can respond differently to the calling application if a 0 byte
4095 record is received with invalid padding compared to if a 0 byte record is
4096 received with an invalid MAC. If the application then behaves differently
4097 based on that in a way that is detectable to the remote peer, then this
4098 amounts to a padding oracle that could be used to decrypt data.
4100 In order for this to be exploitable "non-stitched" ciphersuites must be in
4101 use. Stitched ciphersuites are optimised implementations of certain
4102 commonly used ciphersuites. Also the application must call SSL_shutdown()
4103 twice even if a protocol error has occurred (applications should not do
4104 this but some do anyway).
4106 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4107 Aviram, with additional investigation by Steven Collison and Andrew
4108 Hourselt. It was reported to OpenSSL on 10th December 2018.
4113 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4117 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4119 * Microarchitecture timing vulnerability in ECC scalar multiplication
4121 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4122 shown to be vulnerable to a microarchitecture timing side channel attack.
4123 An attacker with sufficient access to mount local timing attacks during
4124 ECDSA signature generation could recover the private key.
4126 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4127 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4133 * Timing vulnerability in DSA signature generation
4135 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4136 timing side channel attack. An attacker could use variations in the signing
4137 algorithm to recover the private key.
4139 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4144 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4145 Module, accidentally introduced while backporting security fixes from the
4146 development branch and hindering the use of ECC in FIPS mode.
4150 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4152 * Client DoS due to large DH parameter
4154 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4155 malicious server can send a very large prime value to the client. This will
4156 cause the client to spend an unreasonably long period of time generating a
4157 key for this prime resulting in a hang until the client has finished. This
4158 could be exploited in a Denial Of Service attack.
4160 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4165 * Cache timing vulnerability in RSA Key Generation
4167 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4168 a cache timing side channel attack. An attacker with sufficient access to
4169 mount cache timing attacks during the RSA key generation process could
4170 recover the private key.
4172 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4173 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4178 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4179 parameter is no longer accepted, as it leads to a corrupt table. NULL
4180 pem_str is reserved for alias entries only.
4184 * Revert blinding in ECDSA sign and instead make problematic addition
4185 length-invariant. Switch even to fixed-length Montgomery multiplication.
4189 * Change generating and checking of primes so that the error rate of not
4190 being prime depends on the intended use based on the size of the input.
4191 For larger primes this will result in more rounds of Miller-Rabin.
4192 The maximal error rate for primes with more than 1080 bits is lowered
4195 *Kurt Roeckx, Annie Yousar*
4197 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4201 * Add blinding to ECDSA and DSA signatures to protect against side channel
4202 attacks discovered by Keegan Ryan (NCC Group).
4206 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4207 now allow empty (zero character) pass phrases.
4211 * Certificate time validation (X509_cmp_time) enforces stricter
4212 compliance with RFC 5280. Fractional seconds and timezone offsets
4213 are no longer allowed.
4217 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4219 * Constructed ASN.1 types with a recursive definition could exceed the stack
4221 Constructed ASN.1 types with a recursive definition (such as can be found
4222 in PKCS7) could eventually exceed the stack given malicious input with
4223 excessive recursion. This could result in a Denial Of Service attack. There
4224 are no such structures used within SSL/TLS that come from untrusted sources
4225 so this is considered safe.
4227 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4233 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4235 * Read/write after SSL object in error state
4237 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4238 mechanism. The intent was that if a fatal error occurred during a handshake
4239 then OpenSSL would move into the error state and would immediately fail if
4240 you attempted to continue the handshake. This works as designed for the
4241 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4242 SSL_connect()), however due to a bug it does not work correctly if
4243 SSL_read() or SSL_write() is called directly. In that scenario, if the
4244 handshake fails then a fatal error will be returned in the initial function
4245 call. If SSL_read()/SSL_write() is subsequently called by the application
4246 for the same SSL object then it will succeed and the data is passed without
4247 being decrypted/encrypted directly from the SSL/TLS record layer.
4249 In order to exploit this issue an application bug would have to be present
4250 that resulted in a call to SSL_read()/SSL_write() being issued after having
4251 already received a fatal error.
4253 This issue was reported to OpenSSL by David Benjamin (Google).
4258 * rsaz_1024_mul_avx2 overflow bug on x86_64
4260 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4261 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4262 Analysis suggests that attacks against RSA and DSA as a result of this
4263 defect would be very difficult to perform and are not believed likely.
4264 Attacks against DH1024 are considered just feasible, because most of the
4265 work necessary to deduce information about a private key may be performed
4266 offline. The amount of resources required for such an attack would be
4267 significant. However, for an attack on TLS to be meaningful, the server
4268 would have to share the DH1024 private key among multiple clients, which is
4269 no longer an option since CVE-2016-0701.
4271 This only affects processors that support the AVX2 but not ADX extensions
4272 like Intel Haswell (4th generation).
4274 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4275 was originally found via the OSS-Fuzz project.
4280 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4282 * bn_sqrx8x_internal carry bug on x86_64
4284 There is a carry propagating bug in the x86_64 Montgomery squaring
4285 procedure. No EC algorithms are affected. Analysis suggests that attacks
4286 against RSA and DSA as a result of this defect would be very difficult to
4287 perform and are not believed likely. Attacks against DH are considered just
4288 feasible (although very difficult) because most of the work necessary to
4289 deduce information about a private key may be performed offline. The amount
4290 of resources required for such an attack would be very significant and
4291 likely only accessible to a limited number of attackers. An attacker would
4292 additionally need online access to an unpatched system using the target
4293 private key in a scenario with persistent DH parameters and a private
4294 key that is shared between multiple clients.
4296 This only affects processors that support the BMI1, BMI2 and ADX extensions
4297 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4299 This issue was reported to OpenSSL by the OSS-Fuzz project.
4304 * Malformed X.509 IPAddressFamily could cause OOB read
4306 If an X.509 certificate has a malformed IPAddressFamily extension,
4307 OpenSSL could do a one-byte buffer overread. The most likely result
4308 would be an erroneous display of the certificate in text format.
4310 This issue was reported to OpenSSL by the OSS-Fuzz project.
4315 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4317 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4318 platform rather than 'mingw'.
4322 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4324 * Truncated packet could crash via OOB read
4326 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4327 cipher is being used, then a truncated packet can cause that host to
4328 perform an out-of-bounds read, usually resulting in a crash.
4330 This issue was reported to OpenSSL by Robert Święcki of Google.
4335 * BN_mod_exp may produce incorrect results on x86_64
4337 There is a carry propagating bug in the x86_64 Montgomery squaring
4338 procedure. No EC algorithms are affected. Analysis suggests that attacks
4339 against RSA and DSA as a result of this defect would be very difficult to
4340 perform and are not believed likely. Attacks against DH are considered just
4341 feasible (although very difficult) because most of the work necessary to
4342 deduce information about a private key may be performed offline. The amount
4343 of resources required for such an attack would be very significant and
4344 likely only accessible to a limited number of attackers. An attacker would
4345 additionally need online access to an unpatched system using the target
4346 private key in a scenario with persistent DH parameters and a private
4347 key that is shared between multiple clients. For example this can occur by
4348 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4349 similar to CVE-2015-3193 but must be treated as a separate problem.
4351 This issue was reported to OpenSSL by the OSS-Fuzz project.
4356 * Montgomery multiplication may produce incorrect results
4358 There is a carry propagating bug in the Broadwell-specific Montgomery
4359 multiplication procedure that handles input lengths divisible by, but
4360 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4361 and DH private keys are impossible. This is because the subroutine in
4362 question is not used in operations with the private key itself and an input
4363 of the attacker's direct choice. Otherwise the bug can manifest itself as
4364 transient authentication and key negotiation failures or reproducible
4365 erroneous outcome of public-key operations with specially crafted input.
4366 Among EC algorithms only Brainpool P-512 curves are affected and one
4367 presumably can attack ECDH key negotiation. Impact was not analyzed in
4368 detail, because pre-requisites for attack are considered unlikely. Namely
4369 multiple clients have to choose the curve in question and the server has to
4370 share the private key among them, neither of which is default behaviour.
4371 Even then only clients that chose the curve will be affected.
4373 This issue was publicly reported as transient failures and was not
4374 initially recognized as a security issue. Thanks to Richard Morgan for
4375 providing reproducible case.
4380 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4381 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4382 prevent issues where no progress is being made and the peer continually
4383 sends unrecognised record types, using up resources processing them.
4387 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4389 * Missing CRL sanity check
4391 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4392 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4393 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4395 This issue only affects the OpenSSL 1.0.2i
4400 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4402 * OCSP Status Request extension unbounded memory growth
4404 A malicious client can send an excessively large OCSP Status Request
4405 extension. If that client continually requests renegotiation, sending a
4406 large OCSP Status Request extension each time, then there will be unbounded
4407 memory growth on the server. This will eventually lead to a Denial Of
4408 Service attack through memory exhaustion. Servers with a default
4409 configuration are vulnerable even if they do not support OCSP. Builds using
4410 the "no-ocsp" build time option are not affected.
4412 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4417 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4420 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4426 * OOB write in MDC2_Update()
4428 An overflow can occur in MDC2_Update() either if called directly or
4429 through the EVP_DigestUpdate() function using MDC2. If an attacker
4430 is able to supply very large amounts of input data after a previous
4431 call to EVP_EncryptUpdate() with a partial block then a length check
4432 can overflow resulting in a heap corruption.
4434 The amount of data needed is comparable to SIZE_MAX which is impractical
4437 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4442 * Malformed SHA512 ticket DoS
4444 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4445 DoS attack where a malformed ticket will result in an OOB read which will
4448 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4449 a custom server callback and ticket lookup mechanism.
4451 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4456 * OOB write in BN_bn2dec()
4458 The function BN_bn2dec() does not check the return value of BN_div_word().
4459 This can cause an OOB write if an application uses this function with an
4460 overly large BIGNUM. This could be a problem if an overly large certificate
4461 or CRL is printed out from an untrusted source. TLS is not affected because
4462 record limits will reject an oversized certificate before it is parsed.
4464 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4469 * OOB read in TS_OBJ_print_bio()
4471 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4472 the total length the OID text representation would use and not the amount
4473 of data written. This will result in OOB reads when large OIDs are
4476 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4481 * Pointer arithmetic undefined behaviour
4483 Avoid some undefined pointer arithmetic
4485 A common idiom in the codebase is to check limits in the following manner:
4488 Where "p" points to some malloc'd data of SIZE bytes and
4491 "len" here could be from some externally supplied data (e.g. from a TLS
4494 The rules of C pointer arithmetic are such that "p + len" is only well
4495 defined where len <= SIZE. Therefore the above idiom is actually
4496 undefined behaviour.
4498 For example this could cause problems if some malloc implementation
4499 provides an address for "p" such that "p + len" actually overflows for
4500 values of len that are too big and therefore p + len < limit.
4502 This issue was reported to OpenSSL by Guido Vranken
4507 * Constant time flag not preserved in DSA signing
4509 Operations in the DSA signing algorithm should run in constant time in
4510 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4511 implementation means that a non-constant time codepath is followed for
4512 certain operations. This has been demonstrated through a cache-timing
4513 attack to be sufficient for an attacker to recover the private DSA key.
4515 This issue was reported by César Pereida (Aalto University), Billy Brumley
4516 (Tampere University of Technology), and Yuval Yarom (The University of
4517 Adelaide and NICTA).
4522 * DTLS buffered message DoS
4524 In a DTLS connection where handshake messages are delivered out-of-order
4525 those messages that OpenSSL is not yet ready to process will be buffered
4526 for later use. Under certain circumstances, a flaw in the logic means that
4527 those messages do not get removed from the buffer even though the handshake
4528 has been completed. An attacker could force up to approx. 15 messages to
4529 remain in the buffer when they are no longer required. These messages will
4530 be cleared when the DTLS connection is closed. The default maximum size for
4531 a message is 100k. Therefore the attacker could force an additional 1500k
4532 to be consumed per connection. By opening many simulataneous connections an
4533 attacker could cause a DoS attack through memory exhaustion.
4535 This issue was reported to OpenSSL by Quan Luo.
4540 * DTLS replay protection DoS
4542 A flaw in the DTLS replay attack protection mechanism means that records
4543 that arrive for future epochs update the replay protection "window" before
4544 the MAC for the record has been validated. This could be exploited by an
4545 attacker by sending a record for the next epoch (which does not have to
4546 decrypt or have a valid MAC), with a very large sequence number. This means
4547 that all subsequent legitimate packets are dropped causing a denial of
4548 service for a specific DTLS connection.
4550 This issue was reported to OpenSSL by the OCAP audit team.
4555 * Certificate message OOB reads
4557 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4558 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4559 theoretical DoS risk but this has not been observed in practice on common
4562 The messages affected are client certificate, client certificate request
4563 and server certificate. As a result the attack can only be performed
4564 against a client or a server which enables client authentication.
4566 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4571 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4573 * Prevent padding oracle in AES-NI CBC MAC check
4575 A MITM attacker can use a padding oracle attack to decrypt traffic
4576 when the connection uses an AES CBC cipher and the server support
4579 This issue was introduced as part of the fix for Lucky 13 padding
4580 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4581 constant time by making sure that always the same bytes are read and
4582 compared against either the MAC or padding bytes. But it no longer
4583 checked that there was enough data to have both the MAC and padding
4586 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4591 * Fix EVP_EncodeUpdate overflow
4593 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4594 Base64 encoding of binary data. If an attacker is able to supply very large
4595 amounts of input data then a length check can overflow resulting in a heap
4598 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4599 the `PEM_write_bio*` family of functions. These are mainly used within the
4600 OpenSSL command line applications, so any application which processes data
4601 from an untrusted source and outputs it as a PEM file should be considered
4602 vulnerable to this issue. User applications that call these APIs directly
4603 with large amounts of untrusted data may also be vulnerable.
4605 This issue was reported by Guido Vranken.
4610 * Fix EVP_EncryptUpdate overflow
4612 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4613 is able to supply very large amounts of input data after a previous call to
4614 EVP_EncryptUpdate() with a partial block then a length check can overflow
4615 resulting in a heap corruption. Following an analysis of all OpenSSL
4616 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4617 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4618 the first called function after an EVP_EncryptInit(), and therefore that
4619 specific call must be safe. The second form is where the length passed to
4620 EVP_EncryptUpdate() can be seen from the code to be some small value and
4621 therefore there is no possibility of an overflow. Since all instances are
4622 one of these two forms, it is believed that there can be no overflows in
4623 internal code due to this problem. It should be noted that
4624 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4625 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4626 of these calls have also been analysed too and it is believed there are no
4627 instances in internal usage where an overflow could occur.
4629 This issue was reported by Guido Vranken.
4634 * Prevent ASN.1 BIO excessive memory allocation
4636 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4637 a short invalid encoding can cause allocation of large amounts of memory
4638 potentially consuming excessive resources or exhausting memory.
4640 Any application parsing untrusted data through d2i BIO functions is
4641 affected. The memory based functions such as d2i_X509() are *not* affected.
4642 Since the memory based functions are used by the TLS library, TLS
4643 applications are not affected.
4645 This issue was reported by Brian Carpenter.
4652 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4653 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4654 in arbitrary stack data being returned in the buffer.
4656 This issue was reported by Guido Vranken.
4661 * Modify behavior of ALPN to invoke callback after SNI/servername
4662 callback, such that updates to the SSL_CTX affect ALPN.
4666 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4671 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4672 methods are enabled and ssl2 is disabled the methods return NULL.
4676 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4678 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4679 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4680 provide any "EXPORT" or "LOW" strength ciphers.
4684 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4685 is by default disabled at build-time. Builds that are not configured with
4686 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4687 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4688 will need to explicitly call either of:
4690 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4692 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4694 as appropriate. Even if either of those is used, or the application
4695 explicitly uses the version-specific SSLv2_method() or its client and
4696 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4697 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4698 ciphers, and SSLv2 56-bit DES are no longer available.
4703 * Fix a double-free in DSA code
4705 A double free bug was discovered when OpenSSL parses malformed DSA private
4706 keys and could lead to a DoS attack or memory corruption for applications
4707 that receive DSA private keys from untrusted sources. This scenario is
4710 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4716 * Disable SRP fake user seed to address a server memory leak.
4718 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4720 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4721 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4722 was changed to ignore the "fake user" SRP seed, even if the seed
4725 Users should use SRP_VBASE_get1_by_user instead. Note that in
4726 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4727 also that even though configuring the SRP seed attempts to hide
4728 invalid usernames by continuing the handshake with fake
4729 credentials, this behaviour is not constant time and no strong
4730 guarantees are made that the handshake is indistinguishable from
4731 that of a valid user.
4736 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4738 In the BN_hex2bn function the number of hex digits is calculated using an
4739 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4740 large values of `i` this can result in `bn_expand` not allocating any
4741 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4742 field as NULL leading to a subsequent NULL ptr deref. For very large values
4743 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4744 In this case memory is allocated to the internal BIGNUM data field, but it
4745 is insufficiently sized leading to heap corruption. A similar issue exists
4746 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4747 is ever called by user applications with very large untrusted hex/dec data.
4748 This is anticipated to be a rare occurrence.
4750 All OpenSSL internal usage of these functions use data that is not expected
4751 to be untrusted, e.g. config file data or application command line
4752 arguments. If user developed applications generate config file data based
4753 on untrusted data then it is possible that this could also lead to security
4754 consequences. This is also anticipated to be rare.
4756 This issue was reported to OpenSSL by Guido Vranken.
4761 * Fix memory issues in `BIO_*printf` functions
4763 The internal `fmtstr` function used in processing a "%s" format string in
4764 the `BIO_*printf` functions could overflow while calculating the length of a
4765 string and cause an OOB read when printing very long strings.
4767 Additionally the internal `doapr_outch` function can attempt to write to an
4768 OOB memory location (at an offset from the NULL pointer) in the event of a
4769 memory allocation failure. In 1.0.2 and below this could be caused where
4770 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4771 could be in processing a very long "%s" format string. Memory leaks can
4774 The first issue may mask the second issue dependent on compiler behaviour.
4775 These problems could enable attacks where large amounts of untrusted data
4776 is passed to the `BIO_*printf` functions. If applications use these functions
4777 in this way then they could be vulnerable. OpenSSL itself uses these
4778 functions when printing out human-readable dumps of ASN.1 data. Therefore
4779 applications that print this data could be vulnerable if the data is from
4780 untrusted sources. OpenSSL command line applications could also be
4781 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4782 as command line arguments.
4784 Libssl is not considered directly vulnerable. Additionally certificates etc
4785 received via remote connections via libssl are also unlikely to be able to
4786 trigger these issues because of message size limits enforced within libssl.
4788 This issue was reported to OpenSSL Guido Vranken.
4793 * Side channel attack on modular exponentiation
4795 A side-channel attack was found which makes use of cache-bank conflicts on
4796 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4797 of RSA keys. The ability to exploit this issue is limited as it relies on
4798 an attacker who has control of code in a thread running on the same
4799 hyper-threaded core as the victim thread which is performing decryptions.
4801 This issue was reported to OpenSSL by Yuval Yarom, The University of
4802 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4803 Nadia Heninger, University of Pennsylvania with more information at
4804 <http://cachebleed.info>.
4809 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4810 if no keysize is specified with default_bits. This fixes an
4811 omission in an earlier change that changed all RSA/DSA key generation
4812 apps to use 2048 bits by default.
4816 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4818 * DH small subgroups
4820 Historically OpenSSL only ever generated DH parameters based on "safe"
4821 primes. More recently (in version 1.0.2) support was provided for
4822 generating X9.42 style parameter files such as those required for RFC 5114
4823 support. The primes used in such files may not be "safe". Where an
4824 application is using DH configured with parameters based on primes that are
4825 not "safe" then an attacker could use this fact to find a peer's private
4826 DH exponent. This attack requires that the attacker complete multiple
4827 handshakes in which the peer uses the same private DH exponent. For example
4828 this could be used to discover a TLS server's private DH exponent if it's
4829 reusing the private DH exponent or it's using a static DH ciphersuite.
4831 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4832 TLS. It is not on by default. If the option is not set then the server
4833 reuses the same private DH exponent for the life of the server process and
4834 would be vulnerable to this attack. It is believed that many popular
4835 applications do set this option and would therefore not be at risk.
4837 The fix for this issue adds an additional check where a "q" parameter is
4838 available (as is the case in X9.42 based parameters). This detects the
4839 only known attack, and is the only possible defense for static DH
4840 ciphersuites. This could have some performance impact.
4842 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4843 default and cannot be disabled. This could have some performance impact.
4845 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4850 * SSLv2 doesn't block disabled ciphers
4852 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4853 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4854 been disabled, provided that the SSLv2 protocol was not also disabled via
4857 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4858 and Sebastian Schinzel.
4863 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4865 * BN_mod_exp may produce incorrect results on x86_64
4867 There is a carry propagating bug in the x86_64 Montgomery squaring
4868 procedure. No EC algorithms are affected. Analysis suggests that attacks
4869 against RSA and DSA as a result of this defect would be very difficult to
4870 perform and are not believed likely. Attacks against DH are considered just
4871 feasible (although very difficult) because most of the work necessary to
4872 deduce information about a private key may be performed offline. The amount
4873 of resources required for such an attack would be very significant and
4874 likely only accessible to a limited number of attackers. An attacker would
4875 additionally need online access to an unpatched system using the target
4876 private key in a scenario with persistent DH parameters and a private
4877 key that is shared between multiple clients. For example this can occur by
4878 default in OpenSSL DHE based SSL/TLS ciphersuites.
4880 This issue was reported to OpenSSL by Hanno Böck.
4885 * Certificate verify crash with missing PSS parameter
4887 The signature verification routines will crash with a NULL pointer
4888 dereference if presented with an ASN.1 signature using the RSA PSS
4889 algorithm and absent mask generation function parameter. Since these
4890 routines are used to verify certificate signature algorithms this can be
4891 used to crash any certificate verification operation and exploited in a
4892 DoS attack. Any application which performs certificate verification is
4893 vulnerable including OpenSSL clients and servers which enable client
4896 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4901 * X509_ATTRIBUTE memory leak
4903 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4904 memory. This structure is used by the PKCS#7 and CMS routines so any
4905 application which reads PKCS#7 or CMS data from untrusted sources is
4906 affected. SSL/TLS is not affected.
4908 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4914 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4915 This changes the decoding behaviour for some invalid messages,
4916 though the change is mostly in the more lenient direction, and
4917 legacy behaviour is preserved as much as possible.
4921 * In DSA_generate_parameters_ex, if the provided seed is too short,
4924 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4926 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4928 * Alternate chains certificate forgery
4930 During certificate verification, OpenSSL will attempt to find an
4931 alternative certificate chain if the first attempt to build such a chain
4932 fails. An error in the implementation of this logic can mean that an
4933 attacker could cause certain checks on untrusted certificates to be
4934 bypassed, such as the CA flag, enabling them to use a valid leaf
4935 certificate to act as a CA and "issue" an invalid certificate.
4937 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4942 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4944 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4945 incompatibility in the handling of HMAC. The previous ABI has now been
4950 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4952 * Malformed ECParameters causes infinite loop
4954 When processing an ECParameters structure OpenSSL enters an infinite loop
4955 if the curve specified is over a specially malformed binary polynomial
4958 This can be used to perform denial of service against any
4959 system which processes public keys, certificate requests or
4960 certificates. This includes TLS clients and TLS servers with
4961 client authentication enabled.
4963 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4968 * Exploitable out-of-bounds read in X509_cmp_time
4970 X509_cmp_time does not properly check the length of the ASN1_TIME
4971 string and can read a few bytes out of bounds. In addition,
4972 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4975 An attacker can use this to craft malformed certificates and CRLs of
4976 various sizes and potentially cause a segmentation fault, resulting in
4977 a DoS on applications that verify certificates or CRLs. TLS clients
4978 that verify CRLs are affected. TLS clients and servers with client
4979 authentication enabled may be affected if they use custom verification
4982 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4983 independently by Hanno Böck.
4988 * PKCS7 crash with missing EnvelopedContent
4990 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4991 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4992 with missing content and trigger a NULL pointer dereference on parsing.
4994 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4995 structures from untrusted sources are affected. OpenSSL clients and
4996 servers are not affected.
4998 This issue was reported to OpenSSL by Michal Zalewski (Google).
5003 * CMS verify infinite loop with unknown hash function
5005 When verifying a signedData message the CMS code can enter an infinite loop
5006 if presented with an unknown hash function OID. This can be used to perform
5007 denial of service against any system which verifies signedData messages using
5009 This issue was reported to OpenSSL by Johannes Bauer.
5014 * Race condition handling NewSessionTicket
5016 If a NewSessionTicket is received by a multi-threaded client when attempting to
5017 reuse a previous ticket then a race condition can occur potentially leading to
5018 a double free of the ticket data.
5023 * Only support 256-bit or stronger elliptic curves with the
5024 'ecdh_auto' setting (server) or by default (client). Of supported
5025 curves, prefer P-256 (both).
5029 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5031 * ClientHello sigalgs DoS fix
5033 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5034 invalid signature algorithms extension a NULL pointer dereference will
5035 occur. This can be exploited in a DoS attack against the server.
5037 This issue was was reported to OpenSSL by David Ramos of Stanford
5041 *Stephen Henson and Matt Caswell*
5043 * Multiblock corrupted pointer fix
5045 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5046 feature only applies on 64 bit x86 architecture platforms that support AES
5047 NI instructions. A defect in the implementation of "multiblock" can cause
5048 OpenSSL's internal write buffer to become incorrectly set to NULL when
5049 using non-blocking IO. Typically, when the user application is using a
5050 socket BIO for writing, this will only result in a failed connection.
5051 However if some other BIO is used then it is likely that a segmentation
5052 fault will be triggered, thus enabling a potential DoS attack.
5054 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5059 * Segmentation fault in DTLSv1_listen fix
5061 The DTLSv1_listen function is intended to be stateless and processes the
5062 initial ClientHello from many peers. It is common for user code to loop
5063 over the call to DTLSv1_listen until a valid ClientHello is received with
5064 an associated cookie. A defect in the implementation of DTLSv1_listen means
5065 that state is preserved in the SSL object from one invocation to the next
5066 that can lead to a segmentation fault. Errors processing the initial
5067 ClientHello can trigger this scenario. An example of such an error could be
5068 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5071 This issue was reported to OpenSSL by Per Allansson.
5076 * Segmentation fault in ASN1_TYPE_cmp fix
5078 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5079 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5080 certificate signature algorithm consistency this can be used to crash any
5081 certificate verification operation and exploited in a DoS attack. Any
5082 application which performs certificate verification is vulnerable including
5083 OpenSSL clients and servers which enable client authentication.
5088 * Segmentation fault for invalid PSS parameters fix
5090 The signature verification routines will crash with a NULL pointer
5091 dereference if presented with an ASN.1 signature using the RSA PSS
5092 algorithm and invalid parameters. Since these routines are used to verify
5093 certificate signature algorithms this can be used to crash any
5094 certificate verification operation and exploited in a DoS attack. Any
5095 application which performs certificate verification is vulnerable including
5096 OpenSSL clients and servers which enable client authentication.
5098 This issue was was reported to OpenSSL by Brian Carpenter.
5103 * ASN.1 structure reuse memory corruption fix
5105 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5106 memory corruption via an invalid write. Such reuse is and has been
5107 strongly discouraged and is believed to be rare.
5109 Applications that parse structures containing CHOICE or ANY DEFINED BY
5110 components may be affected. Certificate parsing (d2i_X509 and related
5111 functions) are however not affected. OpenSSL clients and servers are
5117 * PKCS7 NULL pointer dereferences fix
5119 The PKCS#7 parsing code does not handle missing outer ContentInfo
5120 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5121 missing content and trigger a NULL pointer dereference on parsing.
5123 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5124 otherwise parse PKCS#7 structures from untrusted sources are
5125 affected. OpenSSL clients and servers are not affected.
5127 This issue was reported to OpenSSL by Michal Zalewski (Google).
5132 * DoS via reachable assert in SSLv2 servers fix
5134 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5135 servers that both support SSLv2 and enable export cipher suites by sending
5136 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5138 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5139 (OpenSSL development team).
5144 * Empty CKE with client auth and DHE fix
5146 If client auth is used then a server can seg fault in the event of a DHE
5147 ciphersuite being selected and a zero length ClientKeyExchange message
5148 being sent by the client. This could be exploited in a DoS attack.
5153 * Handshake with unseeded PRNG fix
5155 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5156 with an unseeded PRNG. The conditions are:
5157 - The client is on a platform where the PRNG has not been seeded
5158 automatically, and the user has not seeded manually
5159 - A protocol specific client method version has been used (i.e. not
5160 SSL_client_methodv23)
5161 - A ciphersuite is used that does not require additional random data from
5162 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5164 If the handshake succeeds then the client random that has been used will
5165 have been generated from a PRNG with insufficient entropy and therefore the
5166 output may be predictable.
5168 For example using the following command with an unseeded openssl will
5169 succeed on an unpatched platform:
5171 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5176 * Use After Free following d2i_ECPrivatekey error fix
5178 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5179 could cause a use after free condition. This, in turn, could cause a double
5180 free in several private key parsing functions (such as d2i_PrivateKey
5181 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5182 for applications that receive EC private keys from untrusted
5183 sources. This scenario is considered rare.
5185 This issue was discovered by the BoringSSL project and fixed in their
5191 * X509_to_X509_REQ NULL pointer deref fix
5193 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5194 the certificate key is invalid. This function is rarely used in practice.
5196 This issue was discovered by Brian Carpenter.
5201 * Removed the export ciphers from the DEFAULT ciphers
5205 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5207 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5208 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5209 So far those who have to target multiple platforms would compromise
5210 and argue that binary targeting say ARMv5 would still execute on
5211 ARMv8. "Universal" build resolves this compromise by providing
5212 near-optimal performance even on newer platforms.
5216 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5217 (other platforms pending).
5219 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5221 * Add support for the SignedCertificateTimestampList certificate and
5222 OCSP response extensions from RFC6962.
5226 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5227 for corner cases. (Certain input points at infinity could lead to
5228 bogus results, with non-infinity inputs mapped to infinity too.)
5232 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5233 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5234 common cases are optimized and there still is room for further
5235 improvements. Vector Permutation AES for Altivec is also added.
5239 * Add support for little-endian ppc64 Linux target.
5241 *Marcelo Cerri (IBM)*
5243 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5244 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5245 are optimized and there still is room for further improvements.
5246 Both 32- and 64-bit modes are supported.
5248 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5250 * Improved ARMv7 NEON support.
5254 * Support for SPARC Architecture 2011 crypto extensions, first
5255 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5256 SHA256/512, MD5, GHASH and modular exponentiation.
5258 *Andy Polyakov, David Miller*
5260 * Accelerated modular exponentiation for Intel processors, a.k.a.
5263 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5265 * Support for new and upcoming Intel processors, including AVX2,
5266 BMI and SHA ISA extensions. This includes additional "stitched"
5267 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5270 This work was sponsored by Intel Corp.
5274 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5275 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5276 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5280 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5281 this fixes a limitation in previous versions of OpenSSL.
5285 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5286 MGF1 digest and OAEP label.
5290 * Add EVP support for key wrapping algorithms, to avoid problems with
5291 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5292 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5293 algorithms and include tests cases.
5297 * Add functions to allocate and set the fields of an ECDSA_METHOD
5300 *Douglas E. Engert, Steve Henson*
5302 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5303 difference in days and seconds between two tm or ASN1_TIME structures.
5307 * Add -rev test option to s_server to just reverse order of characters
5308 received by client and send back to server. Also prints an abbreviated
5309 summary of the connection parameters.
5313 * New option -brief for s_client and s_server to print out a brief summary
5314 of connection parameters.
5318 * Add callbacks for arbitrary TLS extensions.
5320 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5322 * New option -crl_download in several openssl utilities to download CRLs
5323 from CRLDP extension in certificates.
5327 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5331 * New function X509_CRL_diff to generate a delta CRL from the difference
5332 of two full CRLs. Add support to "crl" utility.
5336 * New functions to set lookup_crls function and to retrieve
5337 X509_STORE from X509_STORE_CTX.
5341 * Print out deprecated issuer and subject unique ID fields in
5346 * Extend OCSP I/O functions so they can be used for simple general purpose
5347 HTTP as well as OCSP. New wrapper function which can be used to download
5348 CRLs using the OCSP API.
5352 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5356 * `SSL_CONF*` functions. These provide a common framework for application
5357 configuration using configuration files or command lines.
5361 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5362 message callback and prints the results. Needs compile time option
5363 "enable-ssl-trace". New options to s_client and s_server to enable
5368 * New ctrl and macro to retrieve supported points extensions.
5369 Print out extension in s_server and s_client.
5373 * New functions to retrieve certificate signature and signature
5378 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5383 * New Suite B modes for TLS code. These use and enforce the requirements
5384 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5385 only use Suite B curves. The Suite B modes can be set by using the
5386 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5390 * New chain verification flags for Suite B levels of security. Check
5391 algorithms are acceptable when flags are set in X509_verify_cert.
5395 * Make tls1_check_chain return a set of flags indicating checks passed
5396 by a certificate chain. Add additional tests to handle client
5397 certificates: checks for matching certificate type and issuer name
5402 * If an attempt is made to use a signature algorithm not in the peer
5403 preference list abort the handshake. If client has no suitable
5404 signature algorithms in response to a certificate request do not
5405 use the certificate.
5409 * If server EC tmp key is not in client preference list abort handshake.
5413 * Add support for certificate stores in CERT structure. This makes it
5414 possible to have different stores per SSL structure or one store in
5415 the parent SSL_CTX. Include distinct stores for certificate chain
5416 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5417 to build and store a certificate chain in CERT structure: returning
5418 an error if the chain cannot be built: this will allow applications
5419 to test if a chain is correctly configured.
5421 Note: if the CERT based stores are not set then the parent SSL_CTX
5422 store is used to retain compatibility with existing behaviour.
5426 * New function ssl_set_client_disabled to set a ciphersuite disabled
5427 mask based on the current session, check mask when sending client
5428 hello and checking the requested ciphersuite.
5432 * New ctrls to retrieve and set certificate types in a certificate
5433 request message. Print out received values in s_client. If certificate
5434 types is not set with custom values set sensible values based on
5435 supported signature algorithms.
5439 * Support for distinct client and server supported signature algorithms.
5443 * Add certificate callback. If set this is called whenever a certificate
5444 is required by client or server. An application can decide which
5445 certificate chain to present based on arbitrary criteria: for example
5446 supported signature algorithms. Add very simple example to s_server.
5447 This fixes many of the problems and restrictions of the existing client
5448 certificate callback: for example you can now clear an existing
5449 certificate and specify the whole chain.
5453 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5454 the certificate can be used for (if anything). Set valid_flags field
5455 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5456 to have similar checks in it.
5458 Add new "cert_flags" field to CERT structure and include a "strict mode".
5459 This enforces some TLS certificate requirements (such as only permitting
5460 certificate signature algorithms contained in the supported algorithms
5461 extension) which some implementations ignore: this option should be used
5462 with caution as it could cause interoperability issues.
5466 * Update and tidy signature algorithm extension processing. Work out
5467 shared signature algorithms based on preferences and peer algorithms
5468 and print them out in s_client and s_server. Abort handshake if no
5469 shared signature algorithms.
5473 * Add new functions to allow customised supported signature algorithms
5474 for SSL and SSL_CTX structures. Add options to s_client and s_server
5479 * New function SSL_certs_clear() to delete all references to certificates
5480 from an SSL structure. Before this once a certificate had been added
5481 it couldn't be removed.
5485 * Integrate hostname, email address and IP address checking with certificate
5486 verification. New verify options supporting checking in openssl utility.
5490 * Fixes and wildcard matching support to hostname and email checking
5491 functions. Add manual page.
5493 *Florian Weimer (Red Hat Product Security Team)*
5495 * New functions to check a hostname email or IP address against a
5496 certificate. Add options x509 utility to print results of checks against
5501 * Fix OCSP checking.
5503 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5505 * Initial experimental support for explicitly trusted non-root CAs.
5506 OpenSSL still tries to build a complete chain to a root but if an
5507 intermediate CA has a trust setting included that is used. The first
5508 setting is used: whether to trust (e.g., -addtrust option to the x509
5513 * Add -trusted_first option which attempts to find certificates in the
5514 trusted store even if an untrusted chain is also supplied.
5518 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5519 platform support for Linux and Android.
5523 * Support for linux-x32, ILP32 environment in x86_64 framework.
5527 * Experimental multi-implementation support for FIPS capable OpenSSL.
5528 When in FIPS mode the approved implementations are used as normal,
5529 when not in FIPS mode the internal unapproved versions are used instead.
5530 This means that the FIPS capable OpenSSL isn't forced to use the
5531 (often lower performance) FIPS implementations outside FIPS mode.
5535 * Transparently support X9.42 DH parameters when calling
5536 PEM_read_bio_DHparameters. This means existing applications can handle
5537 the new parameter format automatically.
5541 * Initial experimental support for X9.42 DH parameter format: mainly
5542 to support use of 'q' parameter for RFC5114 parameters.
5546 * Add DH parameters from RFC5114 including test data to dhtest.
5550 * Support for automatic EC temporary key parameter selection. If enabled
5551 the most preferred EC parameters are automatically used instead of
5552 hardcoded fixed parameters. Now a server just has to call:
5553 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5554 support ECDH and use the most appropriate parameters.
5558 * Enhance and tidy EC curve and point format TLS extension code. Use
5559 static structures instead of allocation if default values are used.
5560 New ctrls to set curves we wish to support and to retrieve shared curves.
5561 Print out shared curves in s_server. New options to s_server and s_client
5562 to set list of supported curves.
5566 * New ctrls to retrieve supported signature algorithms and
5567 supported curve values as an array of NIDs. Extend openssl utility
5568 to print out received values.
5572 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5573 between NIDs and the more common NIST names such as "P-256". Enhance
5574 ecparam utility and ECC method to recognise the NIST names for curves.
5578 * Enhance SSL/TLS certificate chain handling to support different
5579 chains for each certificate instead of one chain in the parent SSL_CTX.
5583 * Support for fixed DH ciphersuite client authentication: where both
5584 server and client use DH certificates with common parameters.
5588 * Support for fixed DH ciphersuites: those requiring DH server
5593 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5595 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5596 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5597 X509_CINF_get_signature were reverted post internal team review.
5602 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5604 * OCSP Status Request extension unbounded memory growth
5606 A malicious client can send an excessively large OCSP Status Request
5607 extension. If that client continually requests renegotiation, sending a
5608 large OCSP Status Request extension each time, then there will be unbounded
5609 memory growth on the server. This will eventually lead to a Denial Of
5610 Service attack through memory exhaustion. Servers with a default
5611 configuration are vulnerable even if they do not support OCSP. Builds using
5612 the "no-ocsp" build time option are not affected.
5614 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5619 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5622 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5628 * OOB write in MDC2_Update()
5630 An overflow can occur in MDC2_Update() either if called directly or
5631 through the EVP_DigestUpdate() function using MDC2. If an attacker
5632 is able to supply very large amounts of input data after a previous
5633 call to EVP_EncryptUpdate() with a partial block then a length check
5634 can overflow resulting in a heap corruption.
5636 The amount of data needed is comparable to SIZE_MAX which is impractical
5639 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5644 * Malformed SHA512 ticket DoS
5646 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5647 DoS attack where a malformed ticket will result in an OOB read which will
5650 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5651 a custom server callback and ticket lookup mechanism.
5653 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5658 * OOB write in BN_bn2dec()
5660 The function BN_bn2dec() does not check the return value of BN_div_word().
5661 This can cause an OOB write if an application uses this function with an
5662 overly large BIGNUM. This could be a problem if an overly large certificate
5663 or CRL is printed out from an untrusted source. TLS is not affected because
5664 record limits will reject an oversized certificate before it is parsed.
5666 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5671 * OOB read in TS_OBJ_print_bio()
5673 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5674 the total length the OID text representation would use and not the amount
5675 of data written. This will result in OOB reads when large OIDs are
5678 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5683 * Pointer arithmetic undefined behaviour
5685 Avoid some undefined pointer arithmetic
5687 A common idiom in the codebase is to check limits in the following manner:
5690 Where "p" points to some malloc'd data of SIZE bytes and
5693 "len" here could be from some externally supplied data (e.g. from a TLS
5696 The rules of C pointer arithmetic are such that "p + len" is only well
5697 defined where len <= SIZE. Therefore the above idiom is actually
5698 undefined behaviour.
5700 For example this could cause problems if some malloc implementation
5701 provides an address for "p" such that "p + len" actually overflows for
5702 values of len that are too big and therefore p + len < limit.
5704 This issue was reported to OpenSSL by Guido Vranken
5709 * Constant time flag not preserved in DSA signing
5711 Operations in the DSA signing algorithm should run in constant time in
5712 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5713 implementation means that a non-constant time codepath is followed for
5714 certain operations. This has been demonstrated through a cache-timing
5715 attack to be sufficient for an attacker to recover the private DSA key.
5717 This issue was reported by César Pereida (Aalto University), Billy Brumley
5718 (Tampere University of Technology), and Yuval Yarom (The University of
5719 Adelaide and NICTA).
5724 * DTLS buffered message DoS
5726 In a DTLS connection where handshake messages are delivered out-of-order
5727 those messages that OpenSSL is not yet ready to process will be buffered
5728 for later use. Under certain circumstances, a flaw in the logic means that
5729 those messages do not get removed from the buffer even though the handshake
5730 has been completed. An attacker could force up to approx. 15 messages to
5731 remain in the buffer when they are no longer required. These messages will
5732 be cleared when the DTLS connection is closed. The default maximum size for
5733 a message is 100k. Therefore the attacker could force an additional 1500k
5734 to be consumed per connection. By opening many simulataneous connections an
5735 attacker could cause a DoS attack through memory exhaustion.
5737 This issue was reported to OpenSSL by Quan Luo.
5742 * DTLS replay protection DoS
5744 A flaw in the DTLS replay attack protection mechanism means that records
5745 that arrive for future epochs update the replay protection "window" before
5746 the MAC for the record has been validated. This could be exploited by an
5747 attacker by sending a record for the next epoch (which does not have to
5748 decrypt or have a valid MAC), with a very large sequence number. This means
5749 that all subsequent legitimate packets are dropped causing a denial of
5750 service for a specific DTLS connection.
5752 This issue was reported to OpenSSL by the OCAP audit team.
5757 * Certificate message OOB reads
5759 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5760 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5761 theoretical DoS risk but this has not been observed in practice on common
5764 The messages affected are client certificate, client certificate request
5765 and server certificate. As a result the attack can only be performed
5766 against a client or a server which enables client authentication.
5768 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5773 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5775 * Prevent padding oracle in AES-NI CBC MAC check
5777 A MITM attacker can use a padding oracle attack to decrypt traffic
5778 when the connection uses an AES CBC cipher and the server support
5781 This issue was introduced as part of the fix for Lucky 13 padding
5782 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5783 constant time by making sure that always the same bytes are read and
5784 compared against either the MAC or padding bytes. But it no longer
5785 checked that there was enough data to have both the MAC and padding
5788 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5793 * Fix EVP_EncodeUpdate overflow
5795 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5796 Base64 encoding of binary data. If an attacker is able to supply very large
5797 amounts of input data then a length check can overflow resulting in a heap
5800 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5801 the `PEM_write_bio*` family of functions. These are mainly used within the
5802 OpenSSL command line applications, so any application which processes data
5803 from an untrusted source and outputs it as a PEM file should be considered
5804 vulnerable to this issue. User applications that call these APIs directly
5805 with large amounts of untrusted data may also be vulnerable.
5807 This issue was reported by Guido Vranken.
5812 * Fix EVP_EncryptUpdate overflow
5814 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5815 is able to supply very large amounts of input data after a previous call to
5816 EVP_EncryptUpdate() with a partial block then a length check can overflow
5817 resulting in a heap corruption. Following an analysis of all OpenSSL
5818 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5819 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5820 the first called function after an EVP_EncryptInit(), and therefore that
5821 specific call must be safe. The second form is where the length passed to
5822 EVP_EncryptUpdate() can be seen from the code to be some small value and
5823 therefore there is no possibility of an overflow. Since all instances are
5824 one of these two forms, it is believed that there can be no overflows in
5825 internal code due to this problem. It should be noted that
5826 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5827 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5828 of these calls have also been analysed too and it is believed there are no
5829 instances in internal usage where an overflow could occur.
5831 This issue was reported by Guido Vranken.
5836 * Prevent ASN.1 BIO excessive memory allocation
5838 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5839 a short invalid encoding can casuse allocation of large amounts of memory
5840 potentially consuming excessive resources or exhausting memory.
5842 Any application parsing untrusted data through d2i BIO functions is
5843 affected. The memory based functions such as d2i_X509() are *not* affected.
5844 Since the memory based functions are used by the TLS library, TLS
5845 applications are not affected.
5847 This issue was reported by Brian Carpenter.
5854 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5855 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5856 in arbitrary stack data being returned in the buffer.
5858 This issue was reported by Guido Vranken.
5863 * Modify behavior of ALPN to invoke callback after SNI/servername
5864 callback, such that updates to the SSL_CTX affect ALPN.
5868 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5873 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5874 methods are enabled and ssl2 is disabled the methods return NULL.
5878 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5880 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5881 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5882 provide any "EXPORT" or "LOW" strength ciphers.
5886 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5887 is by default disabled at build-time. Builds that are not configured with
5888 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5889 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5890 will need to explicitly call either of:
5892 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5894 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5896 as appropriate. Even if either of those is used, or the application
5897 explicitly uses the version-specific SSLv2_method() or its client and
5898 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5899 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5900 ciphers, and SSLv2 56-bit DES are no longer available.
5905 * Fix a double-free in DSA code
5907 A double free bug was discovered when OpenSSL parses malformed DSA private
5908 keys and could lead to a DoS attack or memory corruption for applications
5909 that receive DSA private keys from untrusted sources. This scenario is
5912 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5918 * Disable SRP fake user seed to address a server memory leak.
5920 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5922 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5923 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5924 was changed to ignore the "fake user" SRP seed, even if the seed
5927 Users should use SRP_VBASE_get1_by_user instead. Note that in
5928 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5929 also that even though configuring the SRP seed attempts to hide
5930 invalid usernames by continuing the handshake with fake
5931 credentials, this behaviour is not constant time and no strong
5932 guarantees are made that the handshake is indistinguishable from
5933 that of a valid user.
5938 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5940 In the BN_hex2bn function the number of hex digits is calculated using an
5941 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5942 large values of `i` this can result in `bn_expand` not allocating any
5943 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5944 field as NULL leading to a subsequent NULL ptr deref. For very large values
5945 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5946 In this case memory is allocated to the internal BIGNUM data field, but it
5947 is insufficiently sized leading to heap corruption. A similar issue exists
5948 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5949 is ever called by user applications with very large untrusted hex/dec data.
5950 This is anticipated to be a rare occurrence.
5952 All OpenSSL internal usage of these functions use data that is not expected
5953 to be untrusted, e.g. config file data or application command line
5954 arguments. If user developed applications generate config file data based
5955 on untrusted data then it is possible that this could also lead to security
5956 consequences. This is also anticipated to be rare.
5958 This issue was reported to OpenSSL by Guido Vranken.
5963 * Fix memory issues in `BIO_*printf` functions
5965 The internal `fmtstr` function used in processing a "%s" format string in
5966 the `BIO_*printf` functions could overflow while calculating the length of a
5967 string and cause an OOB read when printing very long strings.
5969 Additionally the internal `doapr_outch` function can attempt to write to an
5970 OOB memory location (at an offset from the NULL pointer) in the event of a
5971 memory allocation failure. In 1.0.2 and below this could be caused where
5972 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5973 could be in processing a very long "%s" format string. Memory leaks can
5976 The first issue may mask the second issue dependent on compiler behaviour.
5977 These problems could enable attacks where large amounts of untrusted data
5978 is passed to the `BIO_*printf` functions. If applications use these functions
5979 in this way then they could be vulnerable. OpenSSL itself uses these
5980 functions when printing out human-readable dumps of ASN.1 data. Therefore
5981 applications that print this data could be vulnerable if the data is from
5982 untrusted sources. OpenSSL command line applications could also be
5983 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5984 as command line arguments.
5986 Libssl is not considered directly vulnerable. Additionally certificates etc
5987 received via remote connections via libssl are also unlikely to be able to
5988 trigger these issues because of message size limits enforced within libssl.
5990 This issue was reported to OpenSSL Guido Vranken.
5995 * Side channel attack on modular exponentiation
5997 A side-channel attack was found which makes use of cache-bank conflicts on
5998 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5999 of RSA keys. The ability to exploit this issue is limited as it relies on
6000 an attacker who has control of code in a thread running on the same
6001 hyper-threaded core as the victim thread which is performing decryptions.
6003 This issue was reported to OpenSSL by Yuval Yarom, The University of
6004 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6005 Nadia Heninger, University of Pennsylvania with more information at
6006 <http://cachebleed.info>.
6011 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6012 if no keysize is specified with default_bits. This fixes an
6013 omission in an earlier change that changed all RSA/DSA key generation
6014 apps to use 2048 bits by default.
6018 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6020 * Protection for DH small subgroup attacks
6022 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6023 switched on by default and cannot be disabled. This could have some
6028 * SSLv2 doesn't block disabled ciphers
6030 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6031 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6032 been disabled, provided that the SSLv2 protocol was not also disabled via
6035 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6036 and Sebastian Schinzel.
6041 * Reject DH handshakes with parameters shorter than 1024 bits.
6045 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6047 * Certificate verify crash with missing PSS parameter
6049 The signature verification routines will crash with a NULL pointer
6050 dereference if presented with an ASN.1 signature using the RSA PSS
6051 algorithm and absent mask generation function parameter. Since these
6052 routines are used to verify certificate signature algorithms this can be
6053 used to crash any certificate verification operation and exploited in a
6054 DoS attack. Any application which performs certificate verification is
6055 vulnerable including OpenSSL clients and servers which enable client
6058 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6063 * X509_ATTRIBUTE memory leak
6065 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6066 memory. This structure is used by the PKCS#7 and CMS routines so any
6067 application which reads PKCS#7 or CMS data from untrusted sources is
6068 affected. SSL/TLS is not affected.
6070 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6076 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6077 This changes the decoding behaviour for some invalid messages,
6078 though the change is mostly in the more lenient direction, and
6079 legacy behaviour is preserved as much as possible.
6083 * In DSA_generate_parameters_ex, if the provided seed is too short,
6084 use a random seed, as already documented.
6086 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6088 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6090 * Alternate chains certificate forgery
6092 During certificate verfification, OpenSSL will attempt to find an
6093 alternative certificate chain if the first attempt to build such a chain
6094 fails. An error in the implementation of this logic can mean that an
6095 attacker could cause certain checks on untrusted certificates to be
6096 bypassed, such as the CA flag, enabling them to use a valid leaf
6097 certificate to act as a CA and "issue" an invalid certificate.
6099 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6105 * Race condition handling PSK identify hint
6107 If PSK identity hints are received by a multi-threaded client then
6108 the values are wrongly updated in the parent SSL_CTX structure. This can
6109 result in a race condition potentially leading to a double free of the
6115 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6117 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6118 incompatibility in the handling of HMAC. The previous ABI has now been
6121 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6123 * Malformed ECParameters causes infinite loop
6125 When processing an ECParameters structure OpenSSL enters an infinite loop
6126 if the curve specified is over a specially malformed binary polynomial
6129 This can be used to perform denial of service against any
6130 system which processes public keys, certificate requests or
6131 certificates. This includes TLS clients and TLS servers with
6132 client authentication enabled.
6134 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6139 * Exploitable out-of-bounds read in X509_cmp_time
6141 X509_cmp_time does not properly check the length of the ASN1_TIME
6142 string and can read a few bytes out of bounds. In addition,
6143 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6146 An attacker can use this to craft malformed certificates and CRLs of
6147 various sizes and potentially cause a segmentation fault, resulting in
6148 a DoS on applications that verify certificates or CRLs. TLS clients
6149 that verify CRLs are affected. TLS clients and servers with client
6150 authentication enabled may be affected if they use custom verification
6153 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6154 independently by Hanno Böck.
6159 * PKCS7 crash with missing EnvelopedContent
6161 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6162 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6163 with missing content and trigger a NULL pointer dereference on parsing.
6165 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6166 structures from untrusted sources are affected. OpenSSL clients and
6167 servers are not affected.
6169 This issue was reported to OpenSSL by Michal Zalewski (Google).
6174 * CMS verify infinite loop with unknown hash function
6176 When verifying a signedData message the CMS code can enter an infinite loop
6177 if presented with an unknown hash function OID. This can be used to perform
6178 denial of service against any system which verifies signedData messages using
6180 This issue was reported to OpenSSL by Johannes Bauer.
6185 * Race condition handling NewSessionTicket
6187 If a NewSessionTicket is received by a multi-threaded client when attempting to
6188 reuse a previous ticket then a race condition can occur potentially leading to
6189 a double free of the ticket data.
6194 * Reject DH handshakes with parameters shorter than 768 bits.
6196 *Kurt Roeckx and Emilia Kasper*
6198 * dhparam: generate 2048-bit parameters by default.
6200 *Kurt Roeckx and Emilia Kasper*
6202 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6204 * Segmentation fault in ASN1_TYPE_cmp fix
6206 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6207 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6208 certificate signature algorithm consistency this can be used to crash any
6209 certificate verification operation and exploited in a DoS attack. Any
6210 application which performs certificate verification is vulnerable including
6211 OpenSSL clients and servers which enable client authentication.
6216 * ASN.1 structure reuse memory corruption fix
6218 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6219 memory corruption via an invalid write. Such reuse is and has been
6220 strongly discouraged and is believed to be rare.
6222 Applications that parse structures containing CHOICE or ANY DEFINED BY
6223 components may be affected. Certificate parsing (d2i_X509 and related
6224 functions) are however not affected. OpenSSL clients and servers are
6230 * PKCS7 NULL pointer dereferences fix
6232 The PKCS#7 parsing code does not handle missing outer ContentInfo
6233 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6234 missing content and trigger a NULL pointer dereference on parsing.
6236 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6237 otherwise parse PKCS#7 structures from untrusted sources are
6238 affected. OpenSSL clients and servers are not affected.
6240 This issue was reported to OpenSSL by Michal Zalewski (Google).
6245 * DoS via reachable assert in SSLv2 servers fix
6247 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6248 servers that both support SSLv2 and enable export cipher suites by sending
6249 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6251 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6252 (OpenSSL development team).
6257 * Use After Free following d2i_ECPrivatekey error fix
6259 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6260 could cause a use after free condition. This, in turn, could cause a double
6261 free in several private key parsing functions (such as d2i_PrivateKey
6262 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6263 for applications that receive EC private keys from untrusted
6264 sources. This scenario is considered rare.
6266 This issue was discovered by the BoringSSL project and fixed in their
6272 * X509_to_X509_REQ NULL pointer deref fix
6274 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6275 the certificate key is invalid. This function is rarely used in practice.
6277 This issue was discovered by Brian Carpenter.
6282 * Removed the export ciphers from the DEFAULT ciphers
6286 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6288 * Build fixes for the Windows and OpenVMS platforms
6290 *Matt Caswell and Richard Levitte*
6292 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6294 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6295 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6296 dereference. This could lead to a Denial Of Service attack. Thanks to
6297 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6302 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6303 dtls1_buffer_record function under certain conditions. In particular this
6304 could occur if an attacker sent repeated DTLS records with the same
6305 sequence number but for the next epoch. The memory leak could be exploited
6306 by an attacker in a Denial of Service attack through memory exhaustion.
6307 Thanks to Chris Mueller for reporting this issue.
6312 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6313 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6314 method would be set to NULL which could later result in a NULL pointer
6315 dereference. Thanks to Frank Schmirler for reporting this issue.
6320 * Abort handshake if server key exchange message is omitted for ephemeral
6323 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6324 reporting this issue.
6329 * Remove non-export ephemeral RSA code on client and server. This code
6330 violated the TLS standard by allowing the use of temporary RSA keys in
6331 non-export ciphersuites and could be used by a server to effectively
6332 downgrade the RSA key length used to a value smaller than the server
6333 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6334 INRIA or reporting this issue.
6339 * Fixed issue where DH client certificates are accepted without verification.
6340 An OpenSSL server will accept a DH certificate for client authentication
6341 without the certificate verify message. This effectively allows a client to
6342 authenticate without the use of a private key. This only affects servers
6343 which trust a client certificate authority which issues certificates
6344 containing DH keys: these are extremely rare and hardly ever encountered.
6345 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6351 * Ensure that the session ID context of an SSL is updated when its
6352 SSL_CTX is updated via SSL_set_SSL_CTX.
6354 The session ID context is typically set from the parent SSL_CTX,
6355 and can vary with the CTX.
6359 * Fix various certificate fingerprint issues.
6361 By using non-DER or invalid encodings outside the signed portion of a
6362 certificate the fingerprint can be changed without breaking the signature.
6363 Although no details of the signed portion of the certificate can be changed
6364 this can cause problems with some applications: e.g. those using the
6365 certificate fingerprint for blacklists.
6367 1. Reject signatures with non zero unused bits.
6369 If the BIT STRING containing the signature has non zero unused bits reject
6370 the signature. All current signature algorithms require zero unused bits.
6372 2. Check certificate algorithm consistency.
6374 Check the AlgorithmIdentifier inside TBS matches the one in the
6375 certificate signature. NB: this will result in signature failure
6376 errors for some broken certificates.
6378 Thanks to Konrad Kraszewski from Google for reporting this issue.
6380 3. Check DSA/ECDSA signatures use DER.
6382 Re-encode DSA/ECDSA signatures and compare with the original received
6383 signature. Return an error if there is a mismatch.
6385 This will reject various cases including garbage after signature
6386 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6387 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6388 (negative or with leading zeroes).
6390 Further analysis was conducted and fixes were developed by Stephen Henson
6391 of the OpenSSL core team.
6397 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6398 results on some platforms, including x86_64. This bug occurs at random
6399 with a very low probability, and is not known to be exploitable in any
6400 way, though its exact impact is difficult to determine. Thanks to Pieter
6401 Wuille (Blockstream) who reported this issue and also suggested an initial
6402 fix. Further analysis was conducted by the OpenSSL development team and
6403 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6404 the OpenSSL core team.
6409 * Do not resume sessions on the server if the negotiated protocol
6410 version does not match the session's version. Resuming with a different
6411 version, while not strictly forbidden by the RFC, is of questionable
6412 sanity and breaks all known clients.
6414 *David Benjamin, Emilia Käsper*
6416 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6417 early CCS messages during renegotiation. (Note that because
6418 renegotiation is encrypted, this early CCS was not exploitable.)
6422 * Tighten client-side session ticket handling during renegotiation:
6423 ensure that the client only accepts a session ticket if the server sends
6424 the extension anew in the ServerHello. Previously, a TLS client would
6425 reuse the old extension state and thus accept a session ticket if one was
6426 announced in the initial ServerHello.
6428 Similarly, ensure that the client requires a session ticket if one
6429 was advertised in the ServerHello. Previously, a TLS client would
6430 ignore a missing NewSessionTicket message.
6434 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6438 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6439 sends a carefully crafted handshake message, to cause OpenSSL to fail
6440 to free up to 64k of memory causing a memory leak. This could be
6441 exploited in a Denial Of Service attack. This issue affects OpenSSL
6442 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6443 whether SRTP is used or configured. Implementations of OpenSSL that
6444 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6446 The fix was developed by the OpenSSL team.
6451 * Session Ticket Memory Leak.
6453 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6454 integrity of that ticket is first verified. In the event of a session
6455 ticket integrity check failing, OpenSSL will fail to free memory
6456 causing a memory leak. By sending a large number of invalid session
6457 tickets an attacker could exploit this issue in a Denial Of Service
6463 * Build option no-ssl3 is incomplete.
6465 When OpenSSL is configured with "no-ssl3" as a build option, servers
6466 could accept and complete a SSL 3.0 handshake, and clients could be
6467 configured to send them.
6470 *Akamai and the OpenSSL team*
6472 * Add support for TLS_FALLBACK_SCSV.
6473 Client applications doing fallback retries should call
6474 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6477 *Adam Langley, Bodo Moeller*
6479 * Add additional DigestInfo checks.
6481 Re-encode DigestInto in DER and check against the original when
6482 verifying RSA signature: this will reject any improperly encoded
6483 DigestInfo structures.
6485 Note: this is a precautionary measure and no attacks are currently known.
6489 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6491 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6492 SRP code can be overrun an internal buffer. Add sanity check that
6493 g, A, B < N to SRP code.
6495 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6496 Group for discovering this issue.
6501 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6502 TLS 1.0 instead of higher protocol versions when the ClientHello message
6503 is badly fragmented. This allows a man-in-the-middle attacker to force a
6504 downgrade to TLS 1.0 even if both the server and the client support a
6505 higher protocol version, by modifying the client's TLS records.
6507 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6508 researching this issue.
6513 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6514 to a denial of service attack. A malicious server can crash the client
6515 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6516 ciphersuite and sending carefully crafted handshake messages.
6518 Thanks to Felix Gröbert (Google) for discovering and researching this
6524 * By sending carefully crafted DTLS packets an attacker could cause openssl
6525 to leak memory. This can be exploited through a Denial of Service attack.
6526 Thanks to Adam Langley for discovering and researching this issue.
6531 * An attacker can force openssl to consume large amounts of memory whilst
6532 processing DTLS handshake messages. This can be exploited through a
6533 Denial of Service attack.
6534 Thanks to Adam Langley for discovering and researching this issue.
6539 * An attacker can force an error condition which causes openssl to crash
6540 whilst processing DTLS packets due to memory being freed twice. This
6541 can be exploited through a Denial of Service attack.
6542 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6548 * If a multithreaded client connects to a malicious server using a resumed
6549 session and the server sends an ec point format extension it could write
6550 up to 255 bytes to freed memory.
6552 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6558 * A malicious server can crash an OpenSSL client with a null pointer
6559 dereference (read) by specifying an SRP ciphersuite even though it was not
6560 properly negotiated with the client. This can be exploited through a
6561 Denial of Service attack.
6563 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6564 discovering and researching this issue.
6569 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6570 X509_name_oneline, X509_name_print_ex et al. to leak some information
6571 from the stack. Applications may be affected if they echo pretty printing
6572 output to the attacker.
6574 Thanks to Ivan Fratric (Google) for discovering this issue.
6577 *Emilia Käsper, and Steve Henson*
6579 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6580 for corner cases. (Certain input points at infinity could lead to
6581 bogus results, with non-infinity inputs mapped to infinity too.)
6585 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6587 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6588 handshake can force the use of weak keying material in OpenSSL
6589 SSL/TLS clients and servers.
6591 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6592 researching this issue. ([CVE-2014-0224])
6594 *KIKUCHI Masashi, Steve Henson*
6596 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6597 OpenSSL DTLS client the code can be made to recurse eventually crashing
6600 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6603 *Imre Rad, Steve Henson*
6605 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6606 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6607 client or server. This is potentially exploitable to run arbitrary
6608 code on a vulnerable client or server.
6610 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6612 *Jüri Aedla, Steve Henson*
6614 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6615 are subject to a denial of service attack.
6617 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6618 this issue. ([CVE-2014-3470])
6620 *Felix Gröbert, Ivan Fratric, Steve Henson*
6622 * Harmonize version and its documentation. -f flag is used to display
6625 *mancha <mancha1@zoho.com>*
6627 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6628 in i2d_ECPrivateKey.
6630 *mancha <mancha1@zoho.com>*
6632 * Fix some double frees. These are not thought to be exploitable.
6634 *mancha <mancha1@zoho.com>*
6636 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6638 * A missing bounds check in the handling of the TLS heartbeat extension
6639 can be used to reveal up to 64k of memory to a connected client or
6642 Thanks for Neel Mehta of Google Security for discovering this bug and to
6643 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6644 preparing the fix ([CVE-2014-0160])
6646 *Adam Langley, Bodo Moeller*
6648 * Fix for the attack described in the paper "Recovering OpenSSL
6649 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6650 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6651 <http://eprint.iacr.org/2014/140>
6653 Thanks to Yuval Yarom and Naomi Benger for discovering this
6654 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6656 *Yuval Yarom and Naomi Benger*
6658 * TLS pad extension: draft-agl-tls-padding-03
6660 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6661 TLS client Hello record length value would otherwise be > 255 and
6662 less that 512 pad with a dummy extension containing zeroes so it
6663 is at least 512 bytes long.
6665 *Adam Langley, Steve Henson*
6667 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6669 * Fix for TLS record tampering bug. A carefully crafted invalid
6670 handshake could crash OpenSSL with a NULL pointer exception.
6671 Thanks to Anton Johansson for reporting this issues.
6674 * Keep original DTLS digest and encryption contexts in retransmission
6675 structures so we can use the previous session parameters if they need
6676 to be resent. ([CVE-2013-6450])
6680 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6681 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6682 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6683 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6684 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6685 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6687 *Rob Stradling, Adam Langley*
6689 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6691 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6692 supporting platforms or when small records were transferred.
6694 *Andy Polyakov, Steve Henson*
6696 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6698 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6700 This addresses the flaw in CBC record processing discovered by
6701 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6702 at: <http://www.isg.rhul.ac.uk/tls/>
6704 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6705 Security Group at Royal Holloway, University of London
6706 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6707 Emilia Käsper for the initial patch.
6710 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6712 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6713 ciphersuites which can be exploited in a denial of service attack.
6714 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6715 and detecting this bug and to Wolfgang Ettlinger
6716 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6721 * Return an error when checking OCSP signatures when key is NULL.
6722 This fixes a DoS attack. ([CVE-2013-0166])
6726 * Make openssl verify return errors.
6728 *Chris Palmer <palmer@google.com> and Ben Laurie*
6730 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6731 the right response is stapled. Also change SSL_get_certificate()
6732 so it returns the certificate actually sent.
6733 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6735 *Rob Stradling <rob.stradling@comodo.com>*
6737 * Fix possible deadlock when decoding public keys.
6741 * Don't use TLS 1.0 record version number in initial client hello
6746 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6748 * Sanity check record length before skipping explicit IV in TLS
6749 1.2, 1.1 and DTLS to fix DoS attack.
6751 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6752 fuzzing as a service testing platform.
6757 * Initialise tkeylen properly when encrypting CMS messages.
6758 Thanks to Solar Designer of Openwall for reporting this issue.
6762 * In FIPS mode don't try to use composite ciphers as they are not
6767 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6769 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6770 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6771 mean any application compiled against OpenSSL 1.0.0 headers setting
6772 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6773 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6774 0x10000000L Any application which was previously compiled against
6775 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6776 will need to be recompiled as a result. Letting be results in
6777 inability to disable specifically TLS 1.1 and in client context,
6778 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6782 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6783 disable just protocol X, but all protocols above X *if* there are
6784 protocols *below* X still enabled. In more practical terms it means
6785 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6786 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6787 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6792 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6794 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6795 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6796 in CRYPTO_realloc_clean.
6798 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6799 issue and to Adam Langley <agl@chromium.org> for fixing it.
6802 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6804 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6808 * Workarounds for some broken servers that "hang" if a client hello
6809 record length exceeds 255 bytes.
6811 1. Do not use record version number > TLS 1.0 in initial client
6812 hello: some (but not all) hanging servers will now work.
6813 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6814 the number of ciphers sent in the client hello. This should be
6815 set to an even number, such as 50, for example by passing:
6816 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6817 Most broken servers should now work.
6818 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6819 TLS 1.2 client support entirely.
6823 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6827 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6829 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6830 STRING form instead of a DigestInfo.
6834 * The format used for MDC2 RSA signatures is inconsistent between EVP
6835 and the RSA_sign/RSA_verify functions. This was made more apparent when
6836 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6837 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6838 the correct format in RSA_verify so both forms transparently work.
6842 * Some servers which support TLS 1.0 can choke if we initially indicate
6843 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6844 encrypted premaster secret. As a workaround use the maximum permitted
6845 client version in client hello, this should keep such servers happy
6846 and still work with previous versions of OpenSSL.
6850 * Add support for TLS/DTLS heartbeats.
6852 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6854 * Add support for SCTP.
6856 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6858 * Improved PRNG seeding for VOS.
6860 *Paul Green <Paul.Green@stratus.com>*
6862 * Extensive assembler packs updates, most notably:
6864 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6865 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6866 - x86_64: bit-sliced AES implementation;
6867 - ARM: NEON support, contemporary platforms optimizations;
6868 - s390x: z196 support;
6869 - `*`: GHASH and GF(2^m) multiplication implementations;
6873 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6874 (removal of unnecessary code)
6876 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6878 * Add TLS key material exporter from RFC 5705.
6882 * Add DTLS-SRTP negotiation from RFC 5764.
6886 * Add Next Protocol Negotiation,
6887 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6888 disabled with a no-npn flag to config or Configure. Code donated
6891 *Adam Langley <agl@google.com> and Ben Laurie*
6893 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6894 NIST-P256, NIST-P521, with constant-time single point multiplication on
6895 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6896 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6897 Code made available under Apache License version 2.0.
6899 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6900 line to include this in your build of OpenSSL, and run "make depend" (or
6901 "make update"). This enables the following EC_METHODs:
6903 EC_GFp_nistp224_method()
6904 EC_GFp_nistp256_method()
6905 EC_GFp_nistp521_method()
6907 EC_GROUP_new_by_curve_name() will automatically use these (while
6908 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6911 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6913 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6914 all platforms. Move ssize_t definition from e_os.h to the public
6915 header file e_os2.h as it now appears in public header file cms.h
6919 * New -sigopt option to the ca, req and x509 utilities. Additional
6920 signature parameters can be passed using this option and in
6925 * Add RSA PSS signing function. This will generate and set the
6926 appropriate AlgorithmIdentifiers for PSS based on those in the
6927 corresponding EVP_MD_CTX structure. No application support yet.
6931 * Support for companion algorithm specific ASN1 signing routines.
6932 New function ASN1_item_sign_ctx() signs a pre-initialised
6933 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6934 the appropriate parameters.
6938 * Add new algorithm specific ASN1 verification initialisation function
6939 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6940 handling will be the same no matter what EVP_PKEY_METHOD is used.
6941 Add a PSS handler to support verification of PSS signatures: checked
6942 against a number of sample certificates.
6946 * Add signature printing for PSS. Add PSS OIDs.
6948 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6950 * Add algorithm specific signature printing. An individual ASN1 method
6951 can now print out signatures instead of the standard hex dump.
6953 More complex signatures (e.g. PSS) can print out more meaningful
6954 information. Include DSA version that prints out the signature
6959 * Password based recipient info support for CMS library: implementing
6964 * Split password based encryption into PBES2 and PBKDF2 functions. This
6965 neatly separates the code into cipher and PBE sections and is required
6966 for some algorithms that split PBES2 into separate pieces (such as
6967 password based CMS).
6971 * Session-handling fixes:
6972 - Fix handling of connections that are resuming with a session ID,
6973 but also support Session Tickets.
6974 - Fix a bug that suppressed issuing of a new ticket if the client
6975 presented a ticket with an expired session.
6976 - Try to set the ticket lifetime hint to something reasonable.
6977 - Make tickets shorter by excluding irrelevant information.
6978 - On the client side, don't ignore renewed tickets.
6980 *Adam Langley, Bodo Moeller (Google)*
6982 * Fix PSK session representation.
6986 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6988 This work was sponsored by Intel.
6992 * Add GCM support to TLS library. Some custom code is needed to split
6993 the IV between the fixed (from PRF) and explicit (from TLS record)
6994 portions. This adds all GCM ciphersuites supported by RFC5288 and
6995 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6996 add a special AESGCM string for GCM only.
7000 * Expand range of ctrls for AES GCM. Permit setting invocation
7001 field on decrypt and retrieval of invocation field only on encrypt.
7005 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7006 As required by RFC5289 these ciphersuites cannot be used if for
7007 versions of TLS earlier than 1.2.
7011 * For FIPS capable OpenSSL interpret a NULL default public key method
7012 as unset and return the appropriate default but do *not* set the default.
7013 This means we can return the appropriate method in applications that
7014 switch between FIPS and non-FIPS modes.
7018 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7019 ENGINE is used then we cannot handle that in the FIPS module so we
7020 keep original code iff non-FIPS operations are allowed.
7024 * Add -attime option to openssl utilities.
7026 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7028 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7032 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7033 FIPS EC methods unconditionally for now.
7037 * New build option no-ec2m to disable characteristic 2 code.
7041 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7042 all cases can be covered as some introduce binary incompatibilities.
7046 * Redirect RSA operations to FIPS module including keygen,
7047 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7051 * Add similar low level API blocking to ciphers.
7055 * Low level digest APIs are not approved in FIPS mode: any attempt
7056 to use these will cause a fatal error. Applications that *really* want
7057 to use them can use the `private_*` version instead.
7061 * Redirect cipher operations to FIPS module for FIPS builds.
7065 * Redirect digest operations to FIPS module for FIPS builds.
7069 * Update build system to add "fips" flag which will link in fipscanister.o
7070 for static and shared library builds embedding a signature if needed.
7074 * Output TLS supported curves in preference order instead of numerical
7075 order. This is currently hardcoded for the highest order curves first.
7076 This should be configurable so applications can judge speed vs strength.
7080 * Add TLS v1.2 server support for client authentication.
7084 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7089 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7090 FIPS modules versions.
7094 * Add TLS v1.2 client side support for client authentication. Keep cache
7095 of handshake records longer as we don't know the hash algorithm to use
7096 until after the certificate request message is received.
7100 * Initial TLS v1.2 client support. Add a default signature algorithms
7101 extension including all the algorithms we support. Parse new signature
7102 format in client key exchange. Relax some ECC signing restrictions for
7103 TLS v1.2 as indicated in RFC5246.
7107 * Add server support for TLS v1.2 signature algorithms extension. Switch
7108 to new signature format when needed using client digest preference.
7109 All server ciphersuites should now work correctly in TLS v1.2. No client
7110 support yet and no support for client certificates.
7114 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7115 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7116 ciphersuites. At present only RSA key exchange ciphersuites work with
7117 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7118 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7119 and version checking.
7123 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7124 with this defined it will not be affected by any changes to ssl internal
7125 structures. Add several utility functions to allow openssl application
7126 to work with OPENSSL_NO_SSL_INTERN defined.
7130 * A long standing patch to add support for SRP from EdelWeb (Peter
7131 Sylvester and Christophe Renou) was integrated.
7132 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7133 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7136 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7140 * Permit abbreviated handshakes when renegotiating using the function
7141 SSL_renegotiate_abbreviated().
7143 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7145 * Add call to ENGINE_register_all_complete() to
7146 ENGINE_load_builtin_engines(), so some implementations get used
7147 automatically instead of needing explicit application support.
7151 * Add support for TLS key exporter as described in RFC5705.
7153 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7155 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7156 a few changes are required:
7158 Add SSL_OP_NO_TLSv1_1 flag.
7159 Add TLSv1_1 methods.
7160 Update version checking logic to handle version 1.1.
7161 Add explicit IV handling (ported from DTLS code).
7162 Add command line options to s_client/s_server.
7169 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7171 * X509_ATTRIBUTE memory leak
7173 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7174 memory. This structure is used by the PKCS#7 and CMS routines so any
7175 application which reads PKCS#7 or CMS data from untrusted sources is
7176 affected. SSL/TLS is not affected.
7178 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7184 * Race condition handling PSK identify hint
7186 If PSK identity hints are received by a multi-threaded client then
7187 the values are wrongly updated in the parent SSL_CTX structure. This can
7188 result in a race condition potentially leading to a double free of the
7194 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7196 * Malformed ECParameters causes infinite loop
7198 When processing an ECParameters structure OpenSSL enters an infinite loop
7199 if the curve specified is over a specially malformed binary polynomial
7202 This can be used to perform denial of service against any
7203 system which processes public keys, certificate requests or
7204 certificates. This includes TLS clients and TLS servers with
7205 client authentication enabled.
7207 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7212 * Exploitable out-of-bounds read in X509_cmp_time
7214 X509_cmp_time does not properly check the length of the ASN1_TIME
7215 string and can read a few bytes out of bounds. In addition,
7216 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7219 An attacker can use this to craft malformed certificates and CRLs of
7220 various sizes and potentially cause a segmentation fault, resulting in
7221 a DoS on applications that verify certificates or CRLs. TLS clients
7222 that verify CRLs are affected. TLS clients and servers with client
7223 authentication enabled may be affected if they use custom verification
7226 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7227 independently by Hanno Böck.
7232 * PKCS7 crash with missing EnvelopedContent
7234 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7235 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7236 with missing content and trigger a NULL pointer dereference on parsing.
7238 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7239 structures from untrusted sources are affected. OpenSSL clients and
7240 servers are not affected.
7242 This issue was reported to OpenSSL by Michal Zalewski (Google).
7247 * CMS verify infinite loop with unknown hash function
7249 When verifying a signedData message the CMS code can enter an infinite loop
7250 if presented with an unknown hash function OID. This can be used to perform
7251 denial of service against any system which verifies signedData messages using
7253 This issue was reported to OpenSSL by Johannes Bauer.
7258 * Race condition handling NewSessionTicket
7260 If a NewSessionTicket is received by a multi-threaded client when attempting to
7261 reuse a previous ticket then a race condition can occur potentially leading to
7262 a double free of the ticket data.
7267 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7269 * Segmentation fault in ASN1_TYPE_cmp fix
7271 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7272 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7273 certificate signature algorithm consistency this can be used to crash any
7274 certificate verification operation and exploited in a DoS attack. Any
7275 application which performs certificate verification is vulnerable including
7276 OpenSSL clients and servers which enable client authentication.
7281 * ASN.1 structure reuse memory corruption fix
7283 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7284 memory corruption via an invalid write. Such reuse is and has been
7285 strongly discouraged and is believed to be rare.
7287 Applications that parse structures containing CHOICE or ANY DEFINED BY
7288 components may be affected. Certificate parsing (d2i_X509 and related
7289 functions) are however not affected. OpenSSL clients and servers are
7295 * PKCS7 NULL pointer dereferences fix
7297 The PKCS#7 parsing code does not handle missing outer ContentInfo
7298 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7299 missing content and trigger a NULL pointer dereference on parsing.
7301 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7302 otherwise parse PKCS#7 structures from untrusted sources are
7303 affected. OpenSSL clients and servers are not affected.
7305 This issue was reported to OpenSSL by Michal Zalewski (Google).
7310 * DoS via reachable assert in SSLv2 servers fix
7312 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7313 servers that both support SSLv2 and enable export cipher suites by sending
7314 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7316 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7317 (OpenSSL development team).
7322 * Use After Free following d2i_ECPrivatekey error fix
7324 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7325 could cause a use after free condition. This, in turn, could cause a double
7326 free in several private key parsing functions (such as d2i_PrivateKey
7327 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7328 for applications that receive EC private keys from untrusted
7329 sources. This scenario is considered rare.
7331 This issue was discovered by the BoringSSL project and fixed in their
7337 * X509_to_X509_REQ NULL pointer deref fix
7339 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7340 the certificate key is invalid. This function is rarely used in practice.
7342 This issue was discovered by Brian Carpenter.
7347 * Removed the export ciphers from the DEFAULT ciphers
7351 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7353 * Build fixes for the Windows and OpenVMS platforms
7355 *Matt Caswell and Richard Levitte*
7357 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7359 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7360 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7361 dereference. This could lead to a Denial Of Service attack. Thanks to
7362 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7367 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7368 dtls1_buffer_record function under certain conditions. In particular this
7369 could occur if an attacker sent repeated DTLS records with the same
7370 sequence number but for the next epoch. The memory leak could be exploited
7371 by an attacker in a Denial of Service attack through memory exhaustion.
7372 Thanks to Chris Mueller for reporting this issue.
7377 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7378 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7379 method would be set to NULL which could later result in a NULL pointer
7380 dereference. Thanks to Frank Schmirler for reporting this issue.
7385 * Abort handshake if server key exchange message is omitted for ephemeral
7388 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7389 reporting this issue.
7394 * Remove non-export ephemeral RSA code on client and server. This code
7395 violated the TLS standard by allowing the use of temporary RSA keys in
7396 non-export ciphersuites and could be used by a server to effectively
7397 downgrade the RSA key length used to a value smaller than the server
7398 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7399 INRIA or reporting this issue.
7404 * Fixed issue where DH client certificates are accepted without verification.
7405 An OpenSSL server will accept a DH certificate for client authentication
7406 without the certificate verify message. This effectively allows a client to
7407 authenticate without the use of a private key. This only affects servers
7408 which trust a client certificate authority which issues certificates
7409 containing DH keys: these are extremely rare and hardly ever encountered.
7410 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7416 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7417 results on some platforms, including x86_64. This bug occurs at random
7418 with a very low probability, and is not known to be exploitable in any
7419 way, though its exact impact is difficult to determine. Thanks to Pieter
7420 Wuille (Blockstream) who reported this issue and also suggested an initial
7421 fix. Further analysis was conducted by the OpenSSL development team and
7422 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7423 the OpenSSL core team.
7428 * Fix various certificate fingerprint issues.
7430 By using non-DER or invalid encodings outside the signed portion of a
7431 certificate the fingerprint can be changed without breaking the signature.
7432 Although no details of the signed portion of the certificate can be changed
7433 this can cause problems with some applications: e.g. those using the
7434 certificate fingerprint for blacklists.
7436 1. Reject signatures with non zero unused bits.
7438 If the BIT STRING containing the signature has non zero unused bits reject
7439 the signature. All current signature algorithms require zero unused bits.
7441 2. Check certificate algorithm consistency.
7443 Check the AlgorithmIdentifier inside TBS matches the one in the
7444 certificate signature. NB: this will result in signature failure
7445 errors for some broken certificates.
7447 Thanks to Konrad Kraszewski from Google for reporting this issue.
7449 3. Check DSA/ECDSA signatures use DER.
7451 Reencode DSA/ECDSA signatures and compare with the original received
7452 signature. Return an error if there is a mismatch.
7454 This will reject various cases including garbage after signature
7455 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7456 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7457 (negative or with leading zeroes).
7459 Further analysis was conducted and fixes were developed by Stephen Henson
7460 of the OpenSSL core team.
7466 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7468 * Session Ticket Memory Leak.
7470 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7471 integrity of that ticket is first verified. In the event of a session
7472 ticket integrity check failing, OpenSSL will fail to free memory
7473 causing a memory leak. By sending a large number of invalid session
7474 tickets an attacker could exploit this issue in a Denial Of Service
7480 * Build option no-ssl3 is incomplete.
7482 When OpenSSL is configured with "no-ssl3" as a build option, servers
7483 could accept and complete a SSL 3.0 handshake, and clients could be
7484 configured to send them.
7487 *Akamai and the OpenSSL team*
7489 * Add support for TLS_FALLBACK_SCSV.
7490 Client applications doing fallback retries should call
7491 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7494 *Adam Langley, Bodo Moeller*
7496 * Add additional DigestInfo checks.
7498 Reencode DigestInto in DER and check against the original when
7499 verifying RSA signature: this will reject any improperly encoded
7500 DigestInfo structures.
7502 Note: this is a precautionary measure and no attacks are currently known.
7506 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7508 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7509 to a denial of service attack. A malicious server can crash the client
7510 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7511 ciphersuite and sending carefully crafted handshake messages.
7513 Thanks to Felix Gröbert (Google) for discovering and researching this
7519 * By sending carefully crafted DTLS packets an attacker could cause openssl
7520 to leak memory. This can be exploited through a Denial of Service attack.
7521 Thanks to Adam Langley for discovering and researching this issue.
7526 * An attacker can force openssl to consume large amounts of memory whilst
7527 processing DTLS handshake messages. This can be exploited through a
7528 Denial of Service attack.
7529 Thanks to Adam Langley for discovering and researching this issue.
7534 * An attacker can force an error condition which causes openssl to crash
7535 whilst processing DTLS packets due to memory being freed twice. This
7536 can be exploited through a Denial of Service attack.
7537 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7543 * If a multithreaded client connects to a malicious server using a resumed
7544 session and the server sends an ec point format extension it could write
7545 up to 255 bytes to freed memory.
7547 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7553 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7554 X509_name_oneline, X509_name_print_ex et al. to leak some information
7555 from the stack. Applications may be affected if they echo pretty printing
7556 output to the attacker.
7558 Thanks to Ivan Fratric (Google) for discovering this issue.
7561 *Emilia Käsper, and Steve Henson*
7563 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7564 for corner cases. (Certain input points at infinity could lead to
7565 bogus results, with non-infinity inputs mapped to infinity too.)
7569 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7571 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7572 handshake can force the use of weak keying material in OpenSSL
7573 SSL/TLS clients and servers.
7575 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7576 researching this issue. ([CVE-2014-0224])
7578 *KIKUCHI Masashi, Steve Henson*
7580 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7581 OpenSSL DTLS client the code can be made to recurse eventually crashing
7584 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7587 *Imre Rad, Steve Henson*
7589 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7590 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7591 client or server. This is potentially exploitable to run arbitrary
7592 code on a vulnerable client or server.
7594 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7596 *Jüri Aedla, Steve Henson*
7598 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7599 are subject to a denial of service attack.
7601 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7602 this issue. ([CVE-2014-3470])
7604 *Felix Gröbert, Ivan Fratric, Steve Henson*
7606 * Harmonize version and its documentation. -f flag is used to display
7609 *mancha <mancha1@zoho.com>*
7611 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7612 in i2d_ECPrivateKey.
7614 *mancha <mancha1@zoho.com>*
7616 * Fix some double frees. These are not thought to be exploitable.
7618 *mancha <mancha1@zoho.com>*
7620 * Fix for the attack described in the paper "Recovering OpenSSL
7621 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7622 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7623 <http://eprint.iacr.org/2014/140>
7625 Thanks to Yuval Yarom and Naomi Benger for discovering this
7626 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7628 *Yuval Yarom and Naomi Benger*
7630 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7632 * Keep original DTLS digest and encryption contexts in retransmission
7633 structures so we can use the previous session parameters if they need
7634 to be resent. ([CVE-2013-6450])
7638 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7639 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7640 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7641 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7642 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7643 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7645 *Rob Stradling, Adam Langley*
7647 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7649 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7651 This addresses the flaw in CBC record processing discovered by
7652 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7653 at: <http://www.isg.rhul.ac.uk/tls/>
7655 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7656 Security Group at Royal Holloway, University of London
7657 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7658 Emilia Käsper for the initial patch.
7661 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7663 * Return an error when checking OCSP signatures when key is NULL.
7664 This fixes a DoS attack. ([CVE-2013-0166])
7668 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7669 the right response is stapled. Also change SSL_get_certificate()
7670 so it returns the certificate actually sent.
7671 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7672 (This is a backport)
7674 *Rob Stradling <rob.stradling@comodo.com>*
7676 * Fix possible deadlock when decoding public keys.
7680 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7682 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7685 * Sanity check record length before skipping explicit IV in DTLS
7688 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7689 fuzzing as a service testing platform.
7694 * Initialise tkeylen properly when encrypting CMS messages.
7695 Thanks to Solar Designer of Openwall for reporting this issue.
7699 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7701 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7702 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7703 in CRYPTO_realloc_clean.
7705 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7706 issue and to Adam Langley <agl@chromium.org> for fixing it.
7709 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7711 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7713 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7714 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7715 content decryption and always return the same error. Note: this attack
7716 needs on average 2^20 messages so it only affects automated senders. The
7717 old behaviour can be re-enabled in the CMS code by setting the
7718 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7719 an MMA defence is not necessary.
7720 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7721 this issue. ([CVE-2012-0884])
7725 * Fix CVE-2011-4619: make sure we really are receiving a
7726 client hello before rejecting multiple SGC restarts. Thanks to
7727 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7731 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7733 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7734 Thanks to Antonio Martin, Enterprise Secure Access Research and
7735 Development, Cisco Systems, Inc. for discovering this bug and
7736 preparing a fix. ([CVE-2012-0050])
7740 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7742 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7743 of the Vaudenay padding oracle attack on CBC mode encryption
7744 which enables an efficient plaintext recovery attack against
7745 the OpenSSL implementation of DTLS. Their attack exploits timing
7746 differences arising during decryption processing. A research
7747 paper describing this attack can be found at:
7748 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7749 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7750 Security Group at Royal Holloway, University of London
7751 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7752 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7753 for preparing the fix. ([CVE-2011-4108])
7755 *Robin Seggelmann, Michael Tuexen*
7757 * Clear bytes used for block padding of SSL 3.0 records.
7760 *Adam Langley (Google)*
7762 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7763 Kadianakis <desnacked@gmail.com> for discovering this issue and
7764 Adam Langley for preparing the fix. ([CVE-2011-4619])
7766 *Adam Langley (Google)*
7768 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7770 *Andrey Kulikov <amdeich@gmail.com>*
7772 * Prevent malformed RFC3779 data triggering an assertion failure.
7773 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7774 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7776 *Rob Austein <sra@hactrn.net>*
7778 * Improved PRNG seeding for VOS.
7780 *Paul Green <Paul.Green@stratus.com>*
7782 * Fix ssl_ciph.c set-up race.
7784 *Adam Langley (Google)*
7786 * Fix spurious failures in ecdsatest.c.
7788 *Emilia Käsper (Google)*
7790 * Fix the BIO_f_buffer() implementation (which was mixing different
7791 interpretations of the `..._len` fields).
7793 *Adam Langley (Google)*
7795 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7796 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7797 threads won't reuse the same blinding coefficients.
7799 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7800 lock to call BN_BLINDING_invert_ex, and avoids one use of
7801 BN_BLINDING_update for each BN_BLINDING structure (previously,
7802 the last update always remained unused).
7804 *Emilia Käsper (Google)*
7806 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7808 *Bob Buckholz (Google)*
7810 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7812 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7813 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7815 *Kaspar Brand <ossl@velox.ch>*
7817 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7818 for multi-threaded use of ECDH. ([CVE-2011-3210])
7820 *Adam Langley (Google)*
7822 * Fix x509_name_ex_d2i memory leak on bad inputs.
7826 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7827 signature public key algorithm by using OID xref utilities instead.
7828 Before this you could only use some ECC ciphersuites with SHA1 only.
7832 * Add protection against ECDSA timing attacks as mentioned in the paper
7833 by Billy Bob Brumley and Nicola Tuveri, see:
7834 <http://eprint.iacr.org/2011/232.pdf>
7836 *Billy Bob Brumley and Nicola Tuveri*
7838 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7840 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7842 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7844 * Fix bug in string printing code: if *any* escaping is enabled we must
7845 escape the escape character (backslash) or the resulting string is
7850 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7852 * Disable code workaround for ancient and obsolete Netscape browsers
7853 and servers: an attacker can use it in a ciphersuite downgrade attack.
7854 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7858 * Fixed J-PAKE implementation error, originally discovered by
7859 Sebastien Martini, further info and confirmation from Stefan
7860 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7864 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7866 * Fix extension code to avoid race conditions which can result in a buffer
7867 overrun vulnerability: resumed sessions must not be modified as they can
7868 be shared by multiple threads. CVE-2010-3864
7872 * Fix WIN32 build system to correctly link an ENGINE directory into
7877 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7879 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7882 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7884 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7886 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7887 context. The operation can be customised via the ctrl mechanism in
7888 case ENGINEs want to include additional functionality.
7892 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7896 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7897 output hashes compatible with older versions of OpenSSL.
7899 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7901 * Fix compression algorithm handling: if resuming a session use the
7902 compression algorithm of the resumed session instead of determining
7903 it from client hello again. Don't allow server to change algorithm.
7907 * Add load_crls() function to apps tidying load_certs() too. Add option
7908 to verify utility to allow additional CRLs to be included.
7912 * Update OCSP request code to permit adding custom headers to the request:
7913 some responders need this.
7917 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7920 *Julia Lawall <julia@diku.dk>*
7922 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7923 needlessly dereferenced structures, used obsolete functions and
7924 didn't handle all updated verify codes correctly.
7928 * Disable MD2 in the default configuration.
7932 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7933 indicate the initial BIO being pushed or popped. This makes it possible
7934 to determine whether the BIO is the one explicitly called or as a result
7935 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7936 it handles reference counts correctly and doesn't zero out the I/O bio
7937 when it is not being explicitly popped. WARNING: applications which
7938 included workarounds for the old buggy behaviour will need to be modified
7939 or they could free up already freed BIOs.
7943 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7944 renaming to all platforms (within the 0.9.8 branch, this was
7945 done conditionally on Netware platforms to avoid a name clash).
7947 *Guenter <lists@gknw.net>*
7949 * Add ECDHE and PSK support to DTLS.
7951 *Michael Tuexen <tuexen@fh-muenster.de>*
7953 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7958 * Add "missing" function EVP_MD_flags() (without this the only way to
7959 retrieve a digest flags is by accessing the structure directly. Update
7960 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7961 or cipher is registered as in the "from" argument. Print out all
7962 registered digests in the dgst usage message instead of manually
7963 attempting to work them out.
7967 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7968 this allows the use of compression and extensions. Change default cipher
7969 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7970 by default unless an application cipher string requests it.
7974 * Alter match criteria in PKCS12_parse(). It used to try to use local
7975 key ids to find matching certificates and keys but some PKCS#12 files
7976 don't follow the (somewhat unwritten) rules and this strategy fails.
7977 Now just gather all certificates together and the first private key
7978 then look for the first certificate that matches the key.
7982 * Support use of registered digest and cipher names for dgst and cipher
7983 commands instead of having to add each one as a special case. So now
7990 openssl dgst -sha256 foo
7992 and this works for ENGINE based algorithms too.
7996 * Update Gost ENGINE to support parameter files.
7998 *Victor B. Wagner <vitus@cryptocom.ru>*
8000 * Support GeneralizedTime in ca utility.
8002 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8004 * Enhance the hash format used for certificate directory links. The new
8005 form uses the canonical encoding (meaning equivalent names will work
8006 even if they aren't identical) and uses SHA1 instead of MD5. This form
8007 is incompatible with the older format and as a result c_rehash should
8008 be used to rebuild symbolic links.
8012 * Make PKCS#8 the default write format for private keys, replacing the
8013 traditional format. This form is standardised, more secure and doesn't
8014 include an implicit MD5 dependency.
8018 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8019 committed to OpenSSL should pass this lot as a minimum.
8023 * Add session ticket override functionality for use by EAP-FAST.
8025 *Jouni Malinen <j@w1.fi>*
8027 * Modify HMAC functions to return a value. Since these can be implemented
8028 in an ENGINE errors can occur.
8032 * Type-checked OBJ_bsearch_ex.
8036 * Type-checked OBJ_bsearch. Also some constification necessitated
8037 by type-checking. Still to come: TXT_DB, bsearch(?),
8038 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8043 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8044 seconds to a tm structure directly, instead of going through OS
8045 specific date routines. This avoids any issues with OS routines such
8046 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8047 and X509_time_adj_ex() to cover the extended range. The existing
8048 X509_time_adj() is still usable and will no longer have any date issues.
8052 * Delta CRL support. New use deltas option which will attempt to locate
8053 and search any appropriate delta CRLs available.
8055 This work was sponsored by Google.
8059 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8060 code and add additional score elements. Validate alternate CRL paths
8061 as part of the CRL checking and indicate a new error "CRL path validation
8062 error" in this case. Applications wanting additional details can use
8063 the verify callback and check the new "parent" field. If this is not
8064 NULL CRL path validation is taking place. Existing applications won't
8065 see this because it requires extended CRL support which is off by
8068 This work was sponsored by Google.
8072 * Support for freshest CRL extension.
8074 This work was sponsored by Google.
8078 * Initial indirect CRL support. Currently only supported in the CRLs
8079 passed directly and not via lookup. Process certificate issuer
8080 CRL entry extension and lookup CRL entries by bother issuer name
8081 and serial number. Check and process CRL issuer entry in IDP extension.
8083 This work was sponsored by Google.
8087 * Add support for distinct certificate and CRL paths. The CRL issuer
8088 certificate is validated separately in this case. Only enabled if
8089 an extended CRL support flag is set: this flag will enable additional
8090 CRL functionality in future.
8092 This work was sponsored by Google.
8096 * Add support for policy mappings extension.
8098 This work was sponsored by Google.
8102 * Fixes to pathlength constraint, self issued certificate handling,
8103 policy processing to align with RFC3280 and PKITS tests.
8105 This work was sponsored by Google.
8109 * Support for name constraints certificate extension. DN, email, DNS
8110 and URI types are currently supported.
8112 This work was sponsored by Google.
8116 * To cater for systems that provide a pointer-based thread ID rather
8117 than numeric, deprecate the current numeric thread ID mechanism and
8118 replace it with a structure and associated callback type. This
8119 mechanism allows a numeric "hash" to be extracted from a thread ID in
8120 either case, and on platforms where pointers are larger than 'long',
8121 mixing is done to help ensure the numeric 'hash' is usable even if it
8122 can't be guaranteed unique. The default mechanism is to use "&errno"
8123 as a pointer-based thread ID to distinguish between threads.
8125 Applications that want to provide their own thread IDs should now use
8126 CRYPTO_THREADID_set_callback() to register a callback that will call
8127 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8129 Note that ERR_remove_state() is now deprecated, because it is tied
8130 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8131 to free the current thread's error state should be replaced by
8132 ERR_remove_thread_state(NULL).
8134 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8135 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8136 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8137 application was previously providing a numeric thread callback that
8138 was inappropriate for distinguishing threads, then uniqueness might
8139 have been obtained with &errno that happened immediately in the
8140 intermediate development versions of OpenSSL; this is no longer the
8141 case, the numeric thread callback will now override the automatic use
8144 *Geoff Thorpe, with help from Bodo Moeller*
8146 * Initial support for different CRL issuing certificates. This covers a
8147 simple case where the self issued certificates in the chain exist and
8148 the real CRL issuer is higher in the existing chain.
8150 This work was sponsored by Google.
8154 * Removed effectively defunct crypto/store from the build.
8158 * Revamp of STACK to provide stronger type-checking. Still to come:
8159 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8160 ASN1_STRING, CONF_VALUE.
8164 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8165 RAM on SSL connections. This option can save about 34k per idle SSL.
8169 * Revamp of LHASH to provide stronger type-checking. Still to come:
8170 STACK, TXT_DB, bsearch, qsort.
8174 * Initial support for Cryptographic Message Syntax (aka CMS) based
8175 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8176 support for data, signedData, compressedData, digestedData and
8177 encryptedData, envelopedData types included. Scripts to check against
8178 RFC4134 examples draft and interop and consistency checks of many
8179 content types and variants.
8183 * Add options to enc utility to support use of zlib compression BIO.
8187 * Extend mk1mf to support importing of options and assembly language
8188 files from Configure script, currently only included in VC-WIN32.
8189 The assembly language rules can now optionally generate the source
8190 files from the associated perl scripts.
8194 * Implement remaining functionality needed to support GOST ciphersuites.
8195 Interop testing has been performed using CryptoPro implementations.
8197 *Victor B. Wagner <vitus@cryptocom.ru>*
8199 * s390x assembler pack.
8203 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8208 * Implement Opaque PRF Input TLS extension as specified in
8209 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8210 official specification yet and no extension type assignment by
8211 IANA exists, this extension (for now) will have to be explicitly
8212 enabled when building OpenSSL by providing the extension number
8213 to use. For example, specify an option
8215 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8217 to the "config" or "Configure" script to enable the extension,
8218 assuming extension number 0x9527 (which is a completely arbitrary
8219 and unofficial assignment based on the MD5 hash of the Internet
8220 Draft). Note that by doing so, you potentially lose
8221 interoperability with other TLS implementations since these might
8222 be using the same extension number for other purposes.
8224 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8225 opaque PRF input value to use in the handshake. This will create
8226 an internal copy of the length-'len' string at 'src', and will
8227 return non-zero for success.
8229 To get more control and flexibility, provide a callback function
8232 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8233 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8237 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8240 Callback function 'cb' will be called in handshakes, and is
8241 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8242 Argument 'arg' is for application purposes (the value as given to
8243 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8244 be provided to the callback function). The callback function
8245 has to return non-zero to report success: usually 1 to use opaque
8246 PRF input just if possible, or 2 to enforce use of the opaque PRF
8247 input. In the latter case, the library will abort the handshake
8248 if opaque PRF input is not successfully negotiated.
8250 Arguments 'peerinput' and 'len' given to the callback function
8251 will always be NULL and 0 in the case of a client. A server will
8252 see the client's opaque PRF input through these variables if
8253 available (NULL and 0 otherwise). Note that if the server
8254 provides an opaque PRF input, the length must be the same as the
8255 length of the client's opaque PRF input.
8257 Note that the callback function will only be called when creating
8258 a new session (session resumption can resume whatever was
8259 previously negotiated), and will not be called in SSL 2.0
8260 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8261 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8262 for applications that need to enforce opaque PRF input.
8266 * Update ssl code to support digests other than SHA1+MD5 for handshake
8269 *Victor B. Wagner <vitus@cryptocom.ru>*
8271 * Add RFC4507 support to OpenSSL. This includes the corrections in
8272 RFC4507bis. The encrypted ticket format is an encrypted encoded
8273 SSL_SESSION structure, that way new session features are automatically
8276 If a client application caches session in an SSL_SESSION structure
8277 support is transparent because tickets are now stored in the encoded
8280 The SSL_CTX structure automatically generates keys for ticket
8281 protection in servers so again support should be possible
8282 with no application modification.
8284 If a client or server wishes to disable RFC4507 support then the option
8285 SSL_OP_NO_TICKET can be set.
8287 Add a TLS extension debugging callback to allow the contents of any client
8288 or server extensions to be examined.
8290 This work was sponsored by Google.
8294 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8295 OpenSSL should now compile cleanly on gcc 4.2
8297 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8299 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8300 support including streaming MAC support: this is required for GOST
8301 ciphersuite support.
8303 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8305 * Add option -stream to use PKCS#7 streaming in smime utility. New
8306 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8307 to output in BER and PEM format.
8311 * Experimental support for use of HMAC via EVP_PKEY interface. This
8312 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8313 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8314 ENGINE support for HMAC keys which are unextractable. New -mac and
8315 -macopt options to dgst utility.
8319 * New option -sigopt to dgst utility. Update dgst to use
8320 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8321 alternative signing parameters such as X9.31 or PSS in the dgst
8326 * Change ssl_cipher_apply_rule(), the internal function that does
8327 the work each time a ciphersuite string requests enabling
8328 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8329 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8330 the order of disabled ciphersuites such that those ciphersuites
8331 that most recently went from enabled to disabled not only stay
8332 in order with respect to each other, but also have higher priority
8333 than other disabled ciphersuites the next time ciphersuites are
8336 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8337 the same ciphersuites as with "HIGH" alone, but in a specific
8338 order where the PSK ciphersuites come first (since they are the
8339 most recently disabled ciphersuites when "HIGH" is parsed).
8341 Also, change ssl_create_cipher_list() (using this new
8342 functionality) such that between otherwise identical
8343 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8348 * Change ssl_create_cipher_list() so that it automatically
8349 arranges the ciphersuites in reasonable order before starting
8350 to process the rule string. Thus, the definition for "DEFAULT"
8351 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8352 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8353 This makes it much easier to arrive at a reasonable default order
8354 in applications for which anonymous ciphers are OK (meaning
8355 that you can't actually use DEFAULT).
8357 *Bodo Moeller; suggested by Victor Duchovni*
8359 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8360 processing) into multiple integers instead of setting
8361 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8362 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8363 (These masks as well as the individual bit definitions are hidden
8364 away into the non-exported interface ssl/ssl_locl.h, so this
8365 change to the definition of the SSL_CIPHER structure shouldn't
8366 affect applications.) This give us more bits for each of these
8367 categories, so there is no longer a need to coagulate AES128 and
8368 AES256 into a single algorithm bit, and to coagulate Camellia128
8369 and Camellia256 into a single algorithm bit, which has led to all
8372 Thus, among other things, the kludge introduced in 0.9.7m and
8373 0.9.8e for masking out AES256 independently of AES128 or masking
8374 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8376 With the change, we also introduce new ciphersuite aliases that
8377 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8382 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8383 Use the leftmost N bytes of the signature input if the input is
8384 larger than the prime q (with N being the size in bytes of q).
8388 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8389 it yet and it is largely untested.
8393 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8397 * Initial incomplete changes to avoid need for function casts in OpenSSL
8398 some compilers (gcc 4.2 and later) reject their use. Safestack is
8399 reimplemented. Update ASN1 to avoid use of legacy functions.
8403 * Win32/64 targets are linked with Winsock2.
8407 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8408 to external functions. This can be used to increase CRL handling
8409 efficiency especially when CRLs are very large by (for example) storing
8410 the CRL revoked certificates in a database.
8414 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8415 new CRLs added to a directory can be used. New command line option
8416 -verify_return_error to s_client and s_server. This causes real errors
8417 to be returned by the verify callback instead of carrying on no matter
8418 what. This reflects the way a "real world" verify callback would behave.
8422 * GOST engine, supporting several GOST algorithms and public key formats.
8423 Kindly donated by Cryptocom.
8427 * Partial support for Issuing Distribution Point CRL extension. CRLs
8428 partitioned by DP are handled but no indirect CRL or reason partitioning
8429 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8430 selected via a scoring technique which handles IDP and AKID in CRLs.
8434 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8435 will ultimately be used for all verify operations: this will remove the
8436 X509_STORE dependency on certificate verification and allow alternative
8437 lookup methods. X509_STORE based implementations of these two callbacks.
8441 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8442 Modify get_crl() to find a valid (unexpired) CRL if possible.
8446 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8447 this would be called X509_CRL_cmp() but that name is already used by
8448 a function that just compares CRL issuer names. Cache several CRL
8449 extensions in X509_CRL structure and cache CRLDP in X509.
8453 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8454 this maps equivalent X509_NAME structures into a consistent structure.
8455 Name comparison can then be performed rapidly using memcmp().
8459 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8464 * Allow digests to supply their own micalg string for S/MIME type using
8465 the ctrl EVP_MD_CTRL_MICALG.
8469 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8470 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8471 ctrl. It can then customise the structure before and/or after signing
8476 * New function OBJ_add_sigid() to allow application defined signature OIDs
8477 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8478 to free up any added signature OIDs.
8482 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8483 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8484 digest and cipher tables. New options added to openssl utility:
8485 list-message-digest-algorithms and list-cipher-algorithms.
8489 * Change the array representation of binary polynomials: the list
8490 of degrees of non-zero coefficients is now terminated with -1.
8491 Previously it was terminated with 0, which was also part of the
8492 value; thus, the array representation was not applicable to
8493 polynomials where t^0 has coefficient zero. This change makes
8494 the array representation useful in a more general context.
8498 * Various modifications and fixes to SSL/TLS cipher string
8499 handling. For ECC, the code now distinguishes between fixed ECDH
8500 with RSA certificates on the one hand and with ECDSA certificates
8501 on the other hand, since these are separate ciphersuites. The
8502 unused code for Fortezza ciphersuites has been removed.
8504 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8505 (not "ECDHE"). For consistency with the code for DH
8506 certificates, use of ECDH certificates is now considered ECDH
8507 authentication, not RSA or ECDSA authentication (the latter is
8508 merely the CA's signing algorithm and not actively used in the
8511 The temporary ciphersuite alias "ECCdraft" is no longer
8512 available, and ECC ciphersuites are no longer excluded from "ALL"
8513 and "DEFAULT". The following aliases now exist for RFC 4492
8514 ciphersuites, most of these by analogy with the DH case:
8516 kECDHr - ECDH cert, signed with RSA
8517 kECDHe - ECDH cert, signed with ECDSA
8518 kECDH - ECDH cert (signed with either RSA or ECDSA)
8519 kEECDH - ephemeral ECDH
8520 ECDH - ECDH cert or ephemeral ECDH
8526 AECDH - anonymous ECDH
8527 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8531 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8532 Use correct micalg parameters depending on digest(s) in signed message.
8536 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8537 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8541 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8542 an engine to register a method. Add ENGINE lookups for methods and
8543 functional reference processing.
8547 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8548 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8553 * New -resign option to smime utility. This adds one or more signers
8554 to an existing PKCS#7 signedData structure. Also -md option to use an
8555 alternative message digest algorithm for signing.
8559 * Tidy up PKCS#7 routines and add new functions to make it easier to
8560 create PKCS7 structures containing multiple signers. Update smime
8561 application to support multiple signers.
8565 * New -macalg option to pkcs12 utility to allow setting of an alternative
8570 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8571 Reorganize PBE internals to lookup from a static table using NIDs,
8572 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8573 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8574 PRF which will be automatically used with PBES2.
8578 * Replace the algorithm specific calls to generate keys in "req" with the
8583 * Update PKCS#7 enveloped data routines to use new API. This is now
8584 supported by any public key method supporting the encrypt operation. A
8585 ctrl is added to allow the public key algorithm to examine or modify
8586 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8591 * Add a ctrl to asn1 method to allow a public key algorithm to express
8592 a default digest type to use. In most cases this will be SHA1 but some
8593 algorithms (such as GOST) need to specify an alternative digest. The
8594 return value indicates how strong the preference is 1 means optional and
8595 2 is mandatory (that is it is the only supported type). Modify
8596 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8597 use the default md. Update openssl utilities to use the default digest
8598 type for signing if it is not explicitly indicated.
8602 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8603 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8604 signing method from the key type. This effectively removes the link
8605 between digests and public key types.
8609 * Add an OID cross reference table and utility functions. Its purpose is to
8610 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8611 rsaEncryption. This will allow some of the algorithm specific hackery
8612 needed to use the correct OID to be removed.
8616 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8617 structures for PKCS7_sign(). They are now set up by the relevant public
8622 * Add provisional EC pkey method with support for ECDSA and ECDH.
8626 * Add support for key derivation (agreement) in the API, DH method and
8631 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8632 public and private key formats. As a side effect these add additional
8633 command line functionality not previously available: DSA signatures can be
8634 generated and verified using pkeyutl and DH key support and generation in
8641 *Oliver Tappe <zooey@hirschkaefer.de>*
8643 * New make target "install_html_docs" installs HTML renditions of the
8646 *Oliver Tappe <zooey@hirschkaefer.de>*
8648 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8649 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8650 support key and parameter generation and add initial key generation
8651 functionality for RSA.
8655 * Add functions for main EVP_PKEY_method operations. The undocumented
8656 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8657 `EVP_PKEY_{encrypt,decrypt}_old`.
8661 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8662 key API, doesn't do much yet.
8666 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8667 public key algorithms. New option to openssl utility:
8668 "list-public-key-algorithms" to print out info.
8672 * Implement the Supported Elliptic Curves Extension for
8673 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8677 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8678 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8682 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8683 utilities such as rsa, dsa, dsaparam etc except they process any key
8688 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8689 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8690 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8695 * Initial support for pluggable public key ASN1.
8696 De-spaghettify the public key ASN1 handling. Move public and private
8697 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8698 algorithm specific handling to a single module within the relevant
8699 algorithm directory. Add functions to allow (near) opaque processing
8700 of public and private key structures.
8704 * Implement the Supported Point Formats Extension for
8705 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8709 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8710 for the psk identity [hint] and the psk callback functions to the
8711 SSL_SESSION, SSL and SSL_CTX structure.
8714 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8718 SSL_CTX_use_psk_identity_hint
8719 SSL_get_psk_identity_hint
8720 SSL_get_psk_identity
8721 SSL_use_psk_identity_hint
8723 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8725 * Add RFC 3161 compliant time stamp request creation, response generation
8726 and response verification functionality.
8728 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8730 * Add initial support for TLS extensions, specifically for the server_name
8731 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8732 have new members for a host name. The SSL data structure has an
8733 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8734 stored in that context to allow for session resumption, even after the
8735 SSL has been switched to a new SSL_CTX in reaction to a client's
8736 server_name extension.
8738 New functions (subject to change):
8740 SSL_get_servername()
8741 SSL_get_servername_type()
8744 New CTRL codes and macros (subject to change):
8746 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8747 - SSL_CTX_set_tlsext_servername_callback()
8748 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8749 - SSL_CTX_set_tlsext_servername_arg()
8750 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8752 openssl s_client has a new '-servername ...' option.
8754 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8755 '-key2 ...', '-servername_fatal' (subject to change). This allows
8756 testing the HostName extension for a specific single host name ('-cert'
8757 and '-key' remain fallbacks for handshakes without HostName
8758 negotiation). If the unrecognized_name alert has to be sent, this by
8759 default is a warning; it becomes fatal with the '-servername_fatal'
8762 *Peter Sylvester, Remy Allais, Christophe Renou*
8764 * Whirlpool hash implementation is added.
8768 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8769 bn(64,32). Because of instruction set limitations it doesn't have
8770 any negative impact on performance. This was done mostly in order
8771 to make it possible to share assembler modules, such as bn_mul_mont
8772 implementations, between 32- and 64-bit builds without hassle.
8776 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8777 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8782 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8783 dedicated Montgomery multiplication procedure, is introduced.
8784 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8785 "64-bit" performance on certain 32-bit targets.
8789 * New option SSL_OP_NO_COMP to disable use of compression selectively
8790 in SSL structures. New SSL ctrl to set maximum send fragment size.
8791 Save memory by setting the I/O buffer sizes dynamically instead of
8792 using the maximum available value.
8796 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8797 in addition to the text details.
8801 * Very, very preliminary EXPERIMENTAL support for printing of general
8802 ASN1 structures. This currently produces rather ugly output and doesn't
8803 handle several customised structures at all.
8807 * Integrated support for PVK file format and some related formats such
8808 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8809 these in the 'rsa' and 'dsa' utilities.
8813 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8817 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8818 place for the (very old) "NETSCAPE" format certificates which are now
8819 handled using new ASN1 code equivalents.
8823 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8824 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8825 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8829 * Modify CRL distribution points extension code to print out previously
8830 unsupported fields. Enhance extension setting code to allow setting of
8835 * Add print and set support for Issuing Distribution Point CRL extension.
8839 * Change 'Configure' script to enable Camellia by default.
8846 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8848 * When rejecting SSL/TLS records due to an incorrect version number, never
8849 update s->server with a new major version number. As of
8850 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8851 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8852 the previous behavior could result in a read attempt at NULL when
8853 receiving specific incorrect SSL/TLS records once record payload
8854 protection is active. ([CVE-2010-0740])
8856 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8858 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8859 could be crashed if the relevant tables were not present (e.g. chrooted).
8861 *Tomas Hoger <thoger@redhat.com>*
8863 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8865 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8867 *Martin Olsson, Neel Mehta*
8869 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8870 accommodate for stack sorting, always a write lock!).
8874 * On some versions of WIN32 Heap32Next is very slow. This can cause
8875 excessive delays in the RAND_poll(): over a minute. As a workaround
8876 include a time check in the inner Heap32Next loop too.
8880 * The code that handled flushing of data in SSL/TLS originally used the
8881 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8882 the problem outlined in PR#1949. The fix suggested there however can
8883 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8884 of Apache). So instead simplify the code to flush unconditionally.
8885 This should be fine since flushing with no data to flush is a no op.
8889 * Handle TLS versions 2.0 and later properly and correctly use the
8890 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8891 off ancient servers have a habit of sticking around for a while...
8895 * Modify compression code so it frees up structures without using the
8896 ex_data callbacks. This works around a problem where some applications
8897 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8898 restarting) then use compression (e.g. SSL with compression) later.
8899 This results in significant per-connection memory leaks and
8900 has caused some security issues including CVE-2008-1678 and
8905 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8906 change when encrypting or decrypting.
8910 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8911 connect and renegotiate with servers which do not support RI.
8912 Until RI is more widely deployed this option is enabled by default.
8916 * Add "missing" ssl ctrls to clear options and mode.
8920 * If client attempts to renegotiate and doesn't support RI respond with
8921 a no_renegotiation alert as required by RFC5746. Some renegotiating
8922 TLS clients will continue a connection gracefully when they receive
8923 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8924 waiting for a server hello which it will never receive. Now we treat a
8925 received no_renegotiation alert as a fatal error. This is because
8926 applications requesting a renegotiation might well expect it to succeed
8927 and would have no code in place to handle the server denying it so the
8928 only safe thing to do is to terminate the connection.
8932 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8933 peer supports secure renegotiation and 0 otherwise. Print out peer
8934 renegotiation support in s_client/s_server.
8938 * Replace the highly broken and deprecated SPKAC certification method with
8939 the updated NID creation version. This should correctly handle UTF8.
8943 * Implement RFC5746. Re-enable renegotiation but require the extension
8944 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8945 turns out to be a bad idea. It has been replaced by
8946 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8947 SSL_CTX_set_options(). This is really not recommended unless you
8948 know what you are doing.
8950 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8952 * Fixes to stateless session resumption handling. Use initial_ctx when
8953 issuing and attempting to decrypt tickets in case it has changed during
8954 servername handling. Use a non-zero length session ID when attempting
8955 stateless session resumption: this makes it possible to determine if
8956 a resumption has occurred immediately after receiving server hello
8957 (several places in OpenSSL subtly assume this) instead of later in
8962 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8963 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8964 fixes for a few places where the return code is not checked
8967 *Julia Lawall <julia@diku.dk>*
8969 * Add --strict-warnings option to Configure script to include devteam
8970 warnings in other configurations.
8974 * Add support for --libdir option and LIBDIR variable in makefiles. This
8975 makes it possible to install openssl libraries in locations which
8976 have names other than "lib", for example "/usr/lib64" which some
8979 *Steve Henson, based on patch from Jeremy Utley*
8981 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8982 X690 8.9.12 and can produce some misleading textual output of OIDs.
8984 *Steve Henson, reported by Dan Kaminsky*
8986 * Delete MD2 from algorithm tables. This follows the recommendation in
8987 several standards that it is not used in new applications due to
8988 several cryptographic weaknesses. For binary compatibility reasons
8989 the MD2 API is still compiled in by default.
8993 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8998 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8999 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9002 *Guenter <lists@gknw.net>*
9004 * Fix the server certificate chain building code to use X509_verify_cert(),
9005 it used to have an ad-hoc builder which was unable to cope with anything
9006 other than a simple chain.
9008 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9010 * Don't check self signed certificate signatures in X509_verify_cert()
9011 by default (a flag can override this): it just wastes time without
9012 adding any security. As a useful side effect self signed root CAs
9013 with non-FIPS digests are now usable in FIPS mode.
9017 * In dtls1_process_out_of_seq_message() the check if the current message
9018 is already buffered was missing. For every new message was memory
9019 allocated, allowing an attacker to perform an denial of service attack
9020 with sending out of seq handshake messages until there is no memory
9021 left. Additionally every future message was buffered, even if the
9022 sequence number made no sense and would be part of another handshake.
9023 So only messages with sequence numbers less than 10 in advance will be
9024 buffered. ([CVE-2009-1378])
9026 *Robin Seggelmann, discovered by Daniel Mentz*
9028 * Records are buffered if they arrive with a future epoch to be
9029 processed after finishing the corresponding handshake. There is
9030 currently no limitation to this buffer allowing an attacker to perform
9031 a DOS attack with sending records with future epochs until there is no
9032 memory left. This patch adds the pqueue_size() function to determine
9033 the size of a buffer and limits the record buffer to 100 entries.
9036 *Robin Seggelmann, discovered by Daniel Mentz*
9038 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9039 parent structure is freed. ([CVE-2009-1379])
9043 * Handle non-blocking I/O properly in SSL_shutdown() call.
9045 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9047 * Add `2.5.4.*` OIDs
9049 *Ilya O. <vrghost@gmail.com>*
9051 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9053 * Disable renegotiation completely - this fixes a severe security
9054 problem ([CVE-2009-3555]) at the cost of breaking all
9055 renegotiation. Renegotiation can be re-enabled by setting
9056 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9057 run-time. This is really not recommended unless you know what
9062 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9064 * Don't set val to NULL when freeing up structures, it is freed up by
9065 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9066 zeroing past the valid field. ([CVE-2009-0789])
9068 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9070 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9071 checked correctly. This would allow some invalid signed attributes to
9072 appear to verify correctly. ([CVE-2009-0591])
9074 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9076 * Reject UniversalString and BMPString types with invalid lengths. This
9077 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9078 a legal length. ([CVE-2009-0590])
9082 * Set S/MIME signing as the default purpose rather than setting it
9083 unconditionally. This allows applications to override it at the store
9088 * Permit restricted recursion of ASN1 strings. This is needed in practice
9089 to handle some structures.
9093 * Improve efficiency of mem_gets: don't search whole buffer each time
9096 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9098 * New -hex option for openssl rand.
9102 * Print out UTF8String and NumericString when parsing ASN1.
9106 * Support NumericString type for name components.
9110 * Allow CC in the environment to override the automatically chosen
9111 compiler. Note that nothing is done to ensure flags work with the
9116 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9118 * Properly check EVP_VerifyFinal() and similar return values
9121 *Ben Laurie, Bodo Moeller, Google Security Team*
9123 * Enable TLS extensions by default.
9127 * Allow the CHIL engine to be loaded, whether the application is
9128 multithreaded or not. (This does not release the developer from the
9129 obligation to set up the dynamic locking callbacks.)
9131 *Sander Temme <sander@temme.net>*
9133 * Use correct exit code if there is an error in dgst command.
9135 *Steve Henson; problem pointed out by Roland Dirlewanger*
9137 * Tweak Configure so that you need to say "experimental-jpake" to enable
9138 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9142 * Add experimental JPAKE support, including demo authentication in
9143 s_client and s_server.
9147 * Set the comparison function in v3_addr_canonize().
9149 *Rob Austein <sra@hactrn.net>*
9151 * Add support for XMPP STARTTLS in s_client.
9153 *Philip Paeps <philip@freebsd.org>*
9155 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9156 to ensure that even with this option, only ciphersuites in the
9157 server's preference list will be accepted. (Note that the option
9158 applies only when resuming a session, so the earlier behavior was
9159 just about the algorithm choice for symmetric cryptography.)
9163 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9165 * Fix NULL pointer dereference if a DTLS server received
9166 ChangeCipherSpec as first record ([CVE-2009-1386]).
9170 * Fix a state transition in s3_srvr.c and d1_srvr.c
9171 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9175 * The fix in 0.9.8c that supposedly got rid of unsafe
9176 double-checked locking was incomplete for RSA blinding,
9177 addressing just one layer of what turns out to have been
9178 doubly unsafe triple-checked locking.
9180 So now fix this for real by retiring the MONT_HELPER macro
9181 in crypto/rsa/rsa_eay.c.
9183 *Bodo Moeller; problem pointed out by Marius Schilder*
9185 * Various precautionary measures:
9187 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9189 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9190 (NB: This would require knowledge of the secret session ticket key
9191 to exploit, in which case you'd be SOL either way.)
9193 - Change bn_nist.c so that it will properly handle input BIGNUMs
9194 outside the expected range.
9196 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9199 *Neel Mehta, Bodo Moeller*
9201 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9202 the load fails. Useful for distros.
9204 *Ben Laurie and the FreeBSD team*
9206 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9210 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9214 * Expand ENGINE to support engine supplied SSL client certificate functions.
9216 This work was sponsored by Logica.
9220 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9221 keystores. Support for SSL/TLS client authentication too.
9222 Not compiled unless enable-capieng specified to Configure.
9224 This work was sponsored by Logica.
9228 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9229 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9230 attribute creation routines such as certificate requests and PKCS#12
9235 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9237 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9238 handshake which could lead to a client crash as found using the
9239 Codenomicon TLS test suite ([CVE-2008-1672])
9241 *Steve Henson, Mark Cox*
9243 * Fix double free in TLS server name extensions which could lead to
9244 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9248 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9250 Clear the error queue to ensure that error entries left from
9251 older function calls do not interfere with the correct operation.
9253 *Lutz Jaenicke, Erik de Castro Lopo*
9255 * Remove root CA certificates of commercial CAs:
9257 The OpenSSL project does not recommend any specific CA and does not
9258 have any policy with respect to including or excluding any CA.
9259 Therefore it does not make any sense to ship an arbitrary selection
9260 of root CA certificates with the OpenSSL software.
9264 * RSA OAEP patches to fix two separate invalid memory reads.
9265 The first one involves inputs when 'lzero' is greater than
9266 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9267 before the beginning of from). The second one involves inputs where
9268 the 'db' section contains nothing but zeroes (there is a one-byte
9269 invalid read after the end of 'db').
9271 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9273 * Partial backport from 0.9.9-dev:
9275 Introduce bn_mul_mont (dedicated Montgomery multiplication
9276 procedure) as a candidate for BIGNUM assembler implementation.
9277 While 0.9.9-dev uses assembler for various architectures, only
9278 x86_64 is available by default here in the 0.9.8 branch, and
9279 32-bit x86 is available through a compile-time setting.
9281 To try the 32-bit x86 assembler implementation, use Configure
9282 option "enable-montasm" (which exists only for this backport).
9284 As "enable-montasm" for 32-bit x86 disclaims code stability
9285 anyway, in this constellation we activate additional code
9286 backported from 0.9.9-dev for further performance improvements,
9287 namely BN_from_montgomery_word. (To enable this otherwise,
9288 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9290 *Andy Polyakov (backport partially by Bodo Moeller)*
9292 * Add TLS session ticket callback. This allows an application to set
9293 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9294 values. This is useful for key rollover for example where several key
9295 sets may exist with different names.
9299 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9300 This was broken until now in 0.9.8 releases, such that the only way
9301 a registered ENGINE could be used (assuming it initialises
9302 successfully on the host) was to explicitly set it as the default
9303 for the relevant algorithms. This is in contradiction with 0.9.7
9304 behaviour and the documentation. With this fix, when an ENGINE is
9305 registered into a given algorithm's table of implementations, the
9306 'uptodate' flag is reset so that auto-discovery will be used next
9307 time a new context for that algorithm attempts to select an
9310 *Ian Lister (tweaked by Geoff Thorpe)*
9312 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9313 implementation in the following ways:
9315 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9318 Lack of BER streaming support means one pass streaming processing is
9319 only supported if data is detached: setting the streaming flag is
9320 ignored for embedded content.
9322 CMS support is disabled by default and must be explicitly enabled
9323 with the enable-cms configuration option.
9327 * Update the GMP engine glue to do direct copies between BIGNUM and
9328 mpz_t when openssl and GMP use the same limb size. Otherwise the
9329 existing "conversion via a text string export" trick is still used.
9331 *Paul Sheer <paulsheer@gmail.com>*
9333 * Zlib compression BIO. This is a filter BIO which compressed and
9334 uncompresses any data passed through it.
9338 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9339 RFC3394 compatible AES key wrapping.
9343 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9344 sets string data without copying. X509_ALGOR_set0() and
9345 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9346 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9347 from an X509_ATTRIBUTE structure optionally checking it occurs only
9348 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9353 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9354 to get the expected BN_FLG_CONSTTIME behavior.
9356 *Bodo Moeller (Google)*
9360 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9361 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9362 - added some more tests to do_tests.pl
9363 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9364 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9365 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9366 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9367 - various changes to netware.pl to enable gcc-cross builds on Win32
9369 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9370 - various changes to fix missing prototype warnings
9371 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9372 - added AES, WHIRLPOOL and CPUID assembler code to build files
9373 - added missing AES assembler make rules to mk1mf.pl
9374 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9376 *Guenter Knauf <eflash@gmx.net>*
9378 * Implement certificate status request TLS extension defined in RFC3546.
9379 A client can set the appropriate parameters and receive the encoded
9380 OCSP response via a callback. A server can query the supplied parameters
9381 and set the encoded OCSP response in the callback. Add simplified examples
9382 to s_client and s_server.
9386 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9389 + Binary incompatibility of ssl_ctx_st structure
9390 + DTLS interoperation with non-compliant servers
9391 + Don't call get_session_cb() without proposed session
9392 + Fix ia64 assembler code
9394 *Andy Polyakov, Steve Henson*
9396 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9398 * DTLS Handshake overhaul. There were longstanding issues with
9399 OpenSSL DTLS implementation, which were making it impossible for
9400 RFC 4347 compliant client to communicate with OpenSSL server.
9401 Unfortunately just fixing these incompatibilities would "cut off"
9402 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9403 server keeps tolerating non RFC compliant syntax. The opposite is
9404 not true, 0.9.8f client can not communicate with earlier server.
9405 This update even addresses CVE-2007-4995.
9409 * Changes to avoid need for function casts in OpenSSL: some compilers
9410 (gcc 4.2 and later) reject their use.
9411 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9414 * Add RFC4507 support to OpenSSL. This includes the corrections in
9415 RFC4507bis. The encrypted ticket format is an encrypted encoded
9416 SSL_SESSION structure, that way new session features are automatically
9419 If a client application caches session in an SSL_SESSION structure
9420 support is transparent because tickets are now stored in the encoded
9423 The SSL_CTX structure automatically generates keys for ticket
9424 protection in servers so again support should be possible
9425 with no application modification.
9427 If a client or server wishes to disable RFC4507 support then the option
9428 SSL_OP_NO_TICKET can be set.
9430 Add a TLS extension debugging callback to allow the contents of any client
9431 or server extensions to be examined.
9433 This work was sponsored by Google.
9437 * Add initial support for TLS extensions, specifically for the server_name
9438 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9439 have new members for a host name. The SSL data structure has an
9440 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9441 stored in that context to allow for session resumption, even after the
9442 SSL has been switched to a new SSL_CTX in reaction to a client's
9443 server_name extension.
9445 New functions (subject to change):
9447 SSL_get_servername()
9448 SSL_get_servername_type()
9451 New CTRL codes and macros (subject to change):
9453 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9454 - SSL_CTX_set_tlsext_servername_callback()
9455 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9456 - SSL_CTX_set_tlsext_servername_arg()
9457 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9459 openssl s_client has a new '-servername ...' option.
9461 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9462 '-key2 ...', '-servername_fatal' (subject to change). This allows
9463 testing the HostName extension for a specific single host name ('-cert'
9464 and '-key' remain fallbacks for handshakes without HostName
9465 negotiation). If the unrecognized_name alert has to be sent, this by
9466 default is a warning; it becomes fatal with the '-servername_fatal'
9469 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9471 * Add AES and SSE2 assembly language support to VC++ build.
9475 * Mitigate attack on final subtraction in Montgomery reduction.
9479 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9480 (which previously caused an internal error).
9484 * Squeeze another 10% out of IGE mode when in != out.
9488 * AES IGE mode speedup.
9490 *Dean Gaudet (Google)*
9492 * Add the Korean symmetric 128-bit cipher SEED (see
9493 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9494 add SEED ciphersuites from RFC 4162:
9496 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9497 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9498 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9499 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9501 To minimize changes between patchlevels in the OpenSSL 0.9.8
9502 series, SEED remains excluded from compilation unless OpenSSL
9503 is configured with 'enable-seed'.
9505 *KISA, Bodo Moeller*
9507 * Mitigate branch prediction attacks, which can be practical if a
9508 single processor is shared, allowing a spy process to extract
9509 information. For detailed background information, see
9510 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9511 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9512 and Necessary Software Countermeasures"). The core of the change
9513 are new versions BN_div_no_branch() and
9514 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9515 respectively, which are slower, but avoid the security-relevant
9516 conditional branches. These are automatically called by BN_div()
9517 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9518 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9519 remove a conditional branch.
9521 BN_FLG_CONSTTIME is the new name for the previous
9522 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9523 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9524 in the exponent causes BN_mod_exp_mont() to use the alternative
9525 implementation in BN_mod_exp_mont_consttime().) The old name
9526 remains as a deprecated alias.
9528 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9529 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9530 constant-time implementations for more than just exponentiation.
9531 Here too the old name is kept as a deprecated alias.
9533 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9534 the BN_BLINDING structure gets an independent copy of the
9535 modulus. This means that the previous `BIGNUM *m` argument to
9536 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9537 essentially becomes `const BIGNUM *m`, although we can't actually
9538 change this in the header file before 0.9.9. It allows
9539 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9540 enable BN_FLG_CONSTTIME.
9542 *Matthew D Wood (Intel Corp)*
9544 * In the SSL/TLS server implementation, be strict about session ID
9545 context matching (which matters if an application uses a single
9546 external cache for different purposes). Previously,
9547 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9548 set. This did ensure strict client verification, but meant that,
9549 with applications using a single external cache for quite
9550 different requirements, clients could circumvent ciphersuite
9551 restrictions for a given session ID context by starting a session
9552 in a different context.
9556 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9557 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9558 authentication-only ciphersuites.
9562 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9563 not complete and could lead to a possible single byte overflow
9564 ([CVE-2007-5135]) [Ben Laurie]
9566 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9568 * Since AES128 and AES256 (and similarly Camellia128 and
9569 Camellia256) share a single mask bit in the logic of
9570 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9571 kludge to work properly if AES128 is available and AES256 isn't
9572 (or if Camellia128 is available and Camellia256 isn't).
9576 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9577 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9578 When a point or a seed is encoded in a BIT STRING, we need to
9579 prevent the removal of trailing zero bits to get the proper DER
9580 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9581 of a NamedBitList, for which trailing 0 bits need to be removed.)
9585 * Have SSL/TLS server implementation tolerate "mismatched" record
9586 protocol version while receiving ClientHello even if the
9587 ClientHello is fragmented. (The server can't insist on the
9588 particular protocol version it has chosen before the ServerHello
9589 message has informed the client about his choice.)
9593 * Add RFC 3779 support.
9595 *Rob Austein for ARIN, Ben Laurie*
9597 * Load error codes if they are not already present instead of using a
9598 static variable. This allows them to be cleanly unloaded and reloaded.
9599 Improve header file function name parsing.
9603 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9604 or CAPABILITY handshake as required by RFCs.
9608 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9610 * Introduce limits to prevent malicious keys being able to
9611 cause a denial of service. ([CVE-2006-2940])
9613 *Steve Henson, Bodo Moeller*
9615 * Fix ASN.1 parsing of certain invalid structures that can result
9616 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9618 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9619 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9621 * Fix SSL client code which could crash if connecting to a
9622 malicious SSLv2 server. ([CVE-2006-4343])
9624 *Tavis Ormandy and Will Drewry, Google Security Team*
9626 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9627 match only those. Before that, "AES256-SHA" would be interpreted
9628 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9629 the same strength classification in 0.9.7h) as we currently only
9630 have a single AES bit in the ciphersuite description bitmap.
9631 That change, however, also applied to ciphersuite strings such as
9632 "RC4-MD5" that intentionally matched multiple ciphersuites --
9633 namely, SSL 2.0 ciphersuites in addition to the more common ones
9634 from SSL 3.0/TLS 1.0.
9636 So we change the selection algorithm again: Naming an explicit
9637 ciphersuite selects this one ciphersuite, and any other similar
9638 ciphersuite (same bitmap) from *other* protocol versions.
9639 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9640 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9642 Since SSL 2.0 does not have any ciphersuites for which the
9643 128/256 bit distinction would be relevant, this works for now.
9644 The proper fix will be to use different bits for AES128 and
9645 AES256, which would have avoided the problems from the beginning;
9646 however, bits are scarce, so we can only do this in a new release
9647 (not just a patchlevel) when we can change the SSL_CIPHER
9648 definition to split the single 'unsigned long mask' bitmap into
9649 multiple values to extend the available space.
9653 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9655 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9656 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9658 * Add AES IGE and biIGE modes.
9662 * Change the Unix randomness entropy gathering to use poll() when
9663 possible instead of select(), since the latter has some
9664 undesirable limitations.
9666 *Darryl Miles via Richard Levitte and Bodo Moeller*
9668 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9669 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9670 cannot be implicitly activated as part of, e.g., the "AES" alias.
9671 However, please upgrade to OpenSSL 0.9.9[-dev] for
9672 non-experimental use of the ECC ciphersuites to get TLS extension
9673 support, which is required for curve and point format negotiation
9674 to avoid potential handshake problems.
9678 * Disable rogue ciphersuites:
9680 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9681 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9682 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9684 The latter two were purportedly from
9685 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9688 Also deactivate the remaining ciphersuites from
9689 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9690 unofficial, and the ID has long expired.
9694 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9695 dual-core machines) and other potential thread-safety issues.
9699 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9700 versions), which is now available for royalty-free use
9701 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9702 Also, add Camellia TLS ciphersuites from RFC 4132.
9704 To minimize changes between patchlevels in the OpenSSL 0.9.8
9705 series, Camellia remains excluded from compilation unless OpenSSL
9706 is configured with 'enable-camellia'.
9710 * Disable the padding bug check when compression is in use. The padding
9711 bug check assumes the first packet is of even length, this is not
9712 necessarily true if compression is enabled and can result in false
9713 positives causing handshake failure. The actual bug test is ancient
9714 code so it is hoped that implementations will either have fixed it by
9715 now or any which still have the bug do not support compression.
9719 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9721 * When applying a cipher rule check to see if string match is an explicit
9722 cipher suite and only match that one cipher suite if it is.
9726 * Link in manifests for VC++ if needed.
9728 *Austin Ziegler <halostatue@gmail.com>*
9730 * Update support for ECC-based TLS ciphersuites according to
9731 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9732 TLS extensions, which are supported starting with the 0.9.9
9733 branch, not in the OpenSSL 0.9.8 branch).
9737 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9738 opaque EVP_CIPHER_CTX handling.
9742 * Fixes and enhancements to zlib compression code. We now only use
9743 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9744 to conform with the standards mentioned here:
9745 <http://www.zlib.net/DLL_FAQ.txt>
9746 Static zlib linking now works on Windows and the new --with-zlib-include
9747 --with-zlib-lib options to Configure can be used to supply the location
9748 of the headers and library. Gracefully handle case where zlib library
9753 * Several fixes and enhancements to the OID generation code. The old code
9754 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9755 handle numbers larger than ULONG_MAX, truncated printing and had a
9756 non standard OBJ_obj2txt() behaviour.
9760 * Add support for building of engines under engine/ as shared libraries
9761 under VC++ build system.
9765 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9766 Hopefully, we will not see any false combination of paths any more.
9770 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9772 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9773 (part of SSL_OP_ALL). This option used to disable the
9774 countermeasure against man-in-the-middle protocol-version
9775 rollback in the SSL 2.0 server implementation, which is a bad
9776 idea. ([CVE-2005-2969])
9778 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9779 for Information Security, National Institute of Advanced Industrial
9780 Science and Technology [AIST], Japan)*
9782 * Add two function to clear and return the verify parameter flags.
9786 * Keep cipherlists sorted in the source instead of sorting them at
9787 runtime, thus removing the need for a lock.
9791 * Avoid some small subgroup attacks in Diffie-Hellman.
9793 *Nick Mathewson and Ben Laurie*
9795 * Add functions for well-known primes.
9799 * Extended Windows CE support.
9801 *Satoshi Nakamura and Andy Polyakov*
9803 * Initialize SSL_METHOD structures at compile time instead of during
9804 runtime, thus removing the need for a lock.
9808 * Make PKCS7_decrypt() work even if no certificate is supplied by
9809 attempting to decrypt each encrypted key in turn. Add support to
9814 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9816 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9819 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9823 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9824 key into the same file any more.
9828 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9832 * Add -utf8 command line and config file option to 'ca'.
9834 *Stefan <stf@udoma.org*
9836 * Removed the macro des_crypt(), as it seems to conflict with some
9837 libraries. Use DES_crypt().
9841 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9842 involves renaming the source and generated shared-libs for
9843 both. The engines will accept the corrected or legacy ids
9844 ('ncipher' and '4758_cca' respectively) when binding. NB,
9845 this only applies when building 'shared'.
9847 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9849 * Add attribute functions to EVP_PKEY structure. Modify
9850 PKCS12_create() to recognize a CSP name attribute and
9851 use it. Make -CSP option work again in pkcs12 utility.
9855 * Add new functionality to the bn blinding code:
9856 - automatic re-creation of the BN_BLINDING parameters after
9857 a fixed number of uses (currently 32)
9858 - add new function for parameter creation
9859 - introduce flags to control the update behaviour of the
9860 BN_BLINDING parameters
9861 - hide BN_BLINDING structure
9862 Add a second BN_BLINDING slot to the RSA structure to improve
9863 performance when a single RSA object is shared among several
9868 * Add support for DTLS.
9870 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9872 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9873 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9877 * Remove buggy and incomplete DH cert support from
9878 ssl/ssl_rsa.c and ssl/s3_both.c
9882 * Use SHA-1 instead of MD5 as the default digest algorithm for
9883 the apps/openssl applications.
9887 * Compile clean with "-Wall -Wmissing-prototypes
9888 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9889 DEBUG_SAFESTACK must also be set.
9893 * Change ./Configure so that certain algorithms can be disabled by default.
9894 The new counterpiece to "no-xxx" is "enable-xxx".
9896 The patented RC5 and MDC2 algorithms will now be disabled unless
9897 "enable-rc5" and "enable-mdc2", respectively, are specified.
9899 (IDEA remains enabled despite being patented. This is because IDEA
9900 is frequently required for interoperability, and there is no license
9901 fee for non-commercial use. As before, "no-idea" can be used to
9902 avoid this algorithm.)
9906 * Add processing of proxy certificates (see RFC 3820). This work was
9907 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9908 EGEE (Enabling Grids for E-science in Europe).
9912 * RC4 performance overhaul on modern architectures/implementations, such
9913 as Intel P4, IA-64 and AMD64.
9917 * New utility extract-section.pl. This can be used specify an alternative
9918 section number in a pod file instead of having to treat each file as
9919 a separate case in Makefile. This can be done by adding two lines to the
9922 =for comment openssl_section:XXX
9924 The blank line is mandatory.
9928 * New arguments -certform, -keyform and -pass for s_client and s_server
9929 to allow alternative format key and certificate files and passphrase
9934 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9935 update associated structures and add various utility functions.
9937 Add new policy related verify parameters, include policy checking in
9938 standard verify code. Enhance 'smime' application with extra parameters
9939 to support policy checking and print out.
9943 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9944 Nehemiah processors. These extensions support AES encryption in hardware
9945 as well as RNG (though RNG support is currently disabled).
9947 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9949 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9953 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9955 *Andy Polyakov and a number of other people*
9957 * Improved PowerPC platform support. Most notably BIGNUM assembler
9958 implementation contributed by IBM.
9960 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9962 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9963 exponent rather than 'unsigned long'. There is a corresponding change to
9964 the new 'rsa_keygen' element of the RSA_METHOD structure.
9966 *Jelte Jansen, Geoff Thorpe*
9968 * Functionality for creating the initial serial number file is now
9969 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9971 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9972 number file to 1, which is bound to cause problems. To avoid
9973 the problems while respecting compatibility between different 0.9.7
9974 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9975 CA.pl for serial number initialization. With the new release 0.9.8,
9976 we can fix the problem directly in the 'ca' utility.)
9980 * Reduced header interdependencies by declaring more opaque objects in
9981 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9982 give fewer recursive includes, which could break lazy source code - so
9983 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9984 developers should define this symbol when building and using openssl to
9985 ensure they track the recommended behaviour, interfaces, [etc], but
9986 backwards-compatible behaviour prevails when this isn't defined.
9990 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9994 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9995 This will generate a random key of the appropriate length based on the
9996 cipher context. The EVP_CIPHER can provide its own random key generation
9997 routine to support keys of a specific form. This is used in the des and
9998 3des routines to generate a key of the correct parity. Update S/MIME
9999 code to use new functions and hence generate correct parity DES keys.
10000 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10001 valid (weak or incorrect parity).
10005 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10006 as looking them up. This is useful when the verified structure may contain
10007 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10008 present unless the new PKCS7_NO_CRL flag is asserted.
10012 * Extend ASN1 oid configuration module. It now additionally accepts the
10015 shortName = some long name, 1.2.3.4
10019 * Reimplemented the BN_CTX implementation. There is now no more static
10020 limitation on the number of variables it can handle nor the depth of the
10021 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10022 information can now expand as required, and rather than having a single
10023 static array of bignums, BN_CTX now uses a linked-list of such arrays
10024 allowing it to expand on demand whilst maintaining the usefulness of
10025 BN_CTX's "bundling".
10029 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10030 to allow all RSA operations to function using a single BN_CTX.
10034 * Preliminary support for certificate policy evaluation and checking. This
10035 is initially intended to pass the tests outlined in "Conformance Testing
10036 of Relying Party Client Certificate Path Processing Logic" v1.07.
10040 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10041 remained unused and not that useful. A variety of other little bignum
10042 tweaks and fixes have also been made continuing on from the audit (see
10047 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10048 associated ASN1, EVP and SSL functions and old ASN1 macros.
10052 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10053 and this should never fail. So the return value from the use of
10054 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10055 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10059 * BN_CTX_get() should return zero-valued bignums, providing the same
10060 initialised value as BN_new().
10062 *Geoff Thorpe, suggested by Ulf Möller*
10064 * Support for inhibitAnyPolicy certificate extension.
10068 * An audit of the BIGNUM code is underway, for which debugging code is
10069 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10070 is considered valid when processing BIGNUMs, and causes execution to
10071 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10072 further steps are taken to deliberately pollute unused data in BIGNUM
10073 structures to try and expose faulty code further on. For now, openssl will
10074 (in its default mode of operation) continue to tolerate the inconsistent
10075 forms that it has tolerated in the past, but authors and packagers should
10076 consider trying openssl and their own applications when compiled with
10077 these debugging symbols defined. It will help highlight potential bugs in
10078 their own code, and will improve the test coverage for OpenSSL itself. At
10079 some point, these tighter rules will become openssl's default to improve
10080 maintainability, though the assert()s and other overheads will remain only
10081 in debugging configurations. See bn.h for more details.
10083 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10085 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10086 that can only be obtained through BN_CTX_new() (which implicitly
10087 initialises it). The presence of this function only made it possible
10088 to overwrite an existing structure (and cause memory leaks).
10092 * Because of the callback-based approach for implementing LHASH as a
10093 template type, lh_insert() adds opaque objects to hash-tables and
10094 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10095 to clean up those corresponding objects before destroying the hash table
10096 (and losing the object pointers). So some over-zealous constifications in
10097 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10098 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10099 prototyped to have "const" restrictions on the object pointers they are
10100 given (and so aren't required to cast them away any more).
10104 * The tmdiff.h API was so ugly and minimal that our own timing utility
10105 (speed) prefers to use its own implementation. The two implementations
10106 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10107 its object type properly exposed (MS_TM) instead of casting to/from
10108 `char *`. This may still change yet if someone realises MS_TM and
10110 aren't necessarily the greatest nomenclatures - but this is what was used
10111 internally to the implementation so I've used that for now.
10115 * Ensure that deprecated functions do not get compiled when
10116 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10117 the self-tests were still using deprecated key-generation functions so
10118 these have been updated also.
10122 * Reorganise PKCS#7 code to separate the digest location functionality
10123 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10124 New function PKCS7_set_digest() to set the digest type for PKCS#7
10125 digestedData type. Add additional code to correctly generate the
10126 digestedData type and add support for this type in PKCS7 initialization
10131 * New function PKCS7_set0_type_other() this initializes a PKCS7
10132 structure of type "other".
10136 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10137 sure the loop does correctly stop and breaking ("division by zero")
10138 modulus operations are not performed. The (pre-generated) prime
10139 table crypto/bn/bn_prime.h was already correct, but it could not be
10140 re-generated on some platforms because of the "division by zero"
10141 situation in the script.
10143 *Ralf S. Engelschall*
10145 * Update support for ECC-based TLS ciphersuites according to
10146 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10147 SHA-1 now is only used for "small" curves (where the
10148 representation of a field element takes up to 24 bytes); for
10149 larger curves, the field element resulting from ECDH is directly
10150 used as premaster secret.
10152 *Douglas Stebila (Sun Microsystems Laboratories)*
10154 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10155 curve secp160r1 to the tests.
10157 *Douglas Stebila (Sun Microsystems Laboratories)*
10159 * Add the possibility to load symbols globally with DSO.
10161 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10163 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10164 control of the error stack.
10168 * Add support for STORE in ENGINE.
10172 * Add the STORE type. The intention is to provide a common interface
10173 to certificate and key stores, be they simple file-based stores, or
10174 HSM-type store, or LDAP stores, or...
10175 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10179 * Add a generic structure called OPENSSL_ITEM. This can be used to
10180 pass a list of arguments to any function as well as provide a way
10181 for a function to pass data back to the caller.
10185 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10186 works like BUF_strdup() but can be used to duplicate a portion of
10187 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10192 * Add the function sk_find_ex() which works like sk_find(), but will
10193 return an index to an element even if an exact match couldn't be
10194 found. The index is guaranteed to point at the element where the
10195 searched-for key would be inserted to preserve sorting order.
10199 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10200 takes an extra flags argument for optional functionality. Currently,
10201 the following flags are defined:
10203 OBJ_BSEARCH_VALUE_ON_NOMATCH
10204 This one gets OBJ_bsearch_ex() to return a pointer to the first
10205 element where the comparing function returns a negative or zero
10208 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10209 This one gets OBJ_bsearch_ex() to return a pointer to the first
10210 element where the comparing function returns zero. This is useful
10211 if there are more than one element where the comparing function
10216 * Make it possible to create self-signed certificates with 'openssl ca'
10217 in such a way that the self-signed certificate becomes part of the
10218 CA database and uses the same mechanisms for serial number generation
10219 as all other certificate signing. The new flag '-selfsign' enables
10220 this functionality. Adapt CA.sh and CA.pl.in.
10224 * Add functionality to check the public key of a certificate request
10225 against a given private. This is useful to check that a certificate
10226 request can be signed by that key (self-signing).
10230 * Make it possible to have multiple active certificates with the same
10231 subject in the CA index file. This is done only if the keyword
10232 'unique_subject' is set to 'no' in the main CA section (default
10233 if 'CA_default') of the configuration file. The value is saved
10234 with the database itself in a separate index attribute file,
10235 named like the index file with '.attr' appended to the name.
10239 * Generate multi-valued AVAs using '+' notation in config files for
10244 * Support for nameConstraints certificate extension.
10248 * Support for policyConstraints certificate extension.
10252 * Support for policyMappings certificate extension.
10256 * Make sure the default DSA_METHOD implementation only uses its
10257 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10258 and change its own handlers to be NULL so as to remove unnecessary
10259 indirection. This lets alternative implementations fallback to the
10260 default implementation more easily.
10264 * Support for directoryName in GeneralName related extensions
10269 * Make it possible to link applications using Makefile.shared.
10270 Make that possible even when linking against static libraries!
10274 * Support for single pass processing for S/MIME signing. This now
10275 means that S/MIME signing can be done from a pipe, in addition
10276 cleartext signing (multipart/signed type) is effectively streaming
10277 and the signed data does not need to be all held in memory.
10279 This is done with a new flag PKCS7_STREAM. When this flag is set
10280 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10281 is done after the data is output (and digests calculated) in
10282 SMIME_write_PKCS7().
10286 * Add full support for -rpath/-R, both in shared libraries and
10287 applications, at least on the platforms where it's known how
10292 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10293 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10294 will now compute a table of multiples of the generator that
10295 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10296 faster (notably in the case of a single point multiplication,
10297 scalar * generator).
10299 *Nils Larsch, Bodo Moeller*
10301 * IPv6 support for certificate extensions. The various extensions
10302 which use the IP:a.b.c.d can now take IPv6 addresses using the
10303 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10308 * Added an ENGINE that implements RSA by performing private key
10309 exponentiations with the GMP library. The conversions to and from
10310 GMP's mpz_t format aren't optimised nor are any montgomery forms
10311 cached, and on x86 it appears OpenSSL's own performance has caught up.
10312 However there are likely to be other architectures where GMP could
10313 provide a boost. This ENGINE is not built in by default, but it can be
10314 specified at Configure time and should be accompanied by the necessary
10315 linker additions, eg;
10316 ./config -DOPENSSL_USE_GMP -lgmp
10320 * "openssl engine" will not display ENGINE/DSO load failure errors when
10321 testing availability of engines with "-t" - the old behaviour is
10322 produced by increasing the feature's verbosity with "-tt".
10326 * ECDSA routines: under certain error conditions uninitialized BN objects
10327 could be freed. Solution: make sure initialization is performed early
10328 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10333 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10334 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10335 software implementations. For DSA and DH, parameter generation can
10336 also be overridden by providing the appropriate method callbacks.
10340 * Change the "progress" mechanism used in key-generation and
10341 primality testing to functions that take a new BN_GENCB pointer in
10342 place of callback/argument pairs. The new API functions have `_ex`
10343 postfixes and the older functions are reimplemented as wrappers for
10344 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10345 declarations of the old functions to help (graceful) attempts to
10346 migrate to the new functions. Also, the new key-generation API
10347 functions operate on a caller-supplied key-structure and return
10348 success/failure rather than returning a key or NULL - this is to
10349 help make "keygen" another member function of RSA_METHOD etc.
10351 Example for using the new callback interface:
10353 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10354 void *my_arg = ...;
10357 BN_GENCB_set(&my_cb, my_callback, my_arg);
10359 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10360 /* For the meaning of a, b in calls to my_callback(), see the
10361 * documentation of the function that calls the callback.
10362 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10363 * my_callback should return 1 if it wants BN_is_prime_ex()
10364 * to continue, or 0 to stop.
10369 * Change the ZLIB compression method to be stateful, and make it
10370 available to TLS with the number defined in
10371 draft-ietf-tls-compression-04.txt.
10375 * Add the ASN.1 structures and functions for CertificatePair, which
10376 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10378 CertificatePair ::= SEQUENCE {
10379 forward [0] Certificate OPTIONAL,
10380 reverse [1] Certificate OPTIONAL,
10381 -- at least one of the pair shall be present -- }
10383 Also implement the PEM functions to read and write certificate
10384 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10386 This needed to be defined, mostly for the sake of the LDAP
10387 attribute crossCertificatePair, but may prove useful elsewhere as
10392 * Make it possible to inhibit symlinking of shared libraries in
10393 Makefile.shared, for Cygwin's sake.
10397 * Extend the BIGNUM API by creating a function
10398 void BN_set_negative(BIGNUM *a, int neg);
10399 and a macro that behave like
10400 int BN_is_negative(const BIGNUM *a);
10402 to avoid the need to access 'a->neg' directly in applications.
10406 * Implement fast modular reduction for pseudo-Mersenne primes
10407 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10408 EC_GROUP_new_curve_GFp() will now automatically use this
10411 *Nils Larsch <nla@trustcenter.de>*
10413 * Add new lock type (CRYPTO_LOCK_BN).
10417 * Change the ENGINE framework to automatically load engines
10418 dynamically from specific directories unless they could be
10419 found to already be built in or loaded. Move all the
10420 current engines except for the cryptodev one to a new
10421 directory engines/.
10422 The engines in engines/ are built as shared libraries if
10423 the "shared" options was given to ./Configure or ./config.
10424 Otherwise, they are inserted in libcrypto.a.
10425 /usr/local/ssl/engines is the default directory for dynamic
10426 engines, but that can be overridden at configure time through
10427 the usual use of --prefix and/or --openssldir, and at run
10428 time with the environment variable OPENSSL_ENGINES.
10430 *Geoff Thorpe and Richard Levitte*
10432 * Add Makefile.shared, a helper makefile to build shared
10433 libraries. Adapt Makefile.org.
10437 * Add version info to Win32 DLLs.
10439 *Peter 'Luna' Runestig" <peter@runestig.com>*
10441 * Add new 'medium level' PKCS#12 API. Certificates and keys
10442 can be added using this API to created arbitrary PKCS#12
10443 files while avoiding the low level API.
10445 New options to PKCS12_create(), key or cert can be NULL and
10446 will then be omitted from the output file. The encryption
10447 algorithm NIDs can be set to -1 for no encryption, the mac
10448 iteration count can be set to 0 to omit the mac.
10450 Enhance pkcs12 utility by making the -nokeys and -nocerts
10451 options work when creating a PKCS#12 file. New option -nomac
10452 to omit the mac, NONE can be set for an encryption algorithm.
10453 New code is modified to use the enhanced PKCS12_create()
10454 instead of the low level API.
10458 * Extend ASN1 encoder to support indefinite length constructed
10459 encoding. This can output sequences tags and octet strings in
10460 this form. Modify pk7_asn1.c to support indefinite length
10461 encoding. This is experimental and needs additional code to
10462 be useful, such as an ASN1 bio and some enhanced streaming
10465 Extend template encode functionality so that tagging is passed
10466 down to the template encoder.
10470 * Let 'openssl req' fail if an argument to '-newkey' is not
10471 recognized instead of using RSA as a default.
10475 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10476 As these are not official, they are not included in "ALL";
10477 the "ECCdraft" ciphersuite group alias can be used to select them.
10479 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10481 * Add ECDH engine support.
10483 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10485 * Add ECDH in new directory crypto/ecdh/.
10487 *Douglas Stebila (Sun Microsystems Laboratories)*
10489 * Let BN_rand_range() abort with an error after 100 iterations
10490 without success (which indicates a broken PRNG).
10494 * Change BN_mod_sqrt() so that it verifies that the input value
10495 is really the square of the return value. (Previously,
10496 BN_mod_sqrt would show GIGO behaviour.)
10500 * Add named elliptic curves over binary fields from X9.62, SECG,
10501 and WAP/WTLS; add OIDs that were still missing.
10503 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10505 * Extend the EC library for elliptic curves over binary fields
10506 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10509 EC_GF2m_simple_method
10513 EC_GROUP_new_curve_GF2m
10514 EC_GROUP_set_curve_GF2m
10515 EC_GROUP_get_curve_GF2m
10516 EC_POINT_set_affine_coordinates_GF2m
10517 EC_POINT_get_affine_coordinates_GF2m
10518 EC_POINT_set_compressed_coordinates_GF2m
10520 Point compression for binary fields is disabled by default for
10521 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10524 As binary polynomials are represented as BIGNUMs, various members
10525 of the EC_GROUP and EC_POINT data structures can be shared
10526 between the implementations for prime fields and binary fields;
10527 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10528 are essentially identical to their `..._GFp` counterparts.
10529 (For simplicity, the `..._GFp` prefix has been dropped from
10530 various internal method names.)
10532 An internal 'field_div' method (similar to 'field_mul' and
10533 'field_sqr') has been added; this is used only for binary fields.
10535 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10537 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10538 through methods ('mul', 'precompute_mult').
10540 The generic implementations (now internally called 'ec_wNAF_mul'
10541 and 'ec_wNAF_precomputed_mult') remain the default if these
10542 methods are undefined.
10544 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10546 * New function EC_GROUP_get_degree, which is defined through
10547 EC_METHOD. For curves over prime fields, this returns the bit
10548 length of the modulus.
10550 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10552 * New functions EC_GROUP_dup, EC_POINT_dup.
10553 (These simply call ..._new and ..._copy).
10555 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10557 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10558 Polynomials are represented as BIGNUMs (where the sign bit is not
10559 used) in the following functions [macros]:
10562 BN_GF2m_sub [= BN_GF2m_add]
10563 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10564 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10565 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10567 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10568 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10569 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10570 BN_GF2m_cmp [= BN_ucmp]
10572 (Note that only the 'mod' functions are actually for fields GF(2^m).
10573 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10575 For some functions, an the irreducible polynomial defining a
10576 field can be given as an 'unsigned int[]' with strictly
10577 decreasing elements giving the indices of those bits that are set;
10578 i.e., p[] represents the polynomial
10579 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10581 p[0] > p[1] > ... > p[k] = 0.
10582 This applies to the following functions:
10585 BN_GF2m_mod_mul_arr
10586 BN_GF2m_mod_sqr_arr
10587 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10588 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10589 BN_GF2m_mod_exp_arr
10590 BN_GF2m_mod_sqrt_arr
10591 BN_GF2m_mod_solve_quad_arr
10595 Conversion can be performed by the following functions:
10600 bntest.c has additional tests for binary polynomial arithmetic.
10602 Two implementations for BN_GF2m_mod_div() are available.
10603 The default algorithm simply uses BN_GF2m_mod_inv() and
10604 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10605 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10606 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10608 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10610 * Add new error code 'ERR_R_DISABLED' that can be used when some
10611 functionality is disabled at compile-time.
10613 *Douglas Stebila <douglas.stebila@sun.com>*
10615 * Change default behaviour of 'openssl asn1parse' so that more
10616 information is visible when viewing, e.g., a certificate:
10618 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10619 mode the content of non-printable OCTET STRINGs is output in a
10620 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10621 avoid the appearance of a printable string.
10623 *Nils Larsch <nla@trustcenter.de>*
10625 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10627 EC_GROUP_set_asn1_flag()
10628 EC_GROUP_get_asn1_flag()
10629 EC_GROUP_set_point_conversion_form()
10630 EC_GROUP_get_point_conversion_form()
10631 These control ASN1 encoding details:
10632 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10633 has been set to OPENSSL_EC_NAMED_CURVE.
10634 - Points are encoded in uncompressed form by default; options for
10635 asn1_for are as for point2oct, namely
10636 POINT_CONVERSION_COMPRESSED
10637 POINT_CONVERSION_UNCOMPRESSED
10638 POINT_CONVERSION_HYBRID
10640 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10642 EC_GROUP_set_seed()
10643 EC_GROUP_get0_seed()
10644 EC_GROUP_get_seed_len()
10645 This is used only for ASN1 purposes (so far).
10647 *Nils Larsch <nla@trustcenter.de>*
10649 * Add 'field_type' member to EC_METHOD, which holds the NID
10650 of the appropriate field type OID. The new function
10651 EC_METHOD_get_field_type() returns this value.
10653 *Nils Larsch <nla@trustcenter.de>*
10656 EC_POINT_point2bn()
10657 EC_POINT_bn2point()
10658 EC_POINT_point2hex()
10659 EC_POINT_hex2point()
10660 providing useful interfaces to EC_POINT_point2oct() and
10661 EC_POINT_oct2point().
10663 *Nils Larsch <nla@trustcenter.de>*
10665 * Change internals of the EC library so that the functions
10666 EC_GROUP_set_generator()
10667 EC_GROUP_get_generator()
10668 EC_GROUP_get_order()
10669 EC_GROUP_get_cofactor()
10670 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10671 to methods, which would lead to unnecessary code duplication when
10672 adding different types of curves.
10674 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10676 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10677 arithmetic, and such that modified wNAFs are generated
10678 (which avoid length expansion in many cases).
10682 * Add a function EC_GROUP_check_discriminant() (defined via
10683 EC_METHOD) that verifies that the curve discriminant is non-zero.
10685 Add a function EC_GROUP_check() that makes some sanity tests
10686 on a EC_GROUP, its generator and order. This includes
10687 EC_GROUP_check_discriminant().
10689 *Nils Larsch <nla@trustcenter.de>*
10691 * Add ECDSA in new directory crypto/ecdsa/.
10693 Add applications 'openssl ecparam' and 'openssl ecdsa'
10694 (these are based on 'openssl dsaparam' and 'openssl dsa').
10696 ECDSA support is also included in various other files across the
10697 library. Most notably,
10698 - 'openssl req' now has a '-newkey ecdsa:file' option;
10699 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10700 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10701 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10702 them suitable for ECDSA where domain parameters must be
10703 extracted before the specific public key;
10704 - ECDSA engine support has been added.
10706 *Nils Larsch <nla@trustcenter.de>*
10708 * Include some named elliptic curves, and add OIDs from X9.62,
10709 SECG, and WAP/WTLS. Each curve can be obtained from the new
10711 EC_GROUP_new_by_curve_name(),
10712 and the list of available named curves can be obtained with
10713 EC_get_builtin_curves().
10714 Also add a 'curve_name' member to EC_GROUP objects, which can be
10716 EC_GROUP_set_curve_name()
10717 EC_GROUP_get_curve_name()
10719 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10721 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10722 was actually never needed) and in BN_mul(). The removal in BN_mul()
10723 required a small change in bn_mul_part_recursive() and the addition
10724 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10725 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10726 bn_sub_words() and bn_add_words() except they take arrays with
10731 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10733 * Cleanse PEM buffers before freeing them since they may contain
10736 *Benjamin Bennett <ben@psc.edu>*
10738 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10739 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10740 authentication-only ciphersuites.
10744 * Since AES128 and AES256 share a single mask bit in the logic of
10745 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10746 kludge to work properly if AES128 is available and AES256 isn't.
10750 * Expand security boundary to match 1.1.1 module.
10754 * Remove redundant features: hash file source, editing of test vectors
10755 modify fipsld to use external fips_premain.c signature.
10759 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10760 run algorithm test programs.
10764 * Make algorithm test programs more tolerant of whitespace.
10768 * Have SSL/TLS server implementation tolerate "mismatched" record
10769 protocol version while receiving ClientHello even if the
10770 ClientHello is fragmented. (The server can't insist on the
10771 particular protocol version it has chosen before the ServerHello
10772 message has informed the client about his choice.)
10776 * Load error codes if they are not already present instead of using a
10777 static variable. This allows them to be cleanly unloaded and reloaded.
10781 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10783 * Introduce limits to prevent malicious keys being able to
10784 cause a denial of service. ([CVE-2006-2940])
10786 *Steve Henson, Bodo Moeller*
10788 * Fix ASN.1 parsing of certain invalid structures that can result
10789 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10791 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10792 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10794 * Fix SSL client code which could crash if connecting to a
10795 malicious SSLv2 server. ([CVE-2006-4343])
10797 *Tavis Ormandy and Will Drewry, Google Security Team*
10799 * Change ciphersuite string processing so that an explicit
10800 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10801 will no longer include "AES128-SHA"), and any other similar
10802 ciphersuite (same bitmap) from *other* protocol versions (so that
10803 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10804 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10805 changes from 0.9.8b and 0.9.8d.
10809 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10811 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10812 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10814 * Change the Unix randomness entropy gathering to use poll() when
10815 possible instead of select(), since the latter has some
10816 undesirable limitations.
10818 *Darryl Miles via Richard Levitte and Bodo Moeller*
10820 * Disable rogue ciphersuites:
10822 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10823 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10824 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10826 The latter two were purportedly from
10827 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10830 Also deactivate the remaining ciphersuites from
10831 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10832 unofficial, and the ID has long expired.
10836 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10837 dual-core machines) and other potential thread-safety issues.
10841 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10843 * Adapt fipsld and the build system to link against the validated FIPS
10844 module in FIPS mode.
10848 * Fixes for VC++ 2005 build under Windows.
10852 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10853 from a Windows bash shell such as MSYS. It is autodetected from the
10854 "config" script when run from a VC++ environment. Modify standard VC++
10855 build to use fipscanister.o from the GNU make build.
10859 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10861 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10862 The value now differs depending on if you build for FIPS or not.
10863 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10864 safely run with a non-FIPSed libcrypto, as it may crash because of
10865 the difference induced by this change.
10869 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10871 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10872 (part of SSL_OP_ALL). This option used to disable the
10873 countermeasure against man-in-the-middle protocol-version
10874 rollback in the SSL 2.0 server implementation, which is a bad
10875 idea. ([CVE-2005-2969])
10877 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10878 for Information Security, National Institute of Advanced Industrial
10879 Science and Technology [AIST, Japan)]*
10881 * Minimal support for X9.31 signatures and PSS padding modes. This is
10882 mainly for FIPS compliance and not fully integrated at this stage.
10886 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10887 the exponentiation using a fixed-length exponent. (Otherwise,
10888 the information leaked through timing could expose the secret key
10889 after many signatures; cf. Bleichenbacher's attack on DSA with
10894 * Make a new fixed-window mod_exp implementation the default for
10895 RSA, DSA, and DH private-key operations so that the sequence of
10896 squares and multiplies and the memory access pattern are
10897 independent of the particular secret key. This will mitigate
10898 cache-timing and potential related attacks.
10900 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10901 and this is automatically used by BN_mod_exp_mont() if the new flag
10902 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10903 will use this BN flag for private exponents unless the flag
10904 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10905 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10907 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10909 * Change the client implementation for SSLv23_method() and
10910 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10911 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10912 (Previously, the SSL 2.0 backwards compatible Client Hello
10913 message format would be used even with SSL_OP_NO_SSLv2.)
10917 * Add support for smime-type MIME parameter in S/MIME messages which some
10922 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10923 a threadsafe manner. Modify rsa code to use new function and add calls
10924 to dsa and dh code (which had race conditions before).
10928 * Include the fixed error library code in the C error file definitions
10929 instead of fixing them up at runtime. This keeps the error code
10930 structures constant.
10934 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10936 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10939 * Fixes for newer kerberos headers. NB: the casts are needed because
10940 the 'length' field is signed on one version and unsigned on another
10941 with no (?) obvious way to tell the difference, without these VC++
10942 complains. Also the "definition" of FAR (blank) is no longer included
10943 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10944 some needed definitions.
10948 * Undo Cygwin change.
10952 * Added support for proxy certificates according to RFC 3820.
10953 Because they may be a security thread to unaware applications,
10954 they must be explicitly allowed in run-time. See
10955 docs/HOWTO/proxy_certificates.txt for further information.
10959 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10961 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10962 server and client random values. Previously
10963 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10964 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10966 This change has negligible security impact because:
10968 1. Server and client random values still have 24 bytes of pseudo random
10971 2. Server and client random values are sent in the clear in the initial
10974 3. The master secret is derived using the premaster secret (48 bytes in
10975 size for static RSA ciphersuites) as well as client server and random
10978 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10981 *Stephen Henson, reported by UK NISCC*
10983 * Use Windows randomness collection on Cygwin.
10987 * Fix hang in EGD/PRNGD query when communication socket is closed
10988 prematurely by EGD/PRNGD.
10990 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10992 * Prompt for pass phrases when appropriate for PKCS12 input format.
10996 * Back-port of selected performance improvements from development
10997 branch, as well as improved support for PowerPC platforms.
11001 * Add lots of checks for memory allocation failure, error codes to indicate
11002 failure and freeing up memory if a failure occurs.
11004 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11006 * Add new -passin argument to dgst.
11010 * Perform some character comparisons of different types in X509_NAME_cmp:
11011 this is needed for some certificates that re-encode DNs into UTF8Strings
11012 (in violation of RFC3280) and can't or won't issue name rollover
11017 * Make an explicit check during certificate validation to see that
11018 the CA setting in each certificate on the chain is correct. As a
11019 side effect always do the following basic checks on extensions,
11020 not just when there's an associated purpose to the check:
11022 - if there is an unhandled critical extension (unless the user
11023 has chosen to ignore this fault)
11024 - if the path length has been exceeded (if one is set at all)
11025 - that certain extensions fit the associated purpose (if one has
11030 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11032 * Avoid a race condition when CRLs are checked in a multi threaded
11033 environment. This would happen due to the reordering of the revoked
11034 entries during signature checking and serial number lookup. Now the
11035 encoding is cached and the serial number sort performed under a lock.
11036 Add new STACK function sk_is_sorted().
11040 * Add Delta CRL to the extension code.
11044 * Various fixes to s3_pkt.c so alerts are sent properly.
11046 *David Holmes <d.holmes@f5.com>*
11048 * Reduce the chances of duplicate issuer name and serial numbers (in
11049 violation of RFC3280) using the OpenSSL certificate creation utilities.
11050 This is done by creating a random 64 bit value for the initial serial
11051 number when a serial number file is created or when a self signed
11052 certificate is created using 'openssl req -x509'. The initial serial
11053 number file is created using 'openssl x509 -next_serial' in CA.pl
11054 rather than being initialized to 1.
11058 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11060 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11061 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11063 *Joe Orton, Steve Henson*
11065 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11068 *Joe Orton, Steve Henson*
11070 * Make it possible to have multiple active certificates with the same
11071 subject in the CA index file. This is done only if the keyword
11072 'unique_subject' is set to 'no' in the main CA section (default
11073 if 'CA_default') of the configuration file. The value is saved
11074 with the database itself in a separate index attribute file,
11075 named like the index file with '.attr' appended to the name.
11079 * X509 verify fixes. Disable broken certificate workarounds when
11080 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11081 keyUsage extension present. Don't accept CRLs with unhandled critical
11082 extensions: since verify currently doesn't process CRL extensions this
11083 rejects a CRL with *any* critical extensions. Add new verify error codes
11088 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11089 A clarification of RFC2560 will require the use of OCTET STRINGs and
11090 some implementations cannot handle the current raw format. Since OpenSSL
11091 copies and compares OCSP nonces as opaque blobs without any attempt at
11092 parsing them this should not create any compatibility issues.
11096 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11097 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11098 this HMAC (and other) operations are several times slower than OpenSSL
11103 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11105 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11107 * Use the correct content when signing type "other".
11111 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11113 * Fix various bugs revealed by running the NISCC test suite:
11115 Stop out of bounds reads in the ASN1 code when presented with
11116 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11118 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11120 If verify callback ignores invalid public key errors don't try to check
11121 certificate signature with the NULL public key.
11125 * New -ignore_err option in ocsp application to stop the server
11126 exiting on the first error in a request.
11130 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11131 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11136 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11137 extra data after the compression methods not only for TLS 1.0
11138 but also for SSL 3.0 (as required by the specification).
11140 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11142 * Change X509_certificate_type() to mark the key as exported/exportable
11143 when it's 512 *bits* long, not 512 bytes.
11147 * Change AES_cbc_encrypt() so it outputs exact multiple of
11148 blocks during encryption.
11152 * Various fixes to base64 BIO and non blocking I/O. On write
11153 flushes were not handled properly if the BIO retried. On read
11154 data was not being buffered properly and had various logic bugs.
11155 This also affects blocking I/O when the data being decoded is a
11160 * Various S/MIME bugfixes and compatibility changes:
11161 output correct application/pkcs7 MIME type if
11162 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11163 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11164 of files as .eml work). Correctly handle very long lines in MIME
11169 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11171 * Countermeasure against the Klima-Pokorny-Rosa extension of
11172 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11173 a protocol version number mismatch like a decryption error
11174 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11178 * Turn on RSA blinding by default in the default implementation
11179 to avoid a timing attack. Applications that don't want it can call
11180 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11181 They would be ill-advised to do so in most cases.
11183 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11185 * Change RSA blinding code so that it works when the PRNG is not
11186 seeded (in this case, the secret RSA exponent is abused as
11187 an unpredictable seed -- if it is not unpredictable, there
11188 is no point in blinding anyway). Make RSA blinding thread-safe
11189 by remembering the creator's thread ID in rsa->blinding and
11190 having all other threads use local one-time blinding factors
11191 (this requires more computation than sharing rsa->blinding, but
11192 avoids excessive locking; and if an RSA object is not shared
11193 between threads, blinding will still be very fast).
11197 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11198 ENGINE as defaults for all supported algorithms irrespective of
11199 the 'flags' parameter. 'flags' is now honoured, so applications
11200 should make sure they are passing it correctly.
11204 * Target "mingw" now allows native Windows code to be generated in
11205 the Cygwin environment as well as with the MinGW compiler.
11209 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11211 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11212 via timing by performing a MAC computation even if incorrect
11213 block cipher padding has been found. This is a countermeasure
11214 against active attacks where the attacker has to distinguish
11215 between bad padding and a MAC verification error. ([CVE-2003-0078])
11217 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11218 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11219 Martin Vuagnoux (EPFL, Ilion)*
11221 * Make the no-err option work as intended. The intention with no-err
11222 is not to have the whole error stack handling routines removed from
11223 libcrypto, it's only intended to remove all the function name and
11224 reason texts, thereby removing some of the footprint that may not
11225 be interesting if those errors aren't displayed anyway.
11227 NOTE: it's still possible for any application or module to have its
11228 own set of error texts inserted. The routines are there, just not
11229 used by default when no-err is given.
11233 * Add support for FreeBSD on IA64.
11235 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11237 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11238 Kerberos function mit_des_cbc_cksum(). Before this change,
11239 the value returned by DES_cbc_cksum() was like the one from
11240 mit_des_cbc_cksum(), except the bytes were swapped.
11242 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11244 * Allow an application to disable the automatic SSL chain building.
11245 Before this a rather primitive chain build was always performed in
11246 ssl3_output_cert_chain(): an application had no way to send the
11247 correct chain if the automatic operation produced an incorrect result.
11249 Now the chain builder is disabled if either:
11251 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11253 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11255 The reasoning behind this is that an application would not want the
11256 auto chain building to take place if extra chain certificates are
11257 present and it might also want a means of sending no additional
11258 certificates (for example the chain has two certificates and the
11263 * Add the possibility to build without the ENGINE framework.
11265 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11267 * Under Win32 gmtime() can return NULL: check return value in
11268 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11272 * DSA routines: under certain error conditions uninitialized BN objects
11273 could be freed. Solution: make sure initialization is performed early
11274 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11275 Nils Larsch <nla@trustcenter.de> via PR#459)
11279 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11280 checked on reconnect on the client side, therefore session resumption
11281 could still fail with a "ssl session id is different" error. This
11282 behaviour is masked when SSL_OP_ALL is used due to
11283 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11284 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11285 followup to PR #377.
11289 * IA-32 assembler support enhancements: unified ELF targets, support
11290 for SCO/Caldera platforms, fix for Cygwin shared build.
11294 * Add support for FreeBSD on sparc64. As a consequence, support for
11295 FreeBSD on non-x86 processors is separate from x86 processors on
11296 the config script, much like the NetBSD support.
11298 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11300 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11302 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11305 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11306 code (06) was taken as the first octet of the session ID and the last
11307 octet was ignored consequently. As a result SSLv2 client side session
11308 caching could not have worked due to the session ID mismatch between
11310 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11315 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11316 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11321 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11322 seems that in spite of existing for more than a year, many application
11323 author have done nothing to provide the necessary callbacks, which
11324 means that this particular engine will not work properly anywhere.
11325 This is a very unfortunate situation which forces us, in the name
11326 of usability, to give the hw_ncipher.c a static lock, which is part
11328 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11329 appear in 0.9.8 or later. We EXPECT application authors to have
11330 dealt properly with this when 0.9.8 is released (unless we actually
11331 make such changes in the libcrypto locking code that changes will
11332 have to be made anyway).
11336 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11337 octets have been read, EOF or an error occurs. Without this change
11338 some truncated ASN1 structures will not produce an error.
11342 * Disable Heimdal support, since it hasn't been fully implemented.
11343 Still give the possibility to force the use of Heimdal, but with
11344 warnings and a request that patches get sent to openssl-dev.
11348 * Add the VC-CE target, introduce the WINCE sysname, and add
11349 INSTALL.WCE and appropriate conditionals to make it build.
11351 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11353 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11354 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11355 edit numbers of the version.
11357 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11359 * Introduce safe string copy and catenation functions
11360 (BUF_strlcpy() and BUF_strlcat()).
11362 *Ben Laurie (CHATS) and Richard Levitte*
11364 * Avoid using fixed-size buffers for one-line DNs.
11366 *Ben Laurie (CHATS)*
11368 * Add BUF_MEM_grow_clean() to avoid information leakage when
11369 resizing buffers containing secrets, and use where appropriate.
11371 *Ben Laurie (CHATS)*
11373 * Avoid using fixed size buffers for configuration file location.
11375 *Ben Laurie (CHATS)*
11377 * Avoid filename truncation for various CA files.
11379 *Ben Laurie (CHATS)*
11381 * Use sizeof in preference to magic numbers.
11383 *Ben Laurie (CHATS)*
11385 * Avoid filename truncation in cert requests.
11387 *Ben Laurie (CHATS)*
11389 * Add assertions to check for (supposedly impossible) buffer
11392 *Ben Laurie (CHATS)*
11394 * Don't cache truncated DNS entries in the local cache (this could
11395 potentially lead to a spoofing attack).
11397 *Ben Laurie (CHATS)*
11399 * Fix various buffers to be large enough for hex/decimal
11400 representations in a platform independent manner.
11402 *Ben Laurie (CHATS)*
11404 * Add CRYPTO_realloc_clean() to avoid information leakage when
11405 resizing buffers containing secrets, and use where appropriate.
11407 *Ben Laurie (CHATS)*
11409 * Add BIO_indent() to avoid much slightly worrying code to do
11412 *Ben Laurie (CHATS)*
11414 * Convert sprintf()/BIO_puts() to BIO_printf().
11416 *Ben Laurie (CHATS)*
11418 * buffer_gets() could terminate with the buffer only half
11421 *Ben Laurie (CHATS)*
11423 * Add assertions to prevent user-supplied crypto functions from
11424 overflowing internal buffers by having large block sizes, etc.
11426 *Ben Laurie (CHATS)*
11428 * New OPENSSL_assert() macro (similar to assert(), but enabled
11431 *Ben Laurie (CHATS)*
11433 * Eliminate unused copy of key in RC4.
11435 *Ben Laurie (CHATS)*
11437 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11439 *Ben Laurie (CHATS)*
11441 * Fix off-by-one error in EGD path.
11443 *Ben Laurie (CHATS)*
11445 * If RANDFILE path is too long, ignore instead of truncating.
11447 *Ben Laurie (CHATS)*
11449 * Eliminate unused and incorrectly sized X.509 structure
11452 *Ben Laurie (CHATS)*
11454 * Eliminate unused and dangerous function knumber().
11456 *Ben Laurie (CHATS)*
11458 * Eliminate unused and dangerous structure, KSSL_ERR.
11460 *Ben Laurie (CHATS)*
11462 * Protect against overlong session ID context length in an encoded
11463 session object. Since these are local, this does not appear to be
11466 *Ben Laurie (CHATS)*
11468 * Change from security patch (see 0.9.6e below) that did not affect
11469 the 0.9.6 release series:
11471 Remote buffer overflow in SSL3 protocol - an attacker could
11472 supply an oversized master key in Kerberos-enabled versions.
11475 *Ben Laurie (CHATS)*
11477 * Change the SSL kerb5 codes to match RFC 2712.
11481 * Make -nameopt work fully for req and add -reqopt switch.
11483 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11485 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11487 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11489 * Make sure tests can be performed even if the corresponding algorithms
11490 have been removed entirely. This was also the last step to make
11491 OpenSSL compilable with DJGPP under all reasonable conditions.
11493 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11495 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11496 to allow version independent disabling of normally unselected ciphers,
11497 which may be activated as a side-effect of selecting a single cipher.
11499 (E.g., cipher list string "RSA" enables ciphersuites that are left
11500 out of "ALL" because they do not provide symmetric encryption.
11501 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11503 *Lutz Jaenicke, Bodo Moeller*
11505 * Add appropriate support for separate platform-dependent build
11506 directories. The recommended way to make a platform-dependent
11507 build directory is the following (tested on Linux), maybe with
11510 # Place yourself outside of the OpenSSL source tree. In
11511 # this example, the environment variable OPENSSL_SOURCE
11512 # is assumed to contain the absolute OpenSSL source directory.
11513 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11514 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11515 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11516 mkdir -p `dirname $F`
11517 ln -s $OPENSSL_SOURCE/$F $F
11520 To be absolutely sure not to disturb the source tree, a "make clean"
11521 is a good thing. If it isn't successful, don't worry about it,
11522 it probably means the source directory is very clean.
11526 * Make sure any ENGINE control commands make local copies of string
11527 pointers passed to them whenever necessary. Otherwise it is possible
11528 the caller may have overwritten (or deallocated) the original string
11529 data when a later ENGINE operation tries to use the stored values.
11531 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11533 * Improve diagnostics in file reading and command-line digests.
11535 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11537 * Add AES modes CFB and OFB to the object database. Correct an
11538 error in AES-CFB decryption.
11542 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11543 allows existing EVP_CIPHER_CTX structures to be reused after
11544 calling `EVP_*Final()`. This behaviour is used by encryption
11545 BIOs and some applications. This has the side effect that
11546 applications must explicitly clean up cipher contexts with
11547 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11551 * Check the values of dna and dnb in bn_mul_recursive before calling
11552 bn_mul_comba (a non zero value means the a or b arrays do not contain
11553 n2 elements) and fallback to bn_mul_normal if either is not zero.
11557 * Fix escaping of non-ASCII characters when using the -subj option
11558 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11562 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11563 form for "surname", serialNumber has no short form.
11564 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11565 therefore remove "mail" short name for "internet 7".
11566 The OID for unique identifiers in X509 certificates is
11567 x500UniqueIdentifier, not uniqueIdentifier.
11568 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11572 * Add an "init" command to the ENGINE config module and auto initialize
11573 ENGINEs. Without any "init" command the ENGINE will be initialized
11574 after all ctrl commands have been executed on it. If init=1 the
11575 ENGINE is initialized at that point (ctrls before that point are run
11576 on the uninitialized ENGINE and after on the initialized one). If
11577 init=0 then the ENGINE will not be initialized at all.
11581 * Fix the 'app_verify_callback' interface so that the user-defined
11582 argument is actually passed to the callback: In the
11583 SSL_CTX_set_cert_verify_callback() prototype, the callback
11584 declaration has been changed from
11587 int (*cb)(X509_STORE_CTX *,void *);
11588 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11589 i=s->ctx->app_verify_callback(&ctx)
11590 has been changed into
11591 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11593 To update applications using SSL_CTX_set_cert_verify_callback(),
11594 a dummy argument can be added to their callback functions.
11596 *D. K. Smetters <smetters@parc.xerox.com>*
11598 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11600 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11602 * Add and OPENSSL_LOAD_CONF define which will cause
11603 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11604 This allows older applications to transparently support certain
11605 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11606 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11607 load the config file and OPENSSL_add_all_algorithms_conf() which will
11608 always load it have also been added.
11612 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11613 Adjust NIDs and EVP layer.
11615 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11617 * Config modules support in openssl utility.
11619 Most commands now load modules from the config file,
11620 though in a few (such as version) this isn't done
11621 because it couldn't be used for anything.
11623 In the case of ca and req the config file used is
11624 the same as the utility itself: that is the -config
11625 command line option can be used to specify an
11630 * Move default behaviour from OPENSSL_config(). If appname is NULL
11631 use "openssl_conf" if filename is NULL use default openssl config file.
11635 * Add an argument to OPENSSL_config() to allow the use of an alternative
11636 config section name. Add a new flag to tolerate a missing config file
11637 and move code to CONF_modules_load_file().
11641 * Support for crypto accelerator cards from Accelerated Encryption
11642 Processing, www.aep.ie. (Use engine 'aep')
11643 The support was copied from 0.9.6c [engine] and adapted/corrected
11644 to work with the new engine framework.
11646 *AEP Inc. and Richard Levitte*
11648 * Support for SureWare crypto accelerator cards from Baltimore
11649 Technologies. (Use engine 'sureware')
11650 The support was copied from 0.9.6c [engine] and adapted
11651 to work with the new engine framework.
11655 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11656 make the newer ENGINE framework commands for the CHIL engine work.
11658 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11660 * Make it possible to produce shared libraries on ReliantUNIX.
11662 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11664 * Add the configuration target debug-linux-ppro.
11665 Make 'openssl rsa' use the general key loading routines
11666 implemented in apps.c, and make those routines able to
11667 handle the key format FORMAT_NETSCAPE and the variant
11670 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11672 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11674 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11676 * Add -keyform to rsautl, and document -engine.
11678 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11680 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11681 BIO_R_NO_SUCH_FILE error code rather than the generic
11682 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11686 * Add new functions
11687 ERR_peek_last_error
11688 ERR_peek_last_error_line
11689 ERR_peek_last_error_line_data.
11690 These are similar to
11692 ERR_peek_error_line
11693 ERR_peek_error_line_data,
11694 but report on the latest error recorded rather than the first one
11695 still in the error queue.
11697 *Ben Laurie, Bodo Moeller*
11699 * default_algorithms option in ENGINE config module. This allows things
11701 default_algorithms = ALL
11702 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11706 * Preliminary ENGINE config module.
11710 * New experimental application configuration code.
11714 * Change the AES code to follow the same name structure as all other
11715 symmetric ciphers, and behave the same way. Move everything to
11716 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11718 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11720 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11722 *Ben Laurie and Theo de Raadt*
11724 * Add option to output public keys in req command.
11726 *Massimiliano Pala madwolf@openca.org*
11728 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11729 (up to about 10% better than before for P-192 and P-224).
11733 * New functions/macros
11735 SSL_CTX_set_msg_callback(ctx, cb)
11736 SSL_CTX_set_msg_callback_arg(ctx, arg)
11737 SSL_set_msg_callback(ssl, cb)
11738 SSL_set_msg_callback_arg(ssl, arg)
11740 to request calling a callback function
11742 void cb(int write_p, int version, int content_type,
11743 const void *buf, size_t len, SSL *ssl, void *arg)
11745 whenever a protocol message has been completely received
11746 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11747 protocol version according to which the SSL library interprets
11748 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11749 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11750 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11751 specification (change_cipher_spec(20), alert(21), handshake(22)).
11752 'buf' and 'len' point to the actual message, 'ssl' to the
11753 SSL object, and 'arg' is the application-defined value set by
11754 SSL[_CTX]_set_msg_callback_arg().
11756 'openssl s_client' and 'openssl s_server' have new '-msg' options
11757 to enable a callback that displays all protocol messages.
11761 * Change the shared library support so shared libraries are built as
11762 soon as the corresponding static library is finished, and thereby get
11763 openssl and the test programs linked against the shared library.
11764 This still only happens when the keyword "shard" has been given to
11765 the configuration scripts.
11767 NOTE: shared library support is still an experimental thing, and
11768 backward binary compatibility is still not guaranteed.
11770 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11772 * Add support for Subject Information Access extension.
11774 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11776 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11777 additional bytes when new memory had to be allocated, not just
11778 when reusing an existing buffer.
11782 * New command line and configuration option 'utf8' for the req command.
11783 This allows field values to be specified as UTF8 strings.
11787 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11788 runs for the former and machine-readable output for the latter.
11792 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11793 of the e-mail address in the DN (i.e., it will go into a certificate
11794 extension only). The new configuration file option 'email_in_dn = no'
11795 has the same effect.
11797 *Massimiliano Pala madwolf@openca.org*
11799 * Change all functions with names starting with `des_` to be starting
11800 with `DES_` instead. Add wrappers that are compatible with libdes,
11801 but are named `_ossl_old_des_*`. Finally, add macros that map the
11802 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11803 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11804 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11807 Since we provide two compatibility mappings, the user needs to
11808 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11809 compatibility is desired. The default (i.e., when that macro
11810 isn't defined) is OpenSSL 0.9.6c compatibility.
11812 There are also macros that enable and disable the support of old
11813 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11814 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11815 are defined, the default will apply: to support the old des routines.
11817 In either case, one must include openssl/des.h to get the correct
11818 definitions. Do not try to just include openssl/des_old.h, that
11821 NOTE: This is a major break of an old API into a new one. Software
11822 authors are encouraged to switch to the `DES_` style functions. Some
11823 time in the future, des_old.h and the libdes compatibility functions
11824 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11825 default), and then completely removed.
11829 * Test for certificates which contain unsupported critical extensions.
11830 If such a certificate is found during a verify operation it is
11831 rejected by default: this behaviour can be overridden by either
11832 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11833 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11834 X509_supported_extension() has also been added which returns 1 if a
11835 particular extension is supported.
11839 * Modify the behaviour of EVP cipher functions in similar way to digests
11840 to retain compatibility with existing code.
11844 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11845 compatibility with existing code. In particular the 'ctx' parameter does
11846 not have to be to be initialized before the call to EVP_DigestInit() and
11847 it is tidied up after a call to EVP_DigestFinal(). New function
11848 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11849 EVP_MD_CTX_copy() changed to not require the destination to be
11850 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11851 requires the destination to be valid.
11853 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11854 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11858 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11859 so that complete 'Handshake' protocol structures are kept in memory
11860 instead of overwriting 'msg_type' and 'length' with 'body' data.
11864 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11866 *Massimo Santin via Richard Levitte*
11868 * Major restructuring to the underlying ENGINE code. This includes
11869 reduction of linker bloat, separation of pure "ENGINE" manipulation
11870 (initialisation, etc) from functionality dealing with implementations
11871 of specific crypto interfaces. This change also introduces integrated
11872 support for symmetric ciphers and digest implementations - so ENGINEs
11873 can now accelerate these by providing EVP_CIPHER and EVP_MD
11874 implementations of their own. This is detailed in
11875 [crypto/engine/README.md](crypto/engine/README.md)
11876 as it couldn't be adequately described here. However, there are a few
11877 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11878 were changed in the original introduction of ENGINE code have now
11879 reverted back - the hooking from this code to ENGINE is now a good
11880 deal more passive and at run-time, operations deal directly with
11881 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11882 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11883 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11884 they were not being used by the framework as there is no concept of a
11885 BIGNUM_METHOD and they could not be generalised to the new
11886 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11887 ENGINE_cpy() has been removed as it cannot be consistently defined in
11892 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11896 * Change mkdef.pl to sort symbols that get the same entry number,
11897 and make sure the automatically generated functions `ERR_load_*`
11898 become part of libeay.num as well.
11902 * New function SSL_renegotiate_pending(). This returns true once
11903 renegotiation has been requested (either SSL_renegotiate() call
11904 or HelloRequest/ClientHello received from the peer) and becomes
11905 false once a handshake has been completed.
11906 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11907 sends a HelloRequest, but does not ensure that a handshake takes
11908 place. SSL_renegotiate_pending() is useful for checking if the
11909 client has followed the request.)
11913 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11914 By default, clients may request session resumption even during
11915 renegotiation (if session ID contexts permit); with this option,
11916 session resumption is possible only in the first handshake.
11918 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11919 more bits available for options that should not be part of
11920 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11924 * Add some demos for certificate and certificate request creation.
11928 * Make maximum certificate chain size accepted from the peer application
11929 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11930 "Douglas E. Engert" <deengert@anl.gov>.
11934 * Add support for shared libraries for Unixware-7
11935 (Boyd Lynn Gerber <gerberb@zenez.com>).
11939 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11940 be done prior to destruction. Use this to unload error strings from
11941 ENGINEs that load their own error strings. NB: This adds two new API
11942 functions to "get" and "set" this destroy handler in an ENGINE.
11946 * Alter all existing ENGINE implementations (except "openssl" and
11947 "openbsd") to dynamically instantiate their own error strings. This
11948 makes them more flexible to be built both as statically-linked ENGINEs
11949 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11950 Also, add stub code to each that makes building them as self-contained
11951 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11955 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11956 implementations into applications that are completely implemented in
11957 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11958 commands that can be used to configure what shared-library to load and
11959 to control aspects of the way it is handled. Also, made an update to
11960 the [README-Engine.md](README-Engine.md) file
11961 that brings its information up-to-date and
11962 provides some information and instructions on the "dynamic" ENGINE
11963 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11967 * Make it possible to unload ranges of ERR strings with a new
11968 "ERR_unload_strings" function.
11972 * Add a copy() function to EVP_MD.
11976 * Make EVP_MD routines take a context pointer instead of just the
11977 md_data void pointer.
11981 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11982 that the digest can only process a single chunk of data
11983 (typically because it is provided by a piece of
11984 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11985 is only going to provide a single chunk of data, and hence the
11986 framework needn't accumulate the data for oneshot drivers.
11990 * As with "ERR", make it possible to replace the underlying "ex_data"
11991 functions. This change also alters the storage and management of global
11992 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11993 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11994 index counters. The API functions that use this state have been changed
11995 to take a "class_index" rather than pointers to the class's local STACK
11996 and counter, and there is now an API function to dynamically create new
11997 classes. This centralisation allows us to (a) plug a lot of the
11998 thread-safety problems that existed, and (b) makes it possible to clean
11999 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12000 such data would previously have always leaked in application code and
12001 workarounds were in place to make the memory debugging turn a blind eye
12002 to it. Application code that doesn't use this new function will still
12003 leak as before, but their memory debugging output will announce it now
12004 rather than letting it slide.
12006 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12007 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12008 has a return value to indicate success or failure.
12012 * Make it possible to replace the underlying "ERR" functions such that the
12013 global state (2 LHASH tables and 2 locks) is only used by the "default"
12014 implementation. This change also adds two functions to "get" and "set"
12015 the implementation prior to it being automatically set the first time
12016 any other ERR function takes place. Ie. an application can call "get",
12017 pass the return value to a module it has just loaded, and that module
12018 can call its own "set" function using that value. This means the
12019 module's "ERR" operations will use (and modify) the error state in the
12020 application and not in its own statically linked copy of OpenSSL code.
12024 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12025 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12026 the operation, and provides a more encapsulated way for external code
12027 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12028 to use these functions rather than manually incrementing the counts.
12030 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12034 * Add EVP test program.
12038 * Add symmetric cipher support to ENGINE. Expect the API to change!
12042 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12043 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12044 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12045 These allow a CRL to be built without having to access X509_CRL fields
12046 directly. Modify 'ca' application to use new functions.
12050 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12051 bug workarounds. Rollback attack detection is a security feature.
12052 The problem will only arise on OpenSSL servers when TLSv1 is not
12053 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12054 Software authors not wanting to support TLSv1 will have special reasons
12055 for their choice and can explicitly enable this option.
12057 *Bodo Moeller, Lutz Jaenicke*
12059 * Rationalise EVP so it can be extended: don't include a union of
12060 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12061 (similar to those existing for EVP_CIPHER_CTX).
12066 EVP_MD_CTX_init(&md); /* new function call */
12067 EVP_DigestInit(&md, EVP_sha1());
12068 EVP_DigestUpdate(&md, in, len);
12069 EVP_DigestFinal(&md, out, NULL);
12070 EVP_MD_CTX_cleanup(&md); /* new function call */
12074 * Make DES key schedule conform to the usual scheme, as well as
12075 correcting its structure. This means that calls to DES functions
12076 now have to pass a pointer to a des_key_schedule instead of a
12077 plain des_key_schedule (which was actually always a pointer
12080 des_key_schedule ks;
12082 des_set_key_checked(..., &ks);
12083 des_ncbc_encrypt(..., &ks, ...);
12085 (Note that a later change renames 'des_...' into 'DES_...'.)
12089 * Initial reduction of linker bloat: the use of some functions, such as
12090 PEM causes large amounts of unused functions to be linked in due to
12091 poor organisation. For example pem_all.c contains every PEM function
12092 which has a knock on effect of linking in large amounts of (unused)
12093 ASN1 code. Grouping together similar functions and splitting unrelated
12094 functions prevents this.
12098 * Cleanup of EVP macros.
12102 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12103 correct `_ecb suffix`.
12107 * Add initial OCSP responder support to ocsp application. The
12108 revocation information is handled using the text based index
12109 use by the ca application. The responder can either handle
12110 requests generated internally, supplied in files (for example
12111 via a CGI script) or using an internal minimal server.
12115 * Add configuration choices to get zlib compression for TLS.
12119 * Changes to Kerberos SSL for RFC 2712 compliance:
12120 1. Implemented real KerberosWrapper, instead of just using
12121 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12122 2. Implemented optional authenticator field of KerberosWrapper.
12124 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12125 and authenticator structs; see crypto/krb5/.
12127 Generalized Kerberos calls to support multiple Kerberos libraries.
12128 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12129 via Richard Levitte*
12131 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12132 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12133 values for each of the key sizes rather than having just
12134 parameters (and 'speed' generating keys each time).
12138 * Speed up EVP routines.
12141 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12142 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12143 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12144 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12146 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12147 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12148 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12151 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12153 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12157 * Added the OS2-EMX target.
12159 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12161 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12162 to support NCONF routines in extension code. New function CONF_set_nconf()
12163 to allow functions which take an NCONF to also handle the old LHASH
12164 structure: this means that the old CONF compatible routines can be
12165 retained (in particular wrt extensions) without having to duplicate the
12166 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12170 * Enhance the general user interface with mechanisms for inner control
12171 and with possibilities to have yes/no kind of prompts.
12175 * Change all calls to low level digest routines in the library and
12176 applications to use EVP. Add missing calls to HMAC_cleanup() and
12177 don't assume HMAC_CTX can be copied using memcpy().
12179 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12181 * Add the possibility to control engines through control names but with
12182 arbitrary arguments instead of just a string.
12183 Change the key loaders to take a UI_METHOD instead of a callback
12184 function pointer. NOTE: this breaks binary compatibility with earlier
12185 versions of OpenSSL [engine].
12186 Adapt the nCipher code for these new conditions and add a card insertion
12191 * Enhance the general user interface with mechanisms to better support
12192 dialog box interfaces, application-defined prompts, the possibility
12193 to use defaults (for example default passwords from somewhere else)
12194 and interrupts/cancellations.
12198 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12199 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12203 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12204 tidy up some unnecessarily weird code in 'sk_new()').
12206 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12208 * Change the key loading routines for ENGINEs to use the same kind
12209 callback (pem_password_cb) as all other routines that need this
12214 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12215 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12216 than this minimum value is recommended.
12220 * New random seeder for OpenVMS, using the system process statistics
12221 that are easily reachable.
12225 * Windows apparently can't transparently handle global
12226 variables defined in DLLs. Initialisations such as:
12228 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12230 won't compile. This is used by the any applications that need to
12231 declare their own ASN1 modules. This was fixed by adding the option
12232 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12233 needed for static libraries under Win32.
12237 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12238 setting of purpose and trust fields. New X509_STORE trust and
12239 purpose functions and tidy up setting in other SSL functions.
12243 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12244 structure. These are inherited by X509_STORE_CTX when it is
12245 initialised. This allows various defaults to be set in the
12246 X509_STORE structure (such as flags for CRL checking and custom
12247 purpose or trust settings) for functions which only use X509_STORE_CTX
12248 internally such as S/MIME.
12250 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12251 trust settings if they are not set in X509_STORE. This allows X509_STORE
12252 purposes and trust (in S/MIME for example) to override any set by default.
12254 Add command line options for CRL checking to smime, s_client and s_server
12259 * Initial CRL based revocation checking. If the CRL checking flag(s)
12260 are set then the CRL is looked up in the X509_STORE structure and
12261 its validity and signature checked, then if the certificate is found
12262 in the CRL the verify fails with a revoked error.
12264 Various new CRL related callbacks added to X509_STORE_CTX structure.
12266 Command line options added to 'verify' application to support this.
12268 This needs some additional work, such as being able to handle multiple
12269 CRLs with different times, extension based lookup (rather than just
12270 by subject name) and ultimately more complete V2 CRL extension
12275 * Add a general user interface API (crypto/ui/). This is designed
12276 to replace things like des_read_password and friends (backward
12277 compatibility functions using this new API are provided).
12278 The purpose is to remove prompting functions from the DES code
12279 section as well as provide for prompting through dialog boxes in
12280 a window system and the like.
12284 * Add "ex_data" support to ENGINE so implementations can add state at a
12285 per-structure level rather than having to store it globally.
12289 * Make it possible for ENGINE structures to be copied when retrieved by
12290 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12291 This causes the "original" ENGINE structure to act like a template,
12292 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12293 operational state can be localised to each ENGINE structure, despite the
12294 fact they all share the same "methods". New ENGINE structures returned in
12295 this case have no functional references and the return value is the single
12296 structural reference. This matches the single structural reference returned
12297 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12302 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12303 needs to match any other type at all we need to manually clear the
12308 * Changes to the "openssl engine" utility to include;
12309 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12310 about an ENGINE's available control commands.
12311 - executing control commands from command line arguments using the
12312 '-pre' and '-post' switches. '-post' is only used if '-t' is
12313 specified and the ENGINE is successfully initialised. The syntax for
12314 the individual commands are colon-separated, for example;
12315 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12319 * New dynamic control command support for ENGINEs. ENGINEs can now
12320 declare their own commands (numbers), names (strings), descriptions,
12321 and input types for run-time discovery by calling applications. A
12322 subset of these commands are implicitly classed as "executable"
12323 depending on their input type, and only these can be invoked through
12324 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12325 can be based on user input, config files, etc). The distinction is
12326 that "executable" commands cannot return anything other than a boolean
12327 result and can only support numeric or string input, whereas some
12328 discoverable commands may only be for direct use through
12329 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12330 pointers, or other custom uses. The "executable" commands are to
12331 support parameterisations of ENGINE behaviour that can be
12332 unambiguously defined by ENGINEs and used consistently across any
12333 OpenSSL-based application. Commands have been added to all the
12334 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12335 control over shared-library paths without source code alterations.
12339 * Changed all ENGINE implementations to dynamically allocate their
12340 ENGINEs rather than declaring them statically. Apart from this being
12341 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12342 this also allows the implementations to compile without using the
12343 internal engine_int.h header.
12347 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12348 'const' value. Any code that should be able to modify a RAND_METHOD
12349 should already have non-const pointers to it (ie. they should only
12350 modify their own ones).
12354 * Made a variety of little tweaks to the ENGINE code.
12355 - "atalla" and "ubsec" string definitions were moved from header files
12356 to C code. "nuron" string definitions were placed in variables
12357 rather than hard-coded - allowing parameterisation of these values
12358 later on via ctrl() commands.
12359 - Removed unused "#if 0"'d code.
12360 - Fixed engine list iteration code so it uses ENGINE_free() to release
12361 structural references.
12362 - Constified the RAND_METHOD element of ENGINE structures.
12363 - Constified various get/set functions as appropriate and added
12364 missing functions (including a catch-all ENGINE_cpy that duplicates
12365 all ENGINE values onto a new ENGINE except reference counts/state).
12366 - Removed NULL parameter checks in get/set functions. Setting a method
12367 or function to NULL is a way of cancelling out a previously set
12368 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12369 and doesn't justify the extra error symbols and code.
12370 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12371 flags from engine_int.h to engine.h.
12372 - Changed prototypes for ENGINE handler functions (init(), finish(),
12373 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12377 * Implement binary inversion algorithm for BN_mod_inverse in addition
12378 to the algorithm using long division. The binary algorithm can be
12379 used only if the modulus is odd. On 32-bit systems, it is faster
12380 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12381 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12382 up to 450 bits. In 64-bit environments, the binary algorithm
12383 appears to be advantageous for much longer moduli; here we use it
12384 for moduli up to 2048 bits.
12388 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12389 could not support the combine flag in choice fields.
12393 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12394 extensions from a certificate request to the certificate.
12398 * Allow multiple 'certopt' and 'nameopt' options to be separated
12399 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12400 file: this allows the display of the certificate about to be
12401 signed to be customised, to allow certain fields to be included
12402 or excluded and extension details. The old system didn't display
12403 multicharacter strings properly, omitted fields not in the policy
12404 and couldn't display additional details such as extensions.
12408 * Function EC_POINTs_mul for multiple scalar multiplication
12409 of an arbitrary number of elliptic curve points
12410 \sum scalars[i]*points[i],
12411 optionally including the generator defined for the EC_GROUP:
12412 scalar*generator + \sum scalars[i]*points[i].
12414 EC_POINT_mul is a simple wrapper function for the typical case
12415 that the point list has just one item (besides the optional
12420 * First EC_METHODs for curves over GF(p):
12422 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12423 operations and provides various method functions that can also
12424 operate with faster implementations of modular arithmetic.
12426 EC_GFp_mont_method() reuses most functions that are part of
12427 EC_GFp_simple_method, but uses Montgomery arithmetic.
12429 *Bodo Moeller; point addition and point doubling
12430 implementation directly derived from source code provided by
12431 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12433 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12434 crypto/ec/ec_lib.c):
12436 Curves are EC_GROUP objects (with an optional group generator)
12437 based on EC_METHODs that are built into the library.
12439 Points are EC_POINT objects based on EC_GROUP objects.
12441 Most of the framework would be able to handle curves over arbitrary
12442 finite fields, but as there are no obvious types for fields other
12443 than GF(p), some functions are limited to that for now.
12447 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12448 that the file contains a complete HTTP response.
12452 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12453 change the def and num file printf format specifier from "%-40sXXX"
12454 to "%-39s XXX". The latter will always guarantee a space after the
12455 field while the former will cause them to run together if the field
12456 is 40 of more characters long.
12460 * Constify the cipher and digest 'method' functions and structures
12461 and modify related functions to take constant EVP_MD and EVP_CIPHER
12466 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12467 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12471 * Modify `EVP_Digest*()` routines so they now return values. Although the
12472 internal software routines can never fail additional hardware versions
12477 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12479 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12480 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12483 ERR_R_NESTED_ASN1_ERROR
12485 ERR_R_MISSING_ASN1_EOS
12486 were 4 .. 9, conflicting with
12487 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12489 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12490 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12492 Add new error code 'ERR_R_INTERNAL_ERROR'.
12496 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12501 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12502 sets the subject name for a new request or supersedes the
12503 subject name in a given request. Formats that can be parsed are
12504 'CN=Some Name, OU=myOU, C=IT'
12506 'CN=Some Name/OU=myOU/C=IT'.
12508 Add options '-batch' and '-verbose' to 'openssl req'.
12510 *Massimiliano Pala <madwolf@hackmasters.net>*
12512 * Introduce the possibility to access global variables through
12513 functions on platform were that's the best way to handle exporting
12514 global variables in shared libraries. To enable this functionality,
12515 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12516 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12517 is normally done by Configure or something similar).
12519 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12520 in the source file (foo.c) like this:
12522 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12523 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12525 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12526 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12528 OPENSSL_DECLARE_GLOBAL(int,foo);
12529 #define foo OPENSSL_GLOBAL_REF(foo)
12530 OPENSSL_DECLARE_GLOBAL(double,bar);
12531 #define bar OPENSSL_GLOBAL_REF(bar)
12533 The #defines are very important, and therefore so is including the
12534 header file everywhere where the defined globals are used.
12536 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12537 of ASN.1 items, but that structure is a bit different.
12539 The largest change is in util/mkdef.pl which has been enhanced with
12540 better and easier to understand logic to choose which symbols should
12541 go into the Windows .def files as well as a number of fixes and code
12542 cleanup (among others, algorithm keywords are now sorted
12543 lexicographically to avoid constant rewrites).
12547 * In BN_div() keep a copy of the sign of 'num' before writing the
12548 result to 'rm' because if rm==num the value will be overwritten
12549 and produce the wrong result if 'num' is negative: this caused
12550 problems with BN_mod() and BN_nnmod().
12554 * Function OCSP_request_verify(). This checks the signature on an
12555 OCSP request and verifies the signer certificate. The signer
12556 certificate is just checked for a generic purpose and OCSP request
12561 * Add OCSP_check_validity() function to check the validity of OCSP
12562 responses. OCSP responses are prepared in real time and may only
12563 be a few seconds old. Simply checking that the current time lies
12564 between thisUpdate and nextUpdate max reject otherwise valid responses
12565 caused by either OCSP responder or client clock inaccuracy. Instead
12566 we allow thisUpdate and nextUpdate to fall within a certain period of
12567 the current time. The age of the response can also optionally be
12568 checked. Two new options -validity_period and -status_age added to
12573 * If signature or public key algorithm is unrecognized print out its
12574 OID rather that just UNKNOWN.
12578 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12579 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12580 ID to be generated from the issuer certificate alone which can then be
12581 passed to OCSP_id_issuer_cmp().
12585 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12586 ASN1 modules to export functions returning ASN1_ITEM pointers
12587 instead of the ASN1_ITEM structures themselves. This adds several
12588 new macros which allow the underlying ASN1 function/structure to
12589 be accessed transparently. As a result code should not use ASN1_ITEM
12590 references directly (such as &X509_it) but instead use the relevant
12591 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12592 use of the new ASN1 code on platforms where exporting structures
12593 is problematical (for example in shared libraries) but exporting
12594 functions returning pointers to structures is not.
12598 * Add support for overriding the generation of SSL/TLS session IDs.
12599 These callbacks can be registered either in an SSL_CTX or per SSL.
12600 The purpose of this is to allow applications to control, if they wish,
12601 the arbitrary values chosen for use as session IDs, particularly as it
12602 can be useful for session caching in multiple-server environments. A
12603 command-line switch for testing this (and any client code that wishes
12604 to use such a feature) has been added to "s_server".
12606 *Geoff Thorpe, Lutz Jaenicke*
12608 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12609 of the form `#if defined(...) || defined(...) || ...` and
12610 `#if !defined(...) && !defined(...) && ...`. This also avoids
12611 the growing number of special cases it was previously handling.
12615 * Make all configuration macros available for application by making
12616 sure they are available in opensslconf.h, by giving them names starting
12617 with `OPENSSL_` to avoid conflicts with other packages and by making
12618 sure e_os2.h will cover all platform-specific cases together with
12620 Additionally, it is now possible to define configuration/platform-
12621 specific names (called "system identities"). In the C code, these
12622 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12623 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12624 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12629 * New option -set_serial to 'req' and 'x509' this allows the serial
12630 number to use to be specified on the command line. Previously self
12631 signed certificates were hard coded with serial number 0 and the
12632 CA options of 'x509' had to use a serial number in a file which was
12637 * New options to 'ca' utility to support V2 CRL entry extensions.
12638 Currently CRL reason, invalidity date and hold instruction are
12639 supported. Add new CRL extensions to V3 code and some new objects.
12643 * New function EVP_CIPHER_CTX_set_padding() this is used to
12644 disable standard block padding (aka PKCS#5 padding) in the EVP
12645 API, which was previously mandatory. This means that the data is
12646 not padded in any way and so the total length much be a multiple
12647 of the block size, otherwise an error occurs.
12651 * Initial (incomplete) OCSP SSL support.
12655 * New function OCSP_parse_url(). This splits up a URL into its host,
12656 port and path components: primarily to parse OCSP URLs. New -url
12657 option to ocsp utility.
12661 * New nonce behavior. The return value of OCSP_check_nonce() now
12662 reflects the various checks performed. Applications can decide
12663 whether to tolerate certain situations such as an absent nonce
12664 in a response when one was present in a request: the ocsp application
12665 just prints out a warning. New function OCSP_add1_basic_nonce()
12666 this is to allow responders to include a nonce in a response even if
12667 the request is nonce-less.
12671 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12672 skipped when using openssl x509 multiple times on a single input file,
12673 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12677 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12678 set string type: to handle setting ASN1_TIME structures. Fix ca
12679 utility to correctly initialize revocation date of CRLs.
12683 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12684 the clients preferred ciphersuites and rather use its own preferences.
12685 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12686 Internet Explorer by ensuring unchanged hash method during stepup.
12687 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12691 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12692 to aes and add a new 'exist' option to print out symbols that don't
12697 * Additional options to ocsp utility to allow flags to be set and
12698 additional certificates supplied.
12702 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12703 OCSP client a number of certificate to only verify the response
12708 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12709 handle the new API. Currently only ECB, CBC modes supported. Add new
12712 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12713 Encryption Standard (AES) Ciphersuites for Transport Layer
12714 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12715 not enabled by default and were not part of the "ALL" ciphersuite
12716 alias because they were not yet official; they could be
12717 explicitly requested by specifying the "AESdraft" ciphersuite
12718 group alias. In the final release of OpenSSL 0.9.7, the group
12719 alias is called "AES" and is part of "ALL".)
12721 *Ben Laurie, Steve Henson, Bodo Moeller*
12723 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12724 request to response.
12728 * Functions for OCSP responders. OCSP_request_onereq_count(),
12729 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12730 extract information from a certificate request. OCSP_response_create()
12731 creates a response and optionally adds a basic response structure.
12732 OCSP_basic_add1_status() adds a complete single response to a basic
12733 response and returns the OCSP_SINGLERESP structure just added (to allow
12734 extensions to be included for example). OCSP_basic_add1_cert() adds a
12735 certificate to a basic response and OCSP_basic_sign() signs a basic
12736 response with various flags. New helper functions ASN1_TIME_check()
12737 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12738 (converts ASN1_TIME to GeneralizedTime).
12742 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12743 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12744 structure from a certificate. X509_pubkey_digest() digests the public_key
12745 contents: this is used in various key identifiers.
12749 * Make sk_sort() tolerate a NULL argument.
12751 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12753 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12754 passed by the function are trusted implicitly. If any of them signed the
12755 response then it is assumed to be valid and is not verified.
12759 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12760 to data. This was previously part of the PKCS7 ASN1 code. This
12761 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12762 *Steve Henson, reported by Kenneth R. Robinette
12763 <support@securenetterm.com>*
12765 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12766 routines: without these tracing memory leaks is very painful.
12767 Fix leaks in PKCS12 and PKCS7 routines.
12771 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12772 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12773 effectively meant GeneralizedTime would never be used. Now it
12774 is initialised to -1 but X509_time_adj() now has to check the value
12775 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12776 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12777 *Steve Henson, reported by Kenneth R. Robinette
12778 <support@securenetterm.com>*
12780 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12781 result in a zero length in the ASN1_INTEGER structure which was
12782 not consistent with the structure when d2i_ASN1_INTEGER() was used
12783 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12784 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12785 where it did not print out a minus for negative ASN1_INTEGER.
12789 * Add summary printout to ocsp utility. The various functions which
12790 convert status values to strings have been renamed to:
12791 OCSP_response_status_str(), OCSP_cert_status_str() and
12792 OCSP_crl_reason_str() and are no longer static. New options
12793 to verify nonce values and to disable verification. OCSP response
12794 printout format cleaned up.
12798 * Add additional OCSP certificate checks. These are those specified
12799 in RFC2560. This consists of two separate checks: the CA of the
12800 certificate being checked must either be the OCSP signer certificate
12801 or the issuer of the OCSP signer certificate. In the latter case the
12802 OCSP signer certificate must contain the OCSP signing extended key
12803 usage. This check is performed by attempting to match the OCSP
12804 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12805 in the OCSP_CERTID structures of the response.
12809 * Initial OCSP certificate verification added to OCSP_basic_verify()
12810 and related routines. This uses the standard OpenSSL certificate
12811 verify routines to perform initial checks (just CA validity) and
12812 to obtain the certificate chain. Then additional checks will be
12813 performed on the chain. Currently the root CA is checked to see
12814 if it is explicitly trusted for OCSP signing. This is used to set
12815 a root CA as a global signing root: that is any certificate that
12816 chains to that CA is an acceptable OCSP signing certificate.
12820 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12821 extensions from a separate configuration file.
12822 As when reading extensions from the main configuration file,
12823 the '-extensions ...' option may be used for specifying the
12826 *Massimiliano Pala <madwolf@comune.modena.it>*
12828 * New OCSP utility. Allows OCSP requests to be generated or
12829 read. The request can be sent to a responder and the output
12830 parsed, outputted or printed in text form. Not complete yet:
12831 still needs to check the OCSP response validity.
12835 * New subcommands for 'openssl ca':
12836 `openssl ca -status <serial>` prints the status of the cert with
12837 the given serial number (according to the index file).
12838 `openssl ca -updatedb` updates the expiry status of certificates
12841 *Massimiliano Pala <madwolf@comune.modena.it>*
12843 * New '-newreq-nodes' command option to CA.pl. This is like
12844 '-newreq', but calls 'openssl req' with the '-nodes' option
12845 so that the resulting key is not encrypted.
12847 *Damien Miller <djm@mindrot.org>*
12849 * New configuration for the GNU Hurd.
12851 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12853 * Initial code to implement OCSP basic response verify. This
12854 is currently incomplete. Currently just finds the signer's
12855 certificate and verifies the signature on the response.
12859 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12860 value of OPENSSLDIR. This is available via the new '-d' option
12861 to 'openssl version', and is also included in 'openssl version -a'.
12865 * Allowing defining memory allocation callbacks that will be given
12866 file name and line number information in additional arguments
12867 (a `const char*` and an int). The basic functionality remains, as
12868 well as the original possibility to just replace malloc(),
12869 realloc() and free() by functions that do not know about these
12870 additional arguments. To register and find out the current
12871 settings for extended allocation functions, the following
12872 functions are provided:
12874 CRYPTO_set_mem_ex_functions
12875 CRYPTO_set_locked_mem_ex_functions
12876 CRYPTO_get_mem_ex_functions
12877 CRYPTO_get_locked_mem_ex_functions
12879 These work the same way as CRYPTO_set_mem_functions and friends.
12880 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12881 extended allocation function is enabled.
12882 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12883 a conventional allocation function is enabled.
12885 *Richard Levitte, Bodo Moeller*
12887 * Finish off removing the remaining LHASH function pointer casts.
12888 There should no longer be any prototype-casting required when using
12889 the LHASH abstraction, and any casts that remain are "bugs". See
12890 the callback types and macros at the head of lhash.h for details
12891 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12895 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12896 If /dev/[u]random devices are not available or do not return enough
12897 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12899 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12900 /etc/entropy will be queried once each in this sequence, querying stops
12901 when enough entropy was collected without querying more sockets.
12905 * Change the Unix RAND_poll() variant to be able to poll several
12906 random devices, as specified by DEVRANDOM, until a sufficient amount
12907 of data has been collected. We spend at most 10 ms on each file
12908 (select timeout) and read in non-blocking mode. DEVRANDOM now
12909 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12910 (previously it was just the string "/dev/urandom"), so on typical
12911 platforms the 10 ms delay will never occur.
12912 Also separate out the Unix variant to its own file, rand_unix.c.
12913 For VMS, there's a currently-empty rand_vms.c.
12917 * Move OCSP client related routines to ocsp_cl.c. These
12918 provide utility functions which an application needing
12919 to issue a request to an OCSP responder and analyse the
12920 response will typically need: as opposed to those which an
12921 OCSP responder itself would need which will be added later.
12923 OCSP_request_sign() signs an OCSP request with an API similar
12924 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12925 response. OCSP_response_get1_basic() extracts basic response
12926 from response. OCSP_resp_find_status(): finds and extracts status
12927 information from an OCSP_CERTID structure (which will be created
12928 when the request structure is built). These are built from lower
12929 level functions which work on OCSP_SINGLERESP structures but
12930 won't normally be used unless the application wishes to examine
12931 extensions in the OCSP response for example.
12933 Replace nonce routines with a pair of functions.
12934 OCSP_request_add1_nonce() adds a nonce value and optionally
12935 generates a random value. OCSP_check_nonce() checks the
12936 validity of the nonce in an OCSP response.
12940 * Change function OCSP_request_add() to OCSP_request_add0_id().
12941 This doesn't copy the supplied OCSP_CERTID and avoids the
12942 need to free up the newly created id. Change return type
12943 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12944 This can then be used to add extensions to the request.
12945 Deleted OCSP_request_new(), since most of its functionality
12946 is now in OCSP_REQUEST_new() (and the case insensitive name
12947 clash) apart from the ability to set the request name which
12948 will be added elsewhere.
12952 * Update OCSP API. Remove obsolete extensions argument from
12953 various functions. Extensions are now handled using the new
12954 OCSP extension code. New simple OCSP HTTP function which
12955 can be used to send requests and parse the response.
12959 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12960 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12961 uses the special reorder version of SET OF to sort the attributes
12962 and reorder them to match the encoded order. This resolves a long
12963 standing problem: a verify on a PKCS7 structure just after signing
12964 it used to fail because the attribute order did not match the
12965 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12966 it uses the received order. This is necessary to tolerate some broken
12967 software that does not order SET OF. This is handled by encoding
12968 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12969 to produce the required SET OF.
12973 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12974 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12975 files to get correct declarations of the ASN.1 item variables.
12979 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12980 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12981 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12982 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12983 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12984 ASN1_ITEM and no wrapper functions.
12988 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12989 replace the old function pointer based I/O routines. Change most of
12990 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12994 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12995 lines, recognize more "algorithms" that can be deselected, and make
12996 it complain about algorithm deselection that isn't recognised.
13000 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13001 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13002 to use new functions. Add NO_ASN1_OLD which can be set to remove
13003 some old style ASN1 functions: this can be used to determine if old
13004 code will still work when these eventually go away.
13008 * New extension functions for OCSP structures, these follow the
13009 same conventions as certificates and CRLs.
13013 * New function X509V3_add1_i2d(). This automatically encodes and
13014 adds an extension. Its behaviour can be customised with various
13015 flags to append, replace or delete. Various wrappers added for
13016 certificates and CRLs.
13020 * Fix to avoid calling the underlying ASN1 print routine when
13021 an extension cannot be parsed. Correct a typo in the
13022 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13026 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13027 entries for variables.
13031 * Add functionality to apps/openssl.c for detecting locking
13032 problems: As the program is single-threaded, all we have
13033 to do is register a locking callback using an array for
13034 storing which locks are currently held by the program.
13038 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13039 SSL_get_ex_data_X509_STORE_idx(), which is used in
13040 ssl_verify_cert_chain() and thus can be called at any time
13041 during TLS/SSL handshakes so that thread-safety is essential.
13042 Unfortunately, the ex_data design is not at all suited
13043 for multi-threaded use, so it probably should be abolished.
13047 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13049 *Broadcom, tweaked and integrated by Geoff Thorpe*
13051 * Move common extension printing code to new function
13052 X509V3_print_extensions(). Reorganise OCSP print routines and
13053 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13057 * New function X509_signature_print() to remove duplication in some
13062 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13063 set (this was treated exactly the same as SET OF previously). This
13064 is used to reorder the STACK representing the structure to match the
13065 encoding. This will be used to get round a problem where a PKCS7
13066 structure which was signed could not be verified because the STACK
13067 order did not reflect the encoded order.
13071 * Reimplement the OCSP ASN1 module using the new code.
13075 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13076 for its ASN1 operations. The old style function pointers still exist
13077 for now but they will eventually go away.
13081 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13082 completely replaces the old ASN1 functionality with a table driven
13083 encoder and decoder which interprets an ASN1_ITEM structure describing
13084 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13085 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13086 has also been converted to the new form.
13090 * Change BN_mod_exp_recp so that negative moduli are tolerated
13091 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13092 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13093 for negative moduli.
13097 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13098 of not touching the result's sign bit.
13102 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13107 * Changed the LHASH code to use prototypes for callbacks, and created
13108 macros to declare and implement thin (optionally static) functions
13109 that provide type-safety and avoid function pointer casting for the
13110 type-specific callbacks.
13114 * Added Kerberos Cipher Suites to be used with TLS, as written in
13116 *Veers Staats <staatsvr@asc.hpc.mil>,
13117 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13119 * Reformat the FAQ so the different questions and answers can be divided
13120 in sections depending on the subject.
13124 * Have the zlib compression code load ZLIB.DLL dynamically under
13129 * New function BN_mod_sqrt for computing square roots modulo a prime
13130 (using the probabilistic Tonelli-Shanks algorithm unless
13131 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13132 be handled deterministically).
13134 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13136 * Make BN_mod_inverse faster by explicitly handling small quotients
13137 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13138 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13142 * New function BN_kronecker.
13146 * Fix BN_gcd so that it works on negative inputs; the result is
13147 positive unless both parameters are zero.
13148 Previously something reasonably close to an infinite loop was
13149 possible because numbers could be growing instead of shrinking
13150 in the implementation of Euclid's algorithm.
13154 * Fix BN_is_word() and BN_is_one() macros to take into account the
13155 sign of the number in question.
13157 Fix BN_is_word(a,w) to work correctly for w == 0.
13159 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13160 because its test if the absolute value of 'a' equals 'w'.
13161 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13162 it exists mostly for use in the implementations of BN_is_zero(),
13163 BN_is_one(), and BN_is_word().
13167 * New function BN_swap.
13171 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13172 the exponentiation functions are more likely to produce reasonable
13173 results on negative inputs.
13177 * Change BN_mod_mul so that the result is always non-negative.
13178 Previously, it could be negative if one of the factors was negative;
13179 I don't think anyone really wanted that behaviour.
13183 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13184 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13185 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13186 and add new functions:
13195 BN_mod_lshift1_quick
13197 BN_mod_lshift_quick
13199 These functions always generate non-negative results.
13201 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13202 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13204 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13205 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13206 be reduced modulo `m`.
13208 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13211 The following entry accidentally appeared in the CHANGES file
13212 distributed with OpenSSL 0.9.7. The modifications described in
13213 it do *not* apply to OpenSSL 0.9.7.
13215 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13216 was actually never needed) and in BN_mul(). The removal in BN_mul()
13217 required a small change in bn_mul_part_recursive() and the addition
13218 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13219 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13220 bn_sub_words() and bn_add_words() except they take arrays with
13226 * In 'openssl passwd', verify passwords read from the terminal
13227 unless the '-salt' option is used (which usually means that
13228 verification would just waste user's time since the resulting
13229 hash is going to be compared with some given password hash)
13230 or the new '-noverify' option is used.
13232 This is an incompatible change, but it does not affect
13233 non-interactive use of 'openssl passwd' (passwords on the command
13234 line, '-stdin' option, '-in ...' option) and thus should not
13235 cause any problems.
13239 * Remove all references to RSAref, since there's no more need for it.
13243 * Make DSO load along a path given through an environment variable
13244 (SHLIB_PATH) with shl_load().
13248 * Constify the ENGINE code as a result of BIGNUM constification.
13249 Also constify the RSA code and most things related to it. In a
13250 few places, most notable in the depth of the ASN.1 code, ugly
13251 casts back to non-const were required (to be solved at a later
13256 * Make it so the openssl application has all engines loaded by default.
13260 * Constify the BIGNUM routines a little more.
13264 * Add the following functions:
13266 ENGINE_load_cswift()
13268 ENGINE_load_atalla()
13269 ENGINE_load_nuron()
13270 ENGINE_load_builtin_engines()
13272 That way, an application can itself choose if external engines that
13273 are built-in in OpenSSL shall ever be used or not. The benefit is
13274 that applications won't have to be linked with libdl or other dso
13275 libraries unless it's really needed.
13277 Changed 'openssl engine' to load all engines on demand.
13278 Changed the engine header files to avoid the duplication of some
13279 declarations (they differed!).
13283 * 'openssl engine' can now list capabilities.
13287 * Better error reporting in 'openssl engine'.
13291 * Never call load_dh_param(NULL) in s_server.
13295 * Add engine application. It can currently list engines by name and
13296 identity, and test if they are actually available.
13300 * Improve RPM specification file by forcing symbolic linking and making
13301 sure the installed documentation is also owned by root.root.
13303 *Damien Miller <djm@mindrot.org>*
13305 * Give the OpenSSL applications more possibilities to make use of
13306 keys (public as well as private) handled by engines.
13310 * Add OCSP code that comes from CertCo.
13314 * Add VMS support for the Rijndael code.
13318 * Added untested support for Nuron crypto accelerator.
13322 * Add support for external cryptographic devices. This code was
13323 previously distributed separately as the "engine" branch.
13325 *Geoff Thorpe, Richard Levitte*
13327 * Rework the filename-translation in the DSO code. It is now possible to
13328 have far greater control over how a "name" is turned into a filename
13329 depending on the operating environment and any oddities about the
13330 different shared library filenames on each system.
13334 * Support threads on FreeBSD-elf in Configure.
13338 * Fix for SHA1 assembly problem with MASM: it produces
13339 warnings about corrupt line number information when assembling
13340 with debugging information. This is caused by the overlapping
13343 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13346 NCONF_get_number() has no error checking at all. As a replacement,
13347 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13348 promoted strongly. The old NCONF_get_number is kept around for
13349 binary backward compatibility.
13350 Make it possible for methods to load from something other than a BIO,
13351 by providing a function pointer that is given a name instead of a BIO.
13352 For example, this could be used to load configuration data from an
13357 * Fix for non blocking accept BIOs. Added new I/O special reason
13358 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13359 with non blocking I/O was not possible because no retry code was
13360 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13365 * Added the beginnings of Rijndael support.
13369 * Fix for bug in DirectoryString mask setting. Add support for
13370 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13371 to allow certificate printing to more controllable, additional
13372 'certopt' option to 'x509' to allow new printing options to be
13377 * Clean old EAY MD5 hack from e_os.h.
13381 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13383 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13384 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13386 *Joe Orton, Steve Henson*
13388 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13390 * Fix additional bug revealed by the NISCC test suite:
13392 Stop bug triggering large recursion when presented with
13393 certain ASN.1 tags ([CVE-2003-0851])
13397 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13399 * Fix various bugs revealed by running the NISCC test suite:
13401 Stop out of bounds reads in the ASN1 code when presented with
13402 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13404 If verify callback ignores invalid public key errors don't try to check
13405 certificate signature with the NULL public key.
13409 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13410 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13415 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13416 extra data after the compression methods not only for TLS 1.0
13417 but also for SSL 3.0 (as required by the specification).
13419 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13421 * Change X509_certificate_type() to mark the key as exported/exportable
13422 when it's 512 *bits* long, not 512 bytes.
13426 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13428 * Countermeasure against the Klima-Pokorny-Rosa extension of
13429 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13430 a protocol version number mismatch like a decryption error
13431 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13435 * Turn on RSA blinding by default in the default implementation
13436 to avoid a timing attack. Applications that don't want it can call
13437 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13438 They would be ill-advised to do so in most cases.
13440 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13442 * Change RSA blinding code so that it works when the PRNG is not
13443 seeded (in this case, the secret RSA exponent is abused as
13444 an unpredictable seed -- if it is not unpredictable, there
13445 is no point in blinding anyway). Make RSA blinding thread-safe
13446 by remembering the creator's thread ID in rsa->blinding and
13447 having all other threads use local one-time blinding factors
13448 (this requires more computation than sharing rsa->blinding, but
13449 avoids excessive locking; and if an RSA object is not shared
13450 between threads, blinding will still be very fast).
13454 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13456 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13457 via timing by performing a MAC computation even if incorrect
13458 block cipher padding has been found. This is a countermeasure
13459 against active attacks where the attacker has to distinguish
13460 between bad padding and a MAC verification error. ([CVE-2003-0078])
13462 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13463 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13464 Martin Vuagnoux (EPFL, Ilion)*
13466 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13468 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13469 memory from its contents. This is done with a counter that will
13470 place alternating values in each byte. This can be used to solve
13471 two issues: 1) the removal of calls to memset() by highly optimizing
13472 compilers, and 2) cleansing with other values than 0, since those can
13473 be read through on certain media, for example a swap space on disk.
13477 * Bugfix: client side session caching did not work with external caching,
13478 because the session->cipher setting was not restored when reloading
13479 from the external cache. This problem was masked, when
13480 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13481 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13485 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13486 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13488 *Zeev Lieber <zeev-l@yahoo.com>*
13490 * Undo an undocumented change introduced in 0.9.6e which caused
13491 repeated calls to OpenSSL_add_all_ciphers() and
13492 OpenSSL_add_all_digests() to be ignored, even after calling
13497 * Change the default configuration reader to deal with last line not
13498 being properly terminated.
13502 * Change X509_NAME_cmp() so it applies the special rules on handling
13503 DN values that are of type PrintableString, as well as RDNs of type
13504 emailAddress where the value has the type ia5String.
13506 *stefank@valicert.com via Richard Levitte*
13508 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13509 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13510 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13511 the bitwise-OR of the two for use by the majority of applications
13512 wanting this behaviour, and update the docs. The documented
13513 behaviour and actual behaviour were inconsistent and had been
13514 changing anyway, so this is more a bug-fix than a behavioural
13517 *Geoff Thorpe, diagnosed by Nadav Har'El*
13519 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13520 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13524 * Fix initialization code race conditions in
13525 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13526 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13527 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13528 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13529 ssl2_get_cipher_by_char(),
13530 ssl3_get_cipher_by_char().
13532 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13534 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13535 the cached sessions are flushed, as the remove_cb() might use ex_data
13536 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13537 (see [openssl.org #212]).
13539 *Geoff Thorpe, Lutz Jaenicke*
13541 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13542 length, instead of the encoding length to d2i_ASN1_OBJECT.
13546 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13548 * [In 0.9.6g-engine release:]
13549 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13551 *Lynn Gazis <lgazis@rainbow.com>*
13553 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13555 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13556 and get fix the header length calculation.
13557 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13558 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13560 * Use proper error handling instead of 'assertions' in buffer
13561 overflow checks added in 0.9.6e. This prevents DoS (the
13562 assertions could call abort()).
13564 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13566 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13568 * Add various sanity checks to asn1_get_length() to reject
13569 the ASN1 length bytes if they exceed sizeof(long), will appear
13570 negative or the content length exceeds the length of the
13573 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13575 * Fix cipher selection routines: ciphers without encryption had no flags
13576 for the cipher strength set and where therefore not handled correctly
13577 by the selection routines (PR #130).
13581 * Fix EVP_dsa_sha macro.
13586 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13587 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13588 that was added in OpenSSL 0.9.6d.
13590 As the countermeasure turned out to be incompatible with some
13591 broken SSL implementations, the new option is part of SSL_OP_ALL.
13592 SSL_OP_ALL is usually employed when compatibility with weird SSL
13593 implementations is desired (e.g. '-bugs' option to 's_client' and
13594 's_server'), so the new option is automatically set in many
13599 * Changes in security patch:
13601 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13602 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13603 Air Force Materiel Command, USAF, under agreement number
13606 * Add various sanity checks to asn1_get_length() to reject
13607 the ASN1 length bytes if they exceed sizeof(long), will appear
13608 negative or the content length exceeds the length of the
13609 supplied buffer. ([CVE-2002-0659])
13611 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13613 * Assertions for various potential buffer overflows, not known to
13614 happen in practice.
13616 *Ben Laurie (CHATS)*
13618 * Various temporary buffers to hold ASCII versions of integers were
13619 too small for 64 bit platforms. ([CVE-2002-0655])
13620 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13622 * Remote buffer overflow in SSL3 protocol - an attacker could
13623 supply an oversized session ID to a client. ([CVE-2002-0656])
13625 *Ben Laurie (CHATS)*
13627 * Remote buffer overflow in SSL2 protocol - an attacker could
13628 supply an oversized client master key. ([CVE-2002-0656])
13630 *Ben Laurie (CHATS)*
13632 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13634 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13635 encoded as NULL) with id-dsa-with-sha1.
13637 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13639 * Check various `X509_...()` return values in apps/req.c.
13641 *Nils Larsch <nla@trustcenter.de>*
13643 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13644 an end-of-file condition would erroneously be flagged, when the CRLF
13645 was just at the end of a processed block. The bug was discovered when
13646 processing data through a buffering memory BIO handing the data to a
13647 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13648 <ptsekov@syntrex.com> and Nedelcho Stanev.
13652 * Implement a countermeasure against a vulnerability recently found
13653 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13654 before application data chunks to avoid the use of known IVs
13655 with data potentially chosen by the attacker.
13659 * Fix length checks in ssl3_get_client_hello().
13663 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13664 to prevent ssl3_read_internal() from incorrectly assuming that
13665 ssl3_read_bytes() found application data while handshake
13666 processing was enabled when in fact s->s3->in_read_app_data was
13667 merely automatically cleared during the initial handshake.
13669 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13671 * Fix object definitions for Private and Enterprise: they were not
13672 recognized in their shortname (=lowercase) representation. Extend
13673 obj_dat.pl to issue an error when using undefined keywords instead
13674 of silently ignoring the problem (Svenning Sorensen
13675 <sss@sss.dnsalias.net>).
13679 * Fix DH_generate_parameters() so that it works for 'non-standard'
13680 generators, i.e. generators other than 2 and 5. (Previously, the
13681 code did not properly initialise the 'add' and 'rem' values to
13682 BN_generate_prime().)
13684 In the new general case, we do not insist that 'generator' is
13685 actually a primitive root: This requirement is rather pointless;
13686 a generator of the order-q subgroup is just as good, if not
13691 * Map new X509 verification errors to alerts. Discovered and submitted by
13692 Tom Wu <tom@arcot.com>.
13696 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13697 returning non-zero before the data has been completely received
13698 when using non-blocking I/O.
13700 *Bodo Moeller; problem pointed out by John Hughes*
13702 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13704 *Ben Laurie, Lutz Jaenicke*
13706 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13707 Yoram Zahavi <YoramZ@gilian.com>).
13711 * Add information about CygWin 1.3 and on, and preserve proper
13712 configuration for the versions before that.
13714 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13716 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13717 check whether we deal with a copy of a session and do not delete from
13718 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13719 <izhar@checkpoint.com>.
13723 * Do not store session data into the internal session cache, if it
13724 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13725 flag is set). Proposed by Aslam <aslam@funk.com>.
13729 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13734 * [In 0.9.6d-engine release:]
13735 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13737 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13739 * Add the configuration target linux-s390x.
13741 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13743 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13744 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13745 variable as an indication that a ClientHello message has been
13746 received. As the flag value will be lost between multiple
13747 invocations of ssl3_accept when using non-blocking I/O, the
13748 function may not be aware that a handshake has actually taken
13749 place, thus preventing a new session from being added to the
13752 To avoid this problem, we now set s->new_session to 2 instead of
13753 using a local variable.
13755 *Lutz Jaenicke, Bodo Moeller*
13757 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13758 if the SSL_R_LENGTH_MISMATCH error is detected.
13760 *Geoff Thorpe, Bodo Moeller*
13762 * New 'shared_ldflag' column in Configure platform table.
13766 * Fix EVP_CIPHER_mode macro.
13768 *"Dan S. Camper" <dan@bti.net>*
13770 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13771 type, we must throw them away by setting rr->length to 0.
13773 *D P Chang <dpc@qualys.com>*
13775 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13777 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13778 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13779 worked incorrectly for those cases where range = `10..._2` and
13780 `3*range` is two bits longer than range.)
13784 * Only add signing time to PKCS7 structures if it is not already
13789 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13790 OBJ_ld_ce should be OBJ_id_ce.
13791 Also some ip-pda OIDs in crypto/objects/objects.txt were
13792 incorrect (cf. RFC 3039).
13794 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13796 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13797 returns early because it has nothing to do.
13799 *Andy Schneider <andy.schneider@bjss.co.uk>*
13801 * [In 0.9.6c-engine release:]
13802 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13804 *Andy Schneider <andy.schneider@bjss.co.uk>*
13806 * [In 0.9.6c-engine release:]
13807 Add support for Cryptographic Appliance's keyserver technology.
13808 (Use engine 'keyclient')
13810 *Cryptographic Appliances and Geoff Thorpe*
13812 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13813 is called via tools/c89.sh because arguments have to be
13814 rearranged (all '-L' options must appear before the first object
13817 *Richard Shapiro <rshapiro@abinitio.com>*
13819 * [In 0.9.6c-engine release:]
13820 Add support for Broadcom crypto accelerator cards, backported
13823 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13825 * [In 0.9.6c-engine release:]
13826 Add support for SureWare crypto accelerator cards from
13827 Baltimore Technologies. (Use engine 'sureware')
13829 *Baltimore Technologies and Mark Cox*
13831 * [In 0.9.6c-engine release:]
13832 Add support for crypto accelerator cards from Accelerated
13833 Encryption Processing, www.aep.ie. (Use engine 'aep')
13835 *AEP Inc. and Mark Cox*
13837 * Add a configuration entry for gcc on UnixWare.
13839 *Gary Benson <gbenson@redhat.com>*
13841 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13842 messages are stored in a single piece (fixed-length part and
13843 variable-length part combined) and fix various bugs found on the way.
13847 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13848 instead. BIO_gethostbyname() does not know what timeouts are
13849 appropriate, so entries would stay in cache even when they have
13851 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13853 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13854 faced with a pathologically small ClientHello fragment that does
13855 not contain client_version: Instead of aborting with an error,
13856 simply choose the highest available protocol version (i.e.,
13857 TLS 1.0 unless it is disabled). In practice, ClientHello
13858 messages are never sent like this, but this change gives us
13859 strictly correct behaviour at least for TLS.
13863 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13864 never resets s->method to s->ctx->method when called from within
13865 one of the SSL handshake functions.
13867 *Bodo Moeller; problem pointed out by Niko Baric*
13869 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13870 (sent using the client's version number) if client_version is
13871 smaller than the protocol version in use. Also change
13872 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13873 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13874 the client will at least see that alert.
13878 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13883 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13884 client receives HelloRequest while in a handshake.
13886 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13888 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13889 should end in 'break', not 'goto end' which circumvents various
13890 cleanups done in state SSL_ST_OK. But session related stuff
13891 must be disabled for SSL_ST_OK in the case that we just sent a
13894 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13895 before just sending a HelloRequest.
13897 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13899 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13900 reveal whether illegal block cipher padding was found or a MAC
13901 verification error occurred. (Neither SSLerr() codes nor alerts
13902 are directly visible to potential attackers, but the information
13903 may leak via logfiles.)
13905 Similar changes are not required for the SSL 2.0 implementation
13906 because the number of padding bytes is sent in clear for SSL 2.0,
13907 and the extra bytes are just ignored. However ssl/s2_pkt.c
13908 failed to verify that the purported number of padding bytes is in
13913 * Add OpenUNIX-8 support including shared libraries
13914 (Boyd Lynn Gerber <gerberb@zenez.com>).
13918 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13919 'wristwatch attack' using huge encoding parameters (cf.
13920 James H. Manger's CRYPTO 2001 paper). Note that the
13921 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13922 encoding parameters and hence was not vulnerable.
13926 * BN_sqr() bug fix.
13928 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13930 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13931 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13932 followed by modular reduction.
13934 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13936 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13937 equivalent based on BN_pseudo_rand() instead of BN_rand().
13941 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13942 This function was broken, as the check for a new client hello message
13943 to handle SGC did not allow these large messages.
13944 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13948 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13952 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13953 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13957 * Rework the configuration and shared library support for Tru64 Unix.
13958 The configuration part makes use of modern compiler features and
13959 still retains old compiler behavior for those that run older versions
13960 of the OS. The shared library support part includes a variant that
13961 uses the RPATH feature, and is available through the special
13962 configuration target "alpha-cc-rpath", which will never be selected
13965 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13967 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13968 with the same message size as in ssl3_get_certificate_request().
13969 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13970 messages might inadvertently be reject as too long.
13972 *Petr Lampa <lampa@fee.vutbr.cz>*
13974 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13978 * Modified SSL library such that the verify_callback that has been set
13979 specifically for an SSL object with SSL_set_verify() is actually being
13980 used. Before the change, a verify_callback set with this function was
13981 ignored and the verify_callback() set in the SSL_CTX at the time of
13982 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13983 to allow the necessary settings.
13987 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13988 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13989 done automatically (in contradiction to the requirements of the C
13990 standard). This made problems when used from OpenSSH.
13994 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13995 dh->length and always used
13997 BN_rand_range(priv_key, dh->p).
13999 BN_rand_range() is not necessary for Diffie-Hellman, and this
14000 specific range makes Diffie-Hellman unnecessarily inefficient if
14001 dh->length (recommended exponent length) is much smaller than the
14002 length of dh->p. We could use BN_rand_range() if the order of
14003 the subgroup was stored in the DH structure, but we only have
14008 BN_rand(priv_key, l, ...)
14010 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14017 RSA_eay_public_encrypt
14018 RSA_eay_private_decrypt
14019 RSA_eay_private_encrypt (signing)
14020 RSA_eay_public_decrypt (signature verification)
14022 (default implementations for RSA_public_encrypt,
14023 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14024 always reject numbers >= n.
14028 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14029 to synchronize access to 'locking_thread'. This is necessary on
14030 systems where access to 'locking_thread' (an 'unsigned long'
14031 variable) is not atomic.
14035 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14036 *before* setting the 'crypto_lock_rand' flag. The previous code had
14037 a race condition if 0 is a valid thread ID.
14039 *Travis Vitek <vitek@roguewave.com>*
14041 * Add support for shared libraries under Irix.
14043 *Albert Chin-A-Young <china@thewrittenword.com>*
14045 * Add configuration option to build on Linux on both big-endian and
14046 little-endian MIPS.
14048 *Ralf Baechle <ralf@uni-koblenz.de>*
14050 * Add the possibility to create shared libraries on HP-UX.
14054 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14056 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14057 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14058 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14059 PRNG state recovery was possible based on the output of
14060 one PRNG request appropriately sized to gain knowledge on
14061 'md' followed by enough consecutive 1-byte PRNG requests
14062 to traverse all of 'state'.
14064 1. When updating 'md_local' (the current thread's copy of 'md')
14065 during PRNG output generation, hash all of the previous
14066 'md_local' value, not just the half used for PRNG output.
14068 2. Make the number of bytes from 'state' included into the hash
14069 independent from the number of PRNG bytes requested.
14071 The first measure alone would be sufficient to avoid
14072 Markku-Juhani's attack. (Actually it had never occurred
14073 to me that the half of 'md_local' used for chaining was the
14074 half from which PRNG output bytes were taken -- I had always
14075 assumed that the secret half would be used.) The second
14076 measure makes sure that additional data from 'state' is never
14077 mixed into 'md_local' in small portions; this heuristically
14078 further strengthens the PRNG.
14082 * Fix crypto/bn/asm/mips3.s.
14086 * When only the key is given to "enc", the IV is undefined. Print out
14087 an error message in this case.
14091 * Handle special case when X509_NAME is empty in X509 printing routines.
14095 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14096 positive and less than q.
14100 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14101 used: it isn't thread safe and the add_lock_callback should handle
14104 *Paul Rose <Paul.Rose@bridge.com>*
14106 * Verify that incoming data obeys the block size in
14107 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14113 *Ulf Möller, Bodo Möller*
14115 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14116 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14117 when fixing the server behaviour for backwards-compatible 'client
14118 hello' messages. (Note that the attack is impractical against
14119 SSL 3.0 and TLS 1.0 anyway because length and version checking
14120 means that the probability of guessing a valid ciphertext is
14121 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14124 Before 0.9.5, the countermeasure (hide the error by generating a
14125 random 'decryption result') did not work properly because
14126 ERR_clear_error() was missing, meaning that SSL_get_error() would
14127 detect the supposedly ignored error.
14129 Both problems are now fixed.
14133 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14134 (previously it was 1024).
14138 * Fix for compatibility mode trust settings: ignore trust settings
14139 unless some valid trust or reject settings are present.
14143 * Fix for blowfish EVP: its a variable length cipher.
14147 * Fix various bugs related to DSA S/MIME verification. Handle missing
14148 parameters in DSA public key structures and return an error in the
14149 DSA routines if parameters are absent.
14153 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14154 in the current directory if neither $RANDFILE nor $HOME was set.
14155 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14156 caused some confusion to Windows users who haven't defined $HOME.
14157 Thus RAND_file_name() is changed again: e_os.h can define a
14158 DEFAULT_HOME, which will be used if $HOME is not set.
14159 For Windows, we use "C:"; on other platforms, we still require
14160 environment variables.
14162 * Move 'if (!initialized) RAND_poll()' into regions protected by
14163 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14164 having multiple threads call RAND_poll() concurrently.
14168 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14169 combination of a flag and a thread ID variable.
14170 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14171 flag), *other* threads can enter ssleay_add_bytes without obeying
14172 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14173 that they do not hold after the first thread unsets add_do_not_lock).
14177 * Change bctest again: '-x' expressions are not available in all
14178 versions of 'test'.
14182 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14184 * Fix a couple of memory leaks in PKCS7_dataDecode()
14186 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14188 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14189 the default extension for executables, if any. Also, make the perl
14190 scripts that use symlink() to test if it really exists and use "cp"
14191 if it doesn't. All this made OpenSSL compilable and installable in
14196 * Fix for asn1_GetSequence() for indefinite length constructed data.
14197 If SEQUENCE is length is indefinite just set c->slen to the total
14198 amount of data available.
14200 *Steve Henson, reported by shige@FreeBSD.org*
14202 *This change does not apply to 0.9.7.*
14204 * Change bctest to avoid here-documents inside command substitution
14205 (workaround for FreeBSD /bin/sh bug).
14206 For compatibility with Ultrix, avoid shell functions (introduced
14207 in the bctest version that searches along $PATH).
14211 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14212 with des_encrypt() defined on some operating systems, like Solaris
14217 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14218 On the Importance of Eliminating Errors in Cryptographic
14219 Computations, J. Cryptology 14 (2001) 2, 101-119,
14220 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14224 * MIPS assembler BIGNUM division bug fix.
14228 * Disabled incorrect Alpha assembler code.
14232 * Fix PKCS#7 decode routines so they correctly update the length
14233 after reading an EOC for the EXPLICIT tag.
14237 *This change does not apply to 0.9.7.*
14239 * Fix bug in PKCS#12 key generation routines. This was triggered
14240 if a 3DES key was generated with a 0 initial byte. Include
14241 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14242 (but broken) behaviour.
14246 * Enhance bctest to search for a working bc along $PATH and print
14249 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14251 * Fix memory leaks in err.c: free err_data string if necessary;
14252 don't write to the wrong index in ERR_set_error_data.
14256 * Implement ssl23_peek (analogous to ssl23_read), which previously
14261 * Replace rdtsc with `_emit` statements for VC++ version 5.
14263 *Jeremy Cooper <jeremy@baymoo.org>*
14265 * Make it possible to reuse SSLv2 sessions.
14269 * In copy_email() check for >= 0 as a return value for
14270 X509_NAME_get_index_by_NID() since 0 is a valid index.
14272 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14274 * Avoid coredump with unsupported or invalid public keys by checking if
14275 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14276 PKCS7_verify() fails with non detached data.
14280 * Don't use getenv in library functions when run as setuid/setgid.
14281 New function OPENSSL_issetugid().
14285 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14286 due to incorrect handling of multi-threading:
14288 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14290 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14292 3. Count how many times MemCheck_off() has been called so that
14293 nested use can be treated correctly. This also avoids
14294 inband-signalling in the previous code (which relied on the
14295 assumption that thread ID 0 is impossible).
14299 * Add "-rand" option also to s_client and s_server.
14303 * Fix CPU detection on Irix 6.x.
14304 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14305 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14307 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14312 *This change does not apply to 0.9.7.*
14314 * Use the cached encoding of an X509_NAME structure rather than
14315 copying it. This is apparently the reason for the libsafe "errors"
14316 but the code is actually correct.
14320 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14321 Bleichenbacher's DSA attack.
14322 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14323 to be set and top=0 forces the highest bit to be set; top=-1 is new
14324 and leaves the highest bit random.
14326 *Ulf Moeller, Bodo Moeller*
14328 * In the `NCONF_...`-based implementations for `CONF_...` queries
14329 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14330 a temporary CONF structure with the data component set to NULL
14331 (which gives segmentation faults in lh_retrieve).
14332 Instead, use NULL for the CONF pointer in CONF_get_string and
14333 CONF_get_number (which may use environment variables) and directly
14334 return NULL from CONF_get_section.
14338 * Fix potential buffer overrun for EBCDIC.
14342 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14343 keyUsage if basicConstraints absent for a CA.
14347 * Make SMIME_write_PKCS7() write mail header values with a format that
14348 is more generally accepted (no spaces before the semicolon), since
14349 some programs can't parse those values properly otherwise. Also make
14350 sure BIO's that break lines after each write do not create invalid
14355 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14356 macros previously used would not encode an empty SEQUENCE OF
14357 and break the signature.
14361 *This change does not apply to 0.9.7.*
14363 * Zero the premaster secret after deriving the master secret in
14368 * Add some EVP_add_digest_alias registrations (as found in
14369 OpenSSL_add_all_digests()) to SSL_library_init()
14370 aka OpenSSL_add_ssl_algorithms(). This provides improved
14371 compatibility with peers using X.509 certificates
14372 with unconventional AlgorithmIdentifier OIDs.
14376 * Fix for Irix with NO_ASM.
14378 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14380 * ./config script fixes.
14382 *Ulf Moeller, Richard Levitte*
14384 * Fix 'openssl passwd -1'.
14388 * Change PKCS12_key_gen_asc() so it can cope with non null
14389 terminated strings whose length is passed in the passlen
14390 parameter, for example from PEM callbacks. This was done
14391 by adding an extra length parameter to asc2uni().
14393 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14395 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14396 call failed, free the DSA structure.
14400 * Fix to uni2asc() to cope with zero length Unicode strings.
14401 These are present in some PKCS#12 files.
14405 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14406 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14407 when writing a 32767 byte record.
14409 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14411 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14412 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14414 (RSA objects have a reference count access to which is protected
14415 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14416 so they are meant to be shared between threads.)
14417 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14418 "Reddie, Steven" <Steven.Reddie@ca.com>*
14420 * Fix a deadlock in CRYPTO_mem_leaks().
14424 * Use better test patterns in bntest.
14428 * rand_win.c fix for Borland C.
14432 * BN_rshift bugfix for n == 0.
14436 * Add a 'bctest' script that checks for some known 'bc' bugs
14437 so that 'make test' does not abort just because 'bc' is broken.
14441 * Store verify_result within SSL_SESSION also for client side to
14442 avoid potential security hole. (Re-used sessions on the client side
14443 always resulted in verify_result==X509_V_OK, not using the original
14444 result of the server certificate verification.)
14448 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14449 SSL3_RT_APPLICATION_DATA, return 0.
14450 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14455 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14456 releases, have been re-implemented by renaming the previous
14457 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14458 and ssl3_read_internal, respectively, and adding 'peek' parameters
14459 to them. The new ssl[23]_{read,peek} functions are calls to
14460 ssl[23]_read_internal with the 'peek' flag set appropriately.
14461 A 'peek' parameter has also been added to ssl3_read_bytes, which
14462 does the actual work for ssl3_read_internal.
14466 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14467 the method-specific "init()" handler. Also clean up ex_data after
14468 calling the method-specific "finish()" handler. Previously, this was
14469 happening the other way round.
14473 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14474 The previous value, 12, was not always sufficient for BN_mod_exp().
14478 * Make sure that shared libraries get the internal name engine with
14479 the full version number and not just 0. This should mark the
14480 shared libraries as not backward compatible. Of course, this should
14481 be changed again when we can guarantee backward binary compatibility.
14485 * Fix typo in get_cert_by_subject() in by_dir.c
14487 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14489 * Rework the system to generate shared libraries:
14491 - Make note of the expected extension for the shared libraries and
14492 if there is a need for symbolic links from for example libcrypto.so.0
14493 to libcrypto.so.0.9.7. There is extended info in Configure for
14496 - Make as few rebuilds of the shared libraries as possible.
14498 - Still avoid linking the OpenSSL programs with the shared libraries.
14500 - When installing, install the shared libraries separately from the
14505 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14507 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14508 and not in SSL_clear because the latter is also used by the
14509 accept/connect functions; previously, the settings made by
14510 SSL_set_read_ahead would be lost during the handshake.
14512 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14514 * Correct util/mkdef.pl to be selective about disabled algorithms.
14515 Previously, it would create entries for disabled algorithms no
14520 * Added several new manual pages for SSL_* function.
14524 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14526 * In ssl23_get_client_hello, generate an error message when faced
14527 with an initial SSL 3.0/TLS record that is too small to contain the
14528 first two bytes of the ClientHello message, i.e. client_version.
14529 (Note that this is a pathologic case that probably has never happened
14530 in real life.) The previous approach was to use the version number
14531 from the record header as a substitute; but our protocol choice
14532 should not depend on that one because it is not authenticated
14533 by the Finished messages.
14537 * More robust randomness gathering functions for Windows.
14539 *Jeffrey Altman <jaltman@columbia.edu>*
14541 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14542 not set then we don't setup the error code for issuer check errors
14543 to avoid possibly overwriting other errors which the callback does
14544 handle. If an application does set the flag then we assume it knows
14545 what it is doing and can handle the new informational codes
14550 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14551 a general "ANY" type, as such it should be able to decode anything
14552 including tagged types. However it didn't check the class so it would
14553 wrongly interpret tagged types in the same way as their universal
14554 counterpart and unknown types were just rejected. Changed so that the
14555 tagged and unknown types are handled in the same way as a SEQUENCE:
14556 that is the encoding is stored intact. There is also a new type
14557 "V_ASN1_OTHER" which is used when the class is not universal, in this
14558 case we have no idea what the actual type is so we just lump them all
14563 * On VMS, stdout may very well lead to a file that is written to
14564 in a record-oriented fashion. That means that every write() will
14565 write a separate record, which will be read separately by the
14566 programs trying to read from it. This can be very confusing.
14568 The solution is to put a BIO filter in the way that will buffer
14569 text until a linefeed is reached, and then write everything a
14570 line at a time, so every record written will be an actual line,
14571 not chunks of lines and not (usually doesn't happen, but I've
14572 seen it once) several lines in one record. BIO_f_linebuffer() is
14575 Currently, it's a VMS-only method, because that's where it has
14576 been tested well enough.
14580 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14581 it can return incorrect results.
14582 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14583 but it was in 0.9.6-beta[12].)
14587 * Disable the check for content being present when verifying detached
14588 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14589 include zero length content when signing messages.
14593 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14594 BIO_ctrl (for BIO pairs).
14598 * Add DSO method for VMS.
14602 * Bug fix: Montgomery multiplication could produce results with the
14607 * Add RPM specification openssl.spec and modify it to build three
14608 packages. The default package contains applications, application
14609 documentation and run-time libraries. The devel package contains
14610 include files, static libraries and function documentation. The
14611 doc package contains the contents of the doc directory. The original
14612 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14616 * Add a large number of documentation files for many SSL routines.
14618 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14620 * Add a configuration entry for Sony News 4.
14622 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14624 * Don't set the two most significant bits to one when generating a
14625 random number < q in the DSA library.
14629 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14630 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14631 the underlying transport is blocking) if a handshake took place.
14632 (The default behaviour is needed by applications such as s_client
14633 and s_server that use select() to determine when to use SSL_read;
14634 but for applications that know in advance when to expect data, it
14635 just makes things more complicated.)
14639 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14644 * Add a few more EBCDIC conditionals that make `req` and `x509`
14645 work better on such systems.
14647 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14649 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14650 Update PKCS12_parse() so it copies the friendlyName and the
14651 keyid to the certificates aux info.
14655 * Fix bug in PKCS7_verify() which caused an infinite loop
14656 if there was more than one signature.
14658 *Sven Uszpelkat <su@celocom.de>*
14660 * Major change in util/mkdef.pl to include extra information
14661 about each symbol, as well as presenting variables as well
14662 as functions. This change means that there's n more need
14663 to rebuild the .num files when some algorithms are excluded.
14667 * Allow the verify time to be set by an application,
14668 rather than always using the current time.
14672 * Phase 2 verify code reorganisation. The certificate
14673 verify code now looks up an issuer certificate by a
14674 number of criteria: subject name, authority key id
14675 and key usage. It also verifies self signed certificates
14676 by the same criteria. The main comparison function is
14677 X509_check_issued() which performs these checks.
14679 Lot of changes were necessary in order to support this
14680 without completely rewriting the lookup code.
14682 Authority and subject key identifier are now cached.
14684 The LHASH 'certs' is X509_STORE has now been replaced
14685 by a STACK_OF(X509_OBJECT). This is mainly because an
14686 LHASH can't store or retrieve multiple objects with
14687 the same hash value.
14689 As a result various functions (which were all internal
14690 use only) have changed to handle the new X509_STORE
14691 structure. This will break anything that messed round
14692 with X509_STORE internally.
14694 The functions X509_STORE_add_cert() now checks for an
14695 exact match, rather than just subject name.
14697 The X509_STORE API doesn't directly support the retrieval
14698 of multiple certificates matching a given criteria, however
14699 this can be worked round by performing a lookup first
14700 (which will fill the cache with candidate certificates)
14701 and then examining the cache for matches. This is probably
14702 the best we can do without throwing out X509_LOOKUP
14703 entirely (maybe later...).
14705 The X509_VERIFY_CTX structure has been enhanced considerably.
14707 All certificate lookup operations now go via a get_issuer()
14708 callback. Although this currently uses an X509_STORE it
14709 can be replaced by custom lookups. This is a simple way
14710 to bypass the X509_STORE hackery necessary to make this
14711 work and makes it possible to use more efficient techniques
14712 in future. A very simple version which uses a simple
14713 STACK for its trusted certificate store is also provided
14714 using X509_STORE_CTX_trusted_stack().
14716 The verify_cb() and verify() callbacks now have equivalents
14717 in the X509_STORE_CTX structure.
14719 X509_STORE_CTX also has a 'flags' field which can be used
14720 to customise the verify behaviour.
14724 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14725 excludes S/MIME capabilities.
14729 * When a certificate request is read in keep a copy of the
14730 original encoding of the signed data and use it when outputting
14731 again. Signatures then use the original encoding rather than
14732 a decoded, encoded version which may cause problems if the
14733 request is improperly encoded.
14737 * For consistency with other BIO_puts implementations, call
14738 buffer_write(b, ...) directly in buffer_puts instead of calling
14741 In BIO_puts, increment b->num_write as in BIO_write.
14743 *Peter.Sylvester@EdelWeb.fr*
14745 * Fix BN_mul_word for the case where the word is 0. (We have to use
14746 BN_zero, we may not return a BIGNUM with an array consisting of
14747 words set to zero.)
14751 * Avoid calling abort() from within the library when problems are
14752 detected, except if preprocessor symbols have been defined
14753 (such as REF_CHECK, BN_DEBUG etc.).
14757 * New openssl application 'rsautl'. This utility can be
14758 used for low level RSA operations. DER public key
14759 BIO/fp routines also added.
14763 * New Configure entry and patches for compiling on QNX 4.
14765 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14767 * A demo state-machine implementation was sponsored by
14768 Nuron (<http://www.nuron.com/>) and is now available in
14769 demos/state_machine.
14773 * New options added to the 'dgst' utility for signature
14774 generation and verification.
14778 * Unrecognized PKCS#7 content types are now handled via a
14779 catch all ASN1_TYPE structure. This allows unsupported
14780 types to be stored as a "blob" and an application can
14781 encode and decode it manually.
14785 * Fix various signed/unsigned issues to make a_strex.c
14786 compile under VC++.
14788 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14790 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14791 length if passed a buffer. ASN1_INTEGER_to_BN failed
14792 if passed a NULL BN and its argument was negative.
14794 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14796 * Modification to PKCS#7 encoding routines to output definite
14797 length encoding. Since currently the whole structures are in
14798 memory there's not real point in using indefinite length
14799 constructed encoding. However if OpenSSL is compiled with
14800 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14804 * Added BIO_vprintf() and BIO_vsnprintf().
14808 * Added more prefixes to parse for in the strings written
14809 through a logging bio, to cover all the levels that are available
14810 through syslog. The prefixes are now:
14812 PANIC, EMERG, EMR => LOG_EMERG
14813 ALERT, ALR => LOG_ALERT
14814 CRIT, CRI => LOG_CRIT
14815 ERROR, ERR => LOG_ERR
14816 WARNING, WARN, WAR => LOG_WARNING
14817 NOTICE, NOTE, NOT => LOG_NOTICE
14818 INFO, INF => LOG_INFO
14819 DEBUG, DBG => LOG_DEBUG
14821 and as before, if none of those prefixes are present at the
14822 beginning of the string, LOG_ERR is chosen.
14824 On Win32, the `LOG_*` levels are mapped according to this:
14826 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14827 LOG_WARNING => EVENTLOG_WARNING_TYPE
14828 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14832 * Made it possible to reconfigure with just the configuration
14833 argument "reconf" or "reconfigure". The command line arguments
14834 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14835 and are retrieved from there when reconfiguring.
14841 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14843 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14847 * The obj_dat.pl script was messing up the sorting of object
14848 names. The reason was that it compared the quoted version
14849 of strings as a result "OCSP" > "OCSP Signing" because
14850 " > SPACE. Changed script to store unquoted versions of
14851 names and add quotes on output. It was also omitting some
14852 names from the lookup table if they were given a default
14853 value (that is if SN is missing it is given the same
14854 value as LN and vice versa), these are now added on the
14855 grounds that if an object has a name we should be able to
14856 look it up. Finally added warning output when duplicate
14857 short or long names are found.
14861 * Changes needed for Tandem NSK.
14863 *Scott Uroff <scott@xypro.com>*
14865 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14866 RSA_padding_check_SSLv23(), special padding was never detected
14867 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14868 version rollback attacks was not effective.
14870 In s23_clnt.c, don't use special rollback-attack detection padding
14871 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14872 client; similarly, in s23_srvr.c, don't do the rollback check if
14873 SSL 2.0 is the only protocol enabled in the server.
14877 * Make it possible to get hexdumps of unprintable data with 'openssl
14878 asn1parse'. By implication, the functions ASN1_parse_dump() and
14879 BIO_dump_indent() are added.
14883 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14884 these print out strings and name structures based on various
14885 flags including RFC2253 support and proper handling of
14886 multibyte characters. Added options to the 'x509' utility
14887 to allow the various flags to be set.
14891 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14892 Also change the functions X509_cmp_current_time() and
14893 X509_gmtime_adj() work with an ASN1_TIME structure,
14894 this will enable certificates using GeneralizedTime in validity
14895 dates to be checked.
14899 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14900 negative public key encodings) on by default,
14901 NO_NEG_PUBKEY_BUG can be set to disable it.
14905 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14906 content octets. An i2c_ASN1_OBJECT is unnecessary because
14907 the encoding can be trivially obtained from the structure.
14911 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14912 not read locks (`CRYPTO_r_[un]lock`).
14916 * A first attempt at creating official support for shared
14917 libraries through configuration. I've kept it so the
14918 default is static libraries only, and the OpenSSL programs
14919 are always statically linked for now, but there are
14920 preparations for dynamic linking in place.
14921 This has been tested on Linux and Tru64.
14925 * Randomness polling function for Win9x, as described in:
14926 Peter Gutmann, Software Generation of Practically Strong
14931 * Fix so PRNG is seeded in req if using an already existing
14936 * New options to smime application. -inform and -outform
14937 allow alternative formats for the S/MIME message including
14938 PEM and DER. The -content option allows the content to be
14939 specified separately. This should allow things like Netscape
14940 form signing output easier to verify.
14944 * Fix the ASN1 encoding of tags using the 'long form'.
14948 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14949 STRING types. These convert content octets to and from the
14950 underlying type. The actual tag and length octets are
14951 already assumed to have been read in and checked. These
14952 are needed because all other string types have virtually
14953 identical handling apart from the tag. By having versions
14954 of the ASN1 functions that just operate on content octets
14955 IMPLICIT tagging can be handled properly. It also allows
14956 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14957 and ASN1_INTEGER are identical apart from the tag.
14961 * Change the handling of OID objects as follows:
14963 - New object identifiers are inserted in objects.txt, following
14964 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14965 - objects.pl is used to process obj_mac.num and create a new
14967 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14970 This is currently kind of a hack, and the perl code in objects.pl
14971 isn't very elegant, but it works as I intended. The simplest way
14972 to check that it worked correctly is to look in obj_dat.h and
14973 check the array nid_objs and make sure the objects haven't moved
14974 around (this is important!). Additions are OK, as well as
14975 consistent name changes.
14979 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14983 * Addition of the command line parameter '-rand file' to 'openssl req'.
14984 The given file adds to whatever has already been seeded into the
14985 random pool through the RANDFILE configuration file option or
14986 environment variable, or the default random state file.
14990 * mkstack.pl now sorts each macro group into lexical order.
14991 Previously the output order depended on the order the files
14992 appeared in the directory, resulting in needless rewriting
14997 * Patches to make OpenSSL compile under Win32 again. Mostly
14998 work arounds for the VC++ problem that it treats func() as
14999 func(void). Also stripped out the parts of mkdef.pl that
15000 added extra typesafe functions: these no longer exist.
15004 * Reorganisation of the stack code. The macros are now all
15005 collected in safestack.h . Each macro is defined in terms of
15006 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15007 DEBUG_SAFESTACK is now handled in terms of function casts,
15008 this has the advantage of retaining type safety without the
15009 use of additional functions. If DEBUG_SAFESTACK is not defined
15010 then the non typesafe macros are used instead. Also modified the
15011 mkstack.pl script to handle the new form. Needs testing to see
15012 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15013 the default if no major problems. Similar behaviour for ASN1_SET_OF
15014 and PKCS12_STACK_OF.
15018 * When some versions of IIS use the 'NET' form of private key the
15019 key derivation algorithm is different. Normally MD5(password) is
15020 used as a 128 bit RC4 key. In the modified case
15021 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15022 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15023 as the old Netscape_RSA functions except they have an additional
15024 'sgckey' parameter which uses the modified algorithm. Also added
15025 an -sgckey command line option to the rsa utility. Thanks to
15026 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15027 algorithm to openssl-dev.
15031 * The evp_local.h macros were using 'c.##kname' which resulted in
15032 invalid expansion on some systems (SCO 5.0.5 for example).
15033 Corrected to 'c.kname'.
15035 *Phillip Porch <root@theporch.com>*
15037 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15038 a STACK of email addresses from a certificate or request, these look
15039 in the subject name and the subject alternative name extensions and
15040 omit any duplicate addresses.
15044 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15045 This makes DSA verification about 2 % faster.
15049 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15050 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15051 plus overhead for 1024 bit moduli).
15052 This makes exponentiations about 0.5 % faster for 1024 bit
15053 exponents (as measured by "openssl speed rsa2048").
15057 * Rename memory handling macros to avoid conflicts with other
15059 Malloc => OPENSSL_malloc
15060 Malloc_locked => OPENSSL_malloc_locked
15061 Realloc => OPENSSL_realloc
15062 Free => OPENSSL_free
15066 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15067 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15071 * CygWin32 support.
15073 *John Jarvie <jjarvie@newsguy.com>*
15075 * The type-safe stack code has been rejigged. It is now only compiled
15076 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15077 by default all type-specific stack functions are "#define"d back to
15078 standard stack functions. This results in more streamlined output
15079 but retains the type-safety checking possibilities of the original
15084 * The STACK code has been cleaned up, and certain type declarations
15085 that didn't make a lot of sense have been brought in line. This has
15086 also involved a cleanup of sorts in safestack.h to more correctly
15087 map type-safe stack functions onto their plain stack counterparts.
15088 This work has also resulted in a variety of "const"ifications of
15089 lots of the code, especially `_cmp` operations which should normally
15090 be prototyped with "const" parameters anyway.
15094 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15095 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15096 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15097 where all of 'md' is used each time the PRNG is used, but 'state'
15098 is used only indexed by a cyclic counter. As entropy may not be
15099 well distributed from the beginning, 'md' is important as a
15100 chaining variable. However, the output function chains only half
15101 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15102 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15103 in all of 'state' being rewritten, with the new values depending
15104 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15108 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15109 the handshake is continued after ssl_verify_cert_chain();
15110 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15111 can lead to 'unexplainable' connection aborts later.
15113 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15115 * Major EVP API cipher revision.
15116 Add hooks for extra EVP features. This allows various cipher
15117 parameters to be set in the EVP interface. Support added for variable
15118 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15119 setting of RC2 and RC5 parameters.
15121 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15124 Remove lots of duplicated code from the EVP library. For example *every*
15125 cipher init() function handles the 'iv' in the same way according to the
15126 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15127 for CFB and OFB modes they zero ctx->num.
15129 New functionality allows removal of S/MIME code RC2 hack.
15131 Most of the routines have the same form and so can be declared in terms
15134 By shifting this to the top level EVP_CipherInit() it can be removed from
15135 all individual ciphers. If the cipher wants to handle IVs or keys
15136 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15139 Change lots of functions like EVP_EncryptUpdate() to now return a
15140 value: although software versions of the algorithms cannot fail
15141 any installed hardware versions can.
15145 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15146 this option is set, tolerate broken clients that send the negotiated
15147 protocol version number instead of the requested protocol version
15152 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15153 i.e. non-zero for export ciphersuites, zero otherwise.
15154 Previous versions had this flag inverted, inconsistent with
15155 rsa_tmp_cb (..._TMP_RSA_CB).
15157 *Bodo Moeller; problem reported by Amit Chopra*
15159 * Add missing DSA library text string. Work around for some IIS
15160 key files with invalid SEQUENCE encoding.
15164 * Add a document (doc/standards.txt) that list all kinds of standards
15165 and so on that are implemented in OpenSSL.
15169 * Enhance c_rehash script. Old version would mishandle certificates
15170 with the same subject name hash and wouldn't handle CRLs at all.
15171 Added -fingerprint option to crl utility, to support new c_rehash
15176 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15180 * Fix for SSL server purpose checking. Server checking was
15181 rejecting certificates which had extended key usage present
15182 but no ssl client purpose.
15184 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15186 * Make PKCS#12 code work with no password. The PKCS#12 spec
15187 is a little unclear about how a blank password is handled.
15188 Since the password in encoded as a BMPString with terminating
15189 double NULL a zero length password would end up as just the
15190 double NULL. However no password at all is different and is
15191 handled differently in the PKCS#12 key generation code. NS
15192 treats a blank password as zero length. MSIE treats it as no
15193 password on export: but it will try both on import. We now do
15194 the same: PKCS12_parse() tries zero length and no password if
15195 the password is set to "" or NULL (NULL is now a valid password:
15196 it wasn't before) as does the pkcs12 application.
15200 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15201 perror when PEM_read_bio_X509_REQ fails, the error message must
15202 be obtained from the error queue.
15206 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15207 it in ERR_remove_state if appropriate, and change ERR_get_state
15208 accordingly to avoid race conditions (this is necessary because
15209 thread_hash is no longer constant once set).
15213 * Bugfix for linux-elf makefile.one.
15217 * RSA_get_default_method() will now cause a default
15218 RSA_METHOD to be chosen if one doesn't exist already.
15219 Previously this was only set during a call to RSA_new()
15220 or RSA_new_method(NULL) meaning it was possible for
15221 RSA_get_default_method() to return NULL.
15225 * Added native name translation to the existing DSO code
15226 that will convert (if the flag to do so is set) filenames
15227 that are sufficiently small and have no path information
15228 into a canonical native form. Eg. "blah" converted to
15229 "libblah.so" or "blah.dll" etc.
15233 * New function ERR_error_string_n(e, buf, len) which is like
15234 ERR_error_string(e, buf), but writes at most 'len' bytes
15235 including the 0 terminator. For ERR_error_string_n, 'buf'
15238 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15240 * CONF library reworked to become more general. A new CONF
15241 configuration file reader "class" is implemented as well as a
15242 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15243 old `CONF_*` functions are still there, but are reimplemented to
15244 work in terms of the new functions. Also, a set of functions
15245 to handle the internal storage of the configuration data is
15246 provided to make it easier to write new configuration file
15247 reader "classes" (I can definitely see something reading a
15248 configuration file in XML format, for example), called `_CONF_*`,
15249 or "the configuration storage API"...
15251 The new configuration file reading functions are:
15253 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15254 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15256 NCONF_default, NCONF_WIN32
15258 NCONF_dump_fp, NCONF_dump_bio
15260 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15261 NCONF_new creates a new CONF object. This works in the same way
15262 as other interfaces in OpenSSL, like the BIO interface.
15263 `NCONF_dump_*` dump the internal storage of the configuration file,
15264 which is useful for debugging. All other functions take the same
15265 arguments as the old `CONF_*` functions with the exception of the
15266 first that must be a `CONF *` instead of a `LHASH *`.
15268 To make it easier to use the new classes with the old `CONF_*` functions,
15269 the function CONF_set_default_method is provided.
15273 * Add '-tls1' option to 'openssl ciphers', which was already
15274 mentioned in the documentation but had not been implemented.
15275 (This option is not yet really useful because even the additional
15276 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15280 * Initial DSO code added into libcrypto for letting OpenSSL (and
15281 OpenSSL-based applications) load shared libraries and bind to
15282 them in a portable way.
15284 *Geoff Thorpe, with contributions from Richard Levitte*
15286 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15288 * Make sure _lrotl and _lrotr are only used with MSVC.
15290 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15291 (the default implementation of RAND_status).
15293 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15294 to '-clrext' (= clear extensions), as intended and documented.
15295 *Bodo Moeller; inconsistency pointed out by Michael Attili
15296 <attili@amaxo.com>*
15298 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15299 was larger than the MD block size.
15301 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15303 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15304 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15305 using the passed key: if the passed key was a private key the result
15306 of X509_print(), for example, would be to print out all the private key
15311 * des_quad_cksum() byte order bug fix.
15312 *Ulf Möller, using the problem description in krb4-0.9.7, where
15313 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15315 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15318 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15320 * For easily testing in shell scripts whether some command
15321 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15322 returns with exit code 0 iff no command of the given name is available.
15323 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15324 the output goes to stdout and nothing is printed to stderr.
15325 Additional arguments are always ignored.
15327 Since for each cipher there is a command of the same name,
15328 the 'no-cipher' compilation switches can be tested this way.
15330 ('openssl no-XXX' is not able to detect pseudo-commands such
15331 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15335 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15339 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15340 is set; it will be thrown away anyway because each handshake creates
15342 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15343 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15344 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15345 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15349 * New s_client option -ign_eof: EOF at stdin is ignored, and
15350 'Q' and 'R' lose their special meanings (quit/renegotiate).
15351 This is part of what -quiet does; unlike -quiet, -ign_eof
15352 does not suppress any output.
15356 * Add compatibility options to the purpose and trust code. The
15357 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15358 accepts a certificate or CA, this was the previous behaviour,
15359 with all the associated security issues.
15361 X509_TRUST_COMPAT is the old trust behaviour: only and
15362 automatically trust self signed roots in certificate store. A
15363 new trust setting X509_TRUST_DEFAULT is used to specify that
15364 a purpose has no associated trust setting and it should instead
15365 use the value in the default purpose.
15369 * Fix the PKCS#8 DSA private key code so it decodes keys again
15370 and fix a memory leak.
15374 * In util/mkerr.pl (which implements 'make errors'), preserve
15375 reason strings from the previous version of the .c file, as
15376 the default to have only downcase letters (and digits) in
15377 automatically generated reasons codes is not always appropriate.
15381 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15382 using strerror. Previously, ERR_reason_error_string() returned
15383 library names as reason strings for SYSerr; but SYSerr is a special
15384 case where small numbers are errno values, not library numbers.
15388 * Add '-dsaparam' option to 'openssl dhparam' application. This
15389 converts DSA parameters into DH parameters. (When creating parameters,
15390 DSA_generate_parameters is used.)
15394 * Include 'length' (recommended exponent length) in C code generated
15395 by 'openssl dhparam -C'.
15399 * The second argument to set_label in perlasm was already being used
15400 so couldn't be used as a "file scope" flag. Moved to third argument
15405 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15406 instead of RAND_bytes for encryption IVs and salts.
15410 * Include RAND_status() into RAND_METHOD instead of implementing
15411 it only for md_rand.c Otherwise replacing the PRNG by calling
15412 RAND_set_rand_method would be impossible.
15416 * Don't let DSA_generate_key() enter an infinite loop if the random
15417 number generation fails.
15421 * New 'rand' application for creating pseudo-random output.
15425 * Added configuration support for Linux/IA64
15427 *Rolf Haberrecker <rolf@suse.de>*
15429 * Assembler module support for Mingw32.
15433 * Shared library support for HPUX (in shlib/).
15435 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15437 * Shared library support for Solaris gcc.
15439 *Lutz Behnke <behnke@trustcenter.de>*
15441 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15443 * PKCS7_encrypt() was adding text MIME headers twice because they
15444 were added manually and by SMIME_crlf_copy().
15448 * In bntest.c don't call BN_rand with zero bits argument.
15450 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15452 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15453 case was implemented. This caused BN_div_recp() to fail occasionally.
15457 * Add an optional second argument to the set_label() in the perl
15458 assembly language builder. If this argument exists and is set
15459 to 1 it signals that the assembler should use a symbol whose
15460 scope is the entire file, not just the current function. This
15461 is needed with MASM which uses the format label:: for this scope.
15463 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15465 * Change the ASN1 types so they are typedefs by default. Before
15466 almost all types were #define'd to ASN1_STRING which was causing
15467 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15472 * Change names of new functions to the new get1/get0 naming
15473 convention: After 'get1', the caller owns a reference count
15474 and has to call `..._free`; 'get0' returns a pointer to some
15475 data structure without incrementing reference counters.
15476 (Some of the existing 'get' functions increment a reference
15477 counter, some don't.)
15478 Similarly, 'set1' and 'add1' functions increase reference
15479 counters or duplicate objects.
15483 * Allow for the possibility of temp RSA key generation failure:
15484 the code used to assume it always worked and crashed on failure.
15488 * Fix potential buffer overrun problem in BIO_printf().
15489 *Ulf Möller, using public domain code by Patrick Powell; problem
15490 pointed out by David Sacerdote <das33@cornell.edu>*
15492 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15493 RAND_egd() and RAND_status(). In the command line application,
15494 the EGD socket can be specified like a seed file using RANDFILE
15499 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15500 Some CAs (e.g. Verisign) distribute certificates in this form.
15504 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15505 list to exclude them. This means that no special compilation option
15506 is needed to use anonymous DH: it just needs to be included in the
15511 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15512 EVP_MD_type. The old functionality is available in a new macro called
15513 EVP_MD_md(). Change code that uses it and update docs.
15517 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15518 where the `void *` argument is replaced by a function pointer argument.
15519 Previously `void *` was abused to point to functions, which works on
15520 many platforms, but is not correct. As these functions are usually
15521 called by macros defined in OpenSSL header files, most source code
15522 should work without changes.
15526 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15527 sections with information on -D... compiler switches used for
15528 compiling the library so that applications can see them. To enable
15529 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15530 must be defined. E.g.,
15531 #define OPENSSL_ALGORITHM_DEFINES
15532 #include <openssl/opensslconf.h>
15533 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15535 *Richard Levitte, Ulf and Bodo Möller*
15537 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15542 * Change the 'other' type in certificate aux info to a STACK_OF
15543 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15544 the required ASN1 format: arbitrary types determined by an OID.
15548 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15549 argument to 'req'. This is not because the function is newer or
15550 better than others it just uses the work 'NEW' in the certificate
15551 request header lines. Some software needs this.
15555 * Reorganise password command line arguments: now passwords can be
15556 obtained from various sources. Delete the PEM_cb function and make
15557 it the default behaviour: i.e. if the callback is NULL and the
15558 usrdata argument is not NULL interpret it as a null terminated pass
15559 phrase. If usrdata and the callback are NULL then the pass phrase
15560 is prompted for as usual.
15564 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15565 the support is automatically enabled. The resulting binaries will
15566 autodetect the card and use it if present.
15568 *Ben Laurie and Compaq Inc.*
15570 * Work around for Netscape hang bug. This sends certificate request
15571 and server done in one record. Since this is perfectly legal in the
15572 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15573 the bugs/SSLv3 entry for more info.
15577 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15581 * Add -rand argument to smime and pkcs12 applications and read/write
15586 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15590 * Add command line password options to the remaining applications.
15594 * Bug fix for BN_div_recp() for numerators with an even number of
15599 * More tests in bntest.c, and changed test_bn output.
15603 * ./config recognizes MacOS X now.
15607 * Bug fix for BN_div() when the first words of num and divisor are
15608 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15612 * Add support for various broken PKCS#8 formats, and command line
15613 options to produce them.
15617 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15618 get temporary BIGNUMs from a BN_CTX.
15622 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15627 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15628 include a #define from the old name to the new. The original intent
15629 was that statically linked binaries could for example just call
15630 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15631 link with digests. This never worked because SSLeay_add_all_digests()
15632 and SSLeay_add_all_ciphers() were in the same source file so calling
15633 one would link with the other. They are now in separate source files.
15637 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15641 * Use a less unusual form of the Miller-Rabin primality test (it used
15642 a binary algorithm for exponentiation integrated into the Miller-Rabin
15643 loop, our standard modexp algorithms are faster).
15647 * Support for the EBCDIC character set completed.
15649 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15651 * Source code cleanups: use const where appropriate, eliminate casts,
15652 use `void *` instead of `char *` in lhash.
15656 * Bugfix: ssl3_send_server_key_exchange was not restartable
15657 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15658 this the server could overwrite ephemeral keys that the client
15663 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15664 using 50 iterations of the Rabin-Miller test.
15666 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15667 iterations of the Rabin-Miller test as required by the appendix
15668 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15669 As BN_is_prime_fasttest includes trial division, DSA parameter
15670 generation becomes much faster.
15672 This implies a change for the callback functions in DSA_is_prime
15673 and DSA_generate_parameters: The callback function is called once
15674 for each positive witness in the Rabin-Miller test, not just
15675 occasionally in the inner loop; and the parameters to the
15676 callback function now provide an iteration count for the outer
15677 loop rather than for the current invocation of the inner loop.
15678 DSA_generate_parameters additionally can call the callback
15679 function with an 'iteration count' of -1, meaning that a
15680 candidate has passed the trial division test (when q is generated
15681 from an application-provided seed, trial division is skipped).
15685 * New function BN_is_prime_fasttest that optionally does trial
15686 division before starting the Rabin-Miller test and has
15687 an additional BN_CTX * argument (whereas BN_is_prime always
15688 has to allocate at least one BN_CTX).
15689 'callback(1, -1, cb_arg)' is called when a number has passed the
15690 trial division stage.
15694 * Fix for bug in CRL encoding. The validity dates weren't being handled
15699 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15703 * New function BN_pseudo_rand().
15707 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15708 bignum version of BN_from_montgomery() with the working code from
15709 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15714 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15715 made it impossible to use the same SSL_SESSION data structure in
15716 SSL2 clients in multiple threads.
15720 * The return value of RAND_load_file() no longer counts bytes obtained
15721 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15722 to seed the PRNG (previously an explicit byte count was required).
15724 *Ulf Möller, Bodo Möller*
15726 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15727 used `char *` instead of `void *` and had casts all over the place.
15731 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15735 * Retain source code compatibility for BN_prime_checks macro:
15736 BN_is_prime(..., BN_prime_checks, ...) now uses
15737 BN_prime_checks_for_size to determine the appropriate number of
15738 Rabin-Miller iterations.
15742 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15743 DH_CHECK_P_NOT_SAFE_PRIME.
15744 (Check if this is true? OpenPGP calls them "strong".)
15748 * Merge the functionality of "dh" and "gendh" programs into a new program
15749 "dhparam". The old programs are retained for now but will handle DH keys
15750 (instead of parameters) in future.
15754 * Make the ciphers, s_server and s_client programs check the return values
15755 when a new cipher list is set.
15759 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15760 ciphers. Before when the 56bit ciphers were enabled the sorting was
15763 The syntax for the cipher sorting has been extended to support sorting by
15764 cipher-strength (using the strength_bits hard coded in the tables).
15765 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15767 Fix a bug in the cipher-command parser: when supplying a cipher command
15768 string with an "undefined" symbol (neither command nor alphanumeric
15769 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15770 an error is flagged.
15772 Due to the strength-sorting extension, the code of the
15773 ssl_create_cipher_list() function was completely rearranged. I hope that
15774 the readability was also increased :-)
15776 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15778 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15779 for the first serial number and places 2 in the serial number file. This
15780 avoids problems when the root CA is created with serial number zero and
15781 the first user certificate has the same issuer name and serial number
15786 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15787 the new code. Add documentation for this stuff.
15791 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15792 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15793 structures and behave in an analogous way to the X509v3 functions:
15794 they shouldn't be called directly but wrapper functions should be used
15797 So we also now have some wrapper functions that call the X509at functions
15798 when passed certificate requests. (TO DO: similar things can be done with
15799 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15800 things. Some of these need some d2i or i2d and print functionality
15801 because they handle more complex structures.)
15805 * Add missing #ifndefs that caused missing symbols when building libssl
15806 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15807 NO_RSA in `ssl/s2*.c`.
15809 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15811 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15812 has a return value which indicates the quality of the random data
15813 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15814 error queue. New function RAND_pseudo_bytes() generates output that is
15815 guaranteed to be unique but not unpredictable. RAND_add is like
15816 RAND_seed, but takes an extra argument for an entropy estimate
15817 (RAND_seed always assumes full entropy).
15821 * Do more iterations of Rabin-Miller probable prime test (specifically,
15822 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15823 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15824 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15825 false-positive rate of at most 2^-80 for random input.
15829 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15833 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15834 in the 0.9.5 release), this returns the chain
15835 from an X509_CTX structure with a dup of the stack and all
15836 the X509 reference counts upped: so the stack will exist
15837 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15840 Also make SSL_SESSION_print() print out the verify return
15845 * Add manpage for the pkcs12 command. Also change the default
15846 behaviour so MAC iteration counts are used unless the new
15847 -nomaciter option is used. This improves file security and
15848 only older versions of MSIE (4.0 for example) need it.
15852 * Honor the no-xxx Configure options when creating .DEF files.
15856 * Add PKCS#10 attributes to field table: challengePassword,
15857 unstructuredName and unstructuredAddress. These are taken from
15858 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15859 international characters are used.
15861 More changes to X509_ATTRIBUTE code: allow the setting of types
15862 based on strings. Remove the 'loc' parameter when adding
15863 attributes because these will be a SET OF encoding which is sorted
15868 * Initial changes to the 'req' utility to allow request generation
15869 automation. This will allow an application to just generate a template
15870 file containing all the field values and have req construct the
15873 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15874 used all over the place including certificate requests and PKCS#7
15875 structures. They are currently handled manually where necessary with
15876 some primitive wrappers for PKCS#7. The new functions behave in a
15877 manner analogous to the X509 extension functions: they allow
15878 attributes to be looked up by NID and added.
15880 Later something similar to the X509V3 code would be desirable to
15881 automatically handle the encoding, decoding and printing of the
15882 more complex types. The string types like challengePassword can
15883 be handled by the string table functions.
15885 Also modified the multi byte string table handling. Now there is
15886 a 'global mask' which masks out certain types. The table itself
15887 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15888 is useful when for example there is only one permissible type
15889 (as in countryName) and using the mask might result in no valid
15894 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15895 SSL_get_peer_finished to allow applications to obtain the latest
15896 Finished messages sent to the peer or expected from the peer,
15897 respectively. (SSL_get_peer_finished is usually the Finished message
15898 actually received from the peer, otherwise the protocol will be aborted.)
15900 As the Finished message are message digests of the complete handshake
15901 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15902 be used for external authentication procedures when the authentication
15903 provided by SSL/TLS is not desired or is not enough.
15907 * Enhanced support for Alpha Linux is added. Now ./config checks if
15908 the host supports BWX extension and if Compaq C is present on the
15909 $PATH. Just exploiting of the BWX extension results in 20-30%
15910 performance kick for some algorithms, e.g. DES and RC4 to mention
15911 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15916 * Add support for MS "fast SGC". This is arguably a violation of the
15917 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15918 weak crypto and after checking the certificate is SGC a second one
15919 with strong crypto. MS SGC stops the first handshake after receiving
15920 the server certificate message and sends a second client hello. Since
15921 a server will typically do all the time consuming operations before
15922 expecting any further messages from the client (server key exchange
15923 is the most expensive) there is little difference between the two.
15925 To get OpenSSL to support MS SGC we have to permit a second client
15926 hello message after we have sent server done. In addition we have to
15927 reset the MAC if we do get this second client hello.
15931 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15932 if a DER encoded private key is RSA or DSA traditional format. Changed
15933 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15934 format DER encoded private key. Newer code should use PKCS#8 format which
15935 has the key type encoded in the ASN1 structure. Added DER private key
15936 support to pkcs8 application.
15940 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15941 ciphersuites has been selected (as required by the SSL 3/TLS 1
15942 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15943 is set, we interpret this as a request to violate the specification
15944 (the worst that can happen is a handshake failure, and 'correct'
15945 behaviour would result in a handshake failure anyway).
15949 * In SSL_CTX_add_session, take into account that there might be multiple
15950 SSL_SESSION structures with the same session ID (e.g. when two threads
15951 concurrently obtain them from an external cache).
15952 The internal cache can handle only one SSL_SESSION with a given ID,
15953 so if there's a conflict, we now throw out the old one to achieve
15958 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15959 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15960 some routines that use cipher OIDs: some ciphers do not have OIDs
15961 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15966 * Simplify the trust setting structure and code. Now we just have
15967 two sequences of OIDs for trusted and rejected settings. These will
15968 typically have values the same as the extended key usage extension
15969 and any application specific purposes.
15971 The trust checking code now has a default behaviour: it will just
15972 check for an object with the same NID as the passed id. Functions can
15973 be provided to override either the default behaviour or the behaviour
15974 for a given id. SSL client, server and email already have functions
15975 in place for compatibility: they check the NID and also return "trusted"
15976 if the certificate is self signed.
15980 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15981 traditional format into an EVP_PKEY structure.
15985 * Add a password callback function PEM_cb() which either prompts for
15986 a password if usr_data is NULL or otherwise assumes it is a null
15987 terminated password. Allow passwords to be passed on command line
15988 environment or config files in a few more utilities.
15992 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15993 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15994 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15995 Update documentation.
15999 * Support for ASN1 "NULL" type. This could be handled before by using
16000 ASN1_TYPE but there wasn't any function that would try to read a NULL
16001 and produce an error if it couldn't. For compatibility we also have
16002 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16003 don't allocate anything because they don't need to.
16007 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16010 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16012 * Rebuild of the memory allocation routines used by OpenSSL code and
16013 possibly others as well. The purpose is to make an interface that
16014 provide hooks so anyone can build a separate set of allocation and
16015 deallocation routines to be used by OpenSSL, for example memory
16016 pool implementations, or something else, which was previously hard
16017 since Malloc(), Realloc() and Free() were defined as macros having
16018 the values malloc, realloc and free, respectively (except for Win32
16019 compilations). The same is provided for memory debugging code.
16020 OpenSSL already comes with functionality to find memory leaks, but
16021 this gives people a chance to debug other memory problems.
16023 With these changes, a new set of functions and macros have appeared:
16025 CRYPTO_set_mem_debug_functions() [F]
16026 CRYPTO_get_mem_debug_functions() [F]
16027 CRYPTO_dbg_set_options() [F]
16028 CRYPTO_dbg_get_options() [F]
16029 CRYPTO_malloc_debug_init() [M]
16031 The memory debug functions are NULL by default, unless the library
16032 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16033 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16034 gives the standard debugging functions that come with OpenSSL) or
16035 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16036 provided by the library user) must be used. When the standard
16037 debugging functions are used, CRYPTO_dbg_set_options can be used to
16038 request additional information:
16039 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16040 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16042 Also, things like CRYPTO_set_mem_functions will always give the
16043 expected result (the new set of functions is used for allocation
16044 and deallocation) at all times, regardless of platform and compiler
16047 To finish it up, some functions that were never use in any other
16048 way than through macros have a new API and new semantic:
16050 CRYPTO_dbg_malloc()
16051 CRYPTO_dbg_realloc()
16054 All macros of value have retained their old syntax.
16056 *Richard Levitte and Bodo Moeller*
16058 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16059 ordering of SMIMECapabilities wasn't in "strength order" and there
16060 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16065 * Some ASN1 types with illegal zero length encoding (INTEGER,
16066 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16068 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16070 * Merge in my S/MIME library for OpenSSL. This provides a simple
16071 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16072 functionality to handle multipart/signed properly) and a utility
16073 called 'smime' to call all this stuff. This is based on code I
16074 originally wrote for Celo who have kindly allowed it to be
16075 included in OpenSSL.
16079 * Add variants des_set_key_checked and des_set_key_unchecked of
16080 des_set_key (aka des_key_sched). Global variable des_check_key
16081 decides which of these is called by des_set_key; this way
16082 des_check_key behaves as it always did, but applications and
16083 the library itself, which was buggy for des_check_key == 1,
16084 have a cleaner way to pick the version they need.
16088 * New function PKCS12_newpass() which changes the password of a
16093 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16094 dynamic mix. In both cases the ids can be used as an index into the
16095 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16096 functions so they accept a list of the field values and the
16097 application doesn't need to directly manipulate the X509_TRUST
16102 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16107 * Modify the way the V3 extension code looks up extensions. This now
16108 works in a similar way to the object code: we have some "standard"
16109 extensions in a static table which is searched with OBJ_bsearch()
16110 and the application can add dynamic ones if needed. The file
16111 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16112 updated whenever a new extension is added to the core code and kept
16113 in ext_nid order. There is a simple program 'tabtest.c' which checks
16114 this. New extensions are not added too often so this file can readily
16115 be maintained manually.
16117 There are two big advantages in doing things this way. The extensions
16118 can be looked up immediately and no longer need to be "added" using
16119 X509V3_add_standard_extensions(): this function now does nothing.
16120 Side note: I get *lots* of email saying the extension code doesn't
16121 work because people forget to call this function.
16122 Also no dynamic allocation is done unless new extensions are added:
16123 so if we don't add custom extensions there is no need to call
16124 X509V3_EXT_cleanup().
16128 * Modify enc utility's salting as follows: make salting the default. Add a
16129 magic header, so unsalted files fail gracefully instead of just decrypting
16130 to garbage. This is because not salting is a big security hole, so people
16131 should be discouraged from doing it.
16135 * Fixes and enhancements to the 'x509' utility. It allowed a message
16136 digest to be passed on the command line but it only used this
16137 parameter when signing a certificate. Modified so all relevant
16138 operations are affected by the digest parameter including the
16139 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16140 DSA key was used because it didn't fix the digest.
16144 * Initial certificate chain verify code. Currently tests the untrusted
16145 certificates for consistency with the verify purpose (which is set
16146 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16148 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16149 this is because it will reject chains with invalid extensions whereas
16150 every previous version of OpenSSL and SSLeay made no checks at all.
16152 Trust code: checks the root CA for the relevant trust settings. Trust
16153 settings have an initial value consistent with the verify purpose: e.g.
16154 if the verify purpose is for SSL client use it expects the CA to be
16155 trusted for SSL client use. However the default value can be changed to
16156 permit custom trust settings: one example of this would be to only trust
16157 certificates from a specific "secure" set of CAs.
16159 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16160 which should be used for version portability: especially since the
16161 verify structure is likely to change more often now.
16163 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16164 to set them. If not set then assume SSL clients will verify SSL servers
16167 Two new options to the verify program: -untrusted allows a set of
16168 untrusted certificates to be passed in and -purpose which sets the
16169 intended purpose of the certificate. If a purpose is set then the
16170 new chain verify code is used to check extension consistency.
16174 * Support for the authority information access extension.
16178 * Modify RSA and DSA PEM read routines to transparently handle
16179 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16180 public keys in a format compatible with certificate
16181 SubjectPublicKeyInfo structures. Unfortunately there were already
16182 functions called *_PublicKey_* which used various odd formats so
16183 these are retained for compatibility: however the DSA variants were
16184 never in a public release so they have been deleted. Changed dsa/rsa
16185 utilities to handle the new format: note no releases ever handled public
16186 keys so we should be OK.
16188 The primary motivation for this change is to avoid the same fiasco
16189 that dogs private keys: there are several incompatible private key
16190 formats some of which are standard and some OpenSSL specific and
16191 require various evil hacks to allow partial transparent handling and
16192 even then it doesn't work with DER formats. Given the option anything
16193 other than PKCS#8 should be dumped: but the other formats have to
16194 stay in the name of compatibility.
16196 With public keys and the benefit of hindsight one standard format
16197 is used which works with EVP_PKEY, RSA or DSA structures: though
16198 it clearly returns an error if you try to read the wrong kind of key.
16200 Added a -pubkey option to the 'x509' utility to output the public key.
16201 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16202 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16203 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16204 that do the same as the `EVP_PKEY_assign_*()` except they up the
16205 reference count of the added key (they don't "swallow" the
16210 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16211 CRLs would fail if the file contained no certificates or no CRLs:
16212 added a new function to read in both types and return the number
16213 read: this means that if none are read it will be an error. The
16214 DER versions of the certificate and CRL reader would always fail
16215 because it isn't possible to mix certificates and CRLs in DER format
16216 without choking one or the other routine. Changed this to just read
16217 a certificate: this is the best we can do. Also modified the code
16218 in apps/verify.c to take notice of return codes: it was previously
16219 attempting to read in certificates from NULL pointers and ignoring
16220 any errors: this is one reason why the cert and CRL reader seemed
16221 to work. It doesn't check return codes from the default certificate
16222 routines: these may well fail if the certificates aren't installed.
16226 * Code to support otherName option in GeneralName.
16230 * First update to verify code. Change the verify utility
16231 so it warns if it is passed a self signed certificate:
16232 for consistency with the normal behaviour. X509_verify
16233 has been modified to it will now verify a self signed
16234 certificate if *exactly* the same certificate appears
16235 in the store: it was previously impossible to trust a
16236 single self signed certificate. This means that:
16237 openssl verify ss.pem
16238 now gives a warning about a self signed certificate but
16239 openssl verify -CAfile ss.pem ss.pem
16244 * For servers, store verify_result in SSL_SESSION data structure
16245 (and add it to external session representation).
16246 This is needed when client certificate verifications fails,
16247 but an application-provided verification callback (set by
16248 SSL_CTX_set_cert_verify_callback) allows accepting the session
16249 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16250 but returns 1): When the session is reused, we have to set
16251 ssl->verify_result to the appropriate error code to avoid
16254 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16256 * Fix a bug in the new PKCS#7 code: it didn't consider the
16257 case in PKCS7_dataInit() where the signed PKCS7 structure
16258 didn't contain any existing data because it was being created.
16260 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16262 * Add a salt to the key derivation routines in enc.c. This
16263 forms the first 8 bytes of the encrypted file. Also add a
16264 -S option to allow a salt to be input on the command line.
16268 * New function X509_cmp(). Oddly enough there wasn't a function
16269 to compare two certificates. We do this by working out the SHA1
16270 hash and comparing that. X509_cmp() will be needed by the trust
16275 * SSL_get1_session() is like SSL_get_session(), but increments
16276 the reference count in the SSL_SESSION returned.
16278 *Geoff Thorpe <geoff@eu.c2.net>*
16280 * Fix for 'req': it was adding a null to request attributes.
16281 Also change the X509_LOOKUP and X509_INFO code to handle
16282 certificate auxiliary information.
16286 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16291 * Add the possibility to add extra information to the memory leak
16292 detecting output, to form tracebacks, showing from where each
16293 allocation was originated: CRYPTO_push_info("constant string") adds
16294 the string plus current file name and line number to a per-thread
16295 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16296 is like calling CYRPTO_pop_info() until the stack is empty.
16297 Also updated memory leak detection code to be multi-thread-safe.
16301 * Add options -text and -noout to pkcs7 utility and delete the
16302 encryption options which never did anything. Update docs.
16306 * Add options to some of the utilities to allow the pass phrase
16307 to be included on either the command line (not recommended on
16308 OSes like Unix) or read from the environment. Update the
16309 manpages and fix a few bugs.
16313 * Add a few manpages for some of the openssl commands.
16317 * Fix the -revoke option in ca. It was freeing up memory twice,
16318 leaking and not finding already revoked certificates.
16322 * Extensive changes to support certificate auxiliary information.
16323 This involves the use of X509_CERT_AUX structure and X509_AUX
16324 functions. An X509_AUX function such as PEM_read_X509_AUX()
16325 can still read in a certificate file in the usual way but it
16326 will also read in any additional "auxiliary information". By
16327 doing things this way a fair degree of compatibility can be
16328 retained: existing certificates can have this information added
16329 using the new 'x509' options.
16331 Current auxiliary information includes an "alias" and some trust
16332 settings. The trust settings will ultimately be used in enhanced
16333 certificate chain verification routines: currently a certificate
16334 can only be trusted if it is self signed and then it is trusted
16339 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16340 The problem was that one of the replacement routines had not been working
16341 since SSLeay releases. For now the offending routine has been replaced
16342 with non-optimised assembler. Even so, this now gives around 95%
16343 performance improvement for 1024 bit RSA signs.
16347 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16348 handling. Most clients have the effective key size in bits equal to
16349 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16350 A few however don't do this and instead use the size of the decrypted key
16351 to determine the RC2 key length and the AlgorithmIdentifier to determine
16352 the effective key length. In this case the effective key length can still
16353 be 40 bits but the key length can be 168 bits for example. This is fixed
16354 by manually forcing an RC2 key into the EVP_PKEY structure because the
16355 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16356 the key length and effective key length are equal.
16360 * Add a bunch of functions that should simplify the creation of
16361 X509_NAME structures. Now you should be able to do:
16362 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16363 and have it automatically work out the correct field type and fill in
16364 the structures. The more adventurous can try:
16365 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16366 and it will (hopefully) work out the correct multibyte encoding.
16370 * Change the 'req' utility to use the new field handling and multibyte
16371 copy routines. Before the DN field creation was handled in an ad hoc
16372 way in req, ca, and x509 which was rather broken and didn't support
16373 BMPStrings or UTF8Strings. Since some software doesn't implement
16374 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16375 using the dirstring_type option. See the new comment in the default
16376 openssl.cnf for more info.
16380 * Make crypto/rand/md_rand.c more robust:
16381 - Assure unique random numbers after fork().
16382 - Make sure that concurrent threads access the global counter and
16383 md serializably so that we never lose entropy in them
16384 or use exactly the same state in multiple threads.
16385 Access to the large state is not always serializable because
16386 the additional locking could be a performance killer, and
16387 md should be large enough anyway.
16391 * New file apps/app_rand.c with commonly needed functionality
16392 for handling the random seed file.
16394 Use the random seed file in some applications that previously did not:
16396 dsaparam -genkey (which also ignored its '-rand' option),
16399 x509 (when signing).
16400 Except on systems with /dev/urandom, it is crucial to have a random
16401 seed file at least for key creation, DSA signing, and for DH exchanges;
16402 for RSA signatures we could do without one.
16404 gendh and gendsa (unlike genrsa) used to read only the first byte
16405 of each file listed in the '-rand' option. The function as previously
16406 found in genrsa is now in app_rand.c and is used by all programs
16407 that support '-rand'.
16411 * In RAND_write_file, use mode 0600 for creating files;
16412 don't just chmod when it may be too late.
16416 * Report an error from X509_STORE_load_locations
16417 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16421 * New function ASN1_mbstring_copy() this copies a string in either
16422 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16423 into an ASN1_STRING type. A mask of permissible types is passed
16424 and it chooses the "minimal" type to use or an error if not type
16429 * Add function equivalents to the various macros in asn1.h. The old
16430 macros are retained with an `M_` prefix. Code inside the library can
16431 use the `M_` macros. External code (including the openssl utility)
16432 should *NOT* in order to be "shared library friendly".
16436 * Add various functions that can check a certificate's extensions
16437 to see if it usable for various purposes such as SSL client,
16438 server or S/MIME and CAs of these types. This is currently
16439 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16440 verification. Also added a -purpose flag to x509 utility to
16441 print out all the purposes.
16445 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16450 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16451 for, obtain and decode and extension and obtain its critical flag.
16452 This allows all the necessary extension code to be handled in a
16453 single function call.
16457 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16458 platforms. See crypto/rc4/rc4_enc.c for further details.
16462 * New -noout option to asn1parse. This causes no output to be produced
16463 its main use is when combined with -strparse and -out to extract data
16464 from a file (which may not be in ASN.1 format).
16468 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16469 when producing the local key id.
16471 *Richard Levitte <levitte@stacken.kth.se>*
16473 * New option -dhparam in s_server. This allows a DH parameter file to be
16474 stated explicitly. If it is not stated then it tries the first server
16475 certificate file. The previous behaviour hard coded the filename
16480 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16481 a public key to be input or output. For example:
16482 openssl rsa -in key.pem -pubout -out pubkey.pem
16483 Also added necessary DSA public key functions to handle this.
16487 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16488 in the message. This was handled by allowing
16489 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16491 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16493 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16494 to the end of the strings whereas this didn't. This would cause problems
16495 if strings read with d2i_ASN1_bytes() were later modified.
16497 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16499 * Fix for base64 decode bug. When a base64 bio reads only one line of
16500 data and it contains EOF it will end up returning an error. This is
16501 caused by input 46 bytes long. The cause is due to the way base64
16502 BIOs find the start of base64 encoded data. They do this by trying a
16503 trial decode on each line until they find one that works. When they
16504 do a flag is set and it starts again knowing it can pass all the
16505 data directly through the decoder. Unfortunately it doesn't reset
16506 the context it uses. This means that if EOF is reached an attempt
16507 is made to pass two EOFs through the context and this causes the
16508 resulting error. This can also cause other problems as well. As is
16509 usual with these problems it takes *ages* to find and the fix is
16510 trivial: move one line.
16512 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16514 * Ugly workaround to get s_client and s_server working under Windows. The
16515 old code wouldn't work because it needed to select() on sockets and the
16516 tty (for keypresses and to see if data could be written). Win32 only
16517 supports select() on sockets so we select() with a 1s timeout on the
16518 sockets and then see if any characters are waiting to be read, if none
16519 are present then we retry, we also assume we can always write data to
16520 the tty. This isn't nice because the code then blocks until we've
16521 received a complete line of data and it is effectively polling the
16522 keyboard at 1s intervals: however it's quite a bit better than not
16523 working at all :-) A dedicated Windows application might handle this
16524 with an event loop for example.
16528 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16529 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16530 will be called when RSA_sign() and RSA_verify() are used. This is useful
16531 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16532 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16533 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16534 This necessitated the support of an extra signature type NID_md5_sha1
16535 for SSL signatures and modifications to the SSL library to use it instead
16536 of calling RSA_public_decrypt() and RSA_private_encrypt().
16540 * Add new -verify -CAfile and -CApath options to the crl program, these
16541 will lookup a CRL issuers certificate and verify the signature in a
16542 similar way to the verify program. Tidy up the crl program so it
16543 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16544 less strict. It will now permit CRL extensions even if it is not
16545 a V2 CRL: this will allow it to tolerate some broken CRLs.
16549 * Initialize all non-automatic variables each time one of the openssl
16550 sub-programs is started (this is necessary as they may be started
16551 multiple times from the "OpenSSL>" prompt).
16553 *Lennart Bang, Bodo Moeller*
16555 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16556 removing all other RSA functionality (this is what NO_RSA does). This
16557 is so (for example) those in the US can disable those operations covered
16558 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16563 * Non-copying interface to BIO pairs.
16564 (still largely untested)
16568 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16569 ASCII string. This was handled independently in various places before.
16573 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16574 UTF8 strings a character at a time.
16578 * Use client_version from client hello to select the protocol
16579 (s23_srvr.c) and for RSA client key exchange verification
16580 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16584 * Add various utility functions to handle SPKACs, these were previously
16585 handled by poking round in the structure internals. Added new function
16586 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16587 print, verify and generate SPKACs. Based on an original idea from
16588 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16592 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16596 * Allow the config file extension section to be overwritten on the
16597 command line. Based on an original idea from Massimiliano Pala
16598 <madwolf@comune.modena.it>. The new option is called -extensions
16599 and can be applied to ca, req and x509. Also -reqexts to override
16600 the request extensions in req and -crlexts to override the crl extensions
16605 * Add new feature to the SPKAC handling in ca. Now you can include
16606 the same field multiple times by preceding it by "XXXX." for example:
16609 this is the same syntax as used in the req config file.
16613 * Allow certificate extensions to be added to certificate requests. These
16614 are specified in a 'req_extensions' option of the req section of the
16615 config file. They can be printed out with the -text option to req but
16616 are otherwise ignored at present.
16620 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16621 data read consists of only the final block it would not decrypted because
16622 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16623 A misplaced 'break' also meant the decrypted final block might not be
16624 copied until the next read.
16628 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16629 a few extra parameters to the DH structure: these will be useful if
16630 for example we want the value of 'q' or implement X9.42 DH.
16634 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16635 provides hooks that allow the default DSA functions or functions on a
16636 "per key" basis to be replaced. This allows hardware acceleration and
16637 hardware key storage to be handled without major modification to the
16638 library. Also added low level modexp hooks and CRYPTO_EX structure and
16639 associated functions.
16643 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16644 as "read only": it can't be written to and the buffer it points to will
16645 not be freed. Reading from a read only BIO is much more efficient than
16646 a normal memory BIO. This was added because there are several times when
16647 an area of memory needs to be read from a BIO. The previous method was
16648 to create a memory BIO and write the data to it, this results in two
16649 copies of the data and an O(n^2) reading algorithm. There is a new
16650 function BIO_new_mem_buf() which creates a read only memory BIO from
16651 an area of memory. Also modified the PKCS#7 routines to use read only
16656 * Bugfix: ssl23_get_client_hello did not work properly when called in
16657 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16658 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16659 but a retry condition occurred while trying to read the rest.
16663 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16664 NID_pkcs7_encrypted by default: this was wrong since this should almost
16665 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16666 the encrypted data type: this is a more sensible place to put it and it
16667 allows the PKCS#12 code to be tidied up that duplicated this
16672 * Changed obj_dat.pl script so it takes its input and output files on
16673 the command line. This should avoid shell escape redirection problems
16678 * Initial support for certificate extension requests, these are included
16679 in things like Xenroll certificate requests. Included functions to allow
16680 extensions to be obtained and added.
16684 * -crlf option to s_client and s_server for sending newlines as
16685 CRLF (as required by many protocols).
16689 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16691 * Install libRSAglue.a when OpenSSL is built with RSAref.
16693 *Ralf S. Engelschall*
16695 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16697 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16699 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16704 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16705 DH parameters/keys (q is lost during that conversion, but the resulting
16706 DH parameters contain its length).
16708 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16709 much faster than DH_generate_parameters (which creates parameters
16710 where `p = 2*q + 1`), and also the smaller q makes DH computations
16711 much more efficient (160-bit exponentiation instead of 1024-bit
16712 exponentiation); so this provides a convenient way to support DHE
16713 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16714 utter importance to use
16715 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16717 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16718 when such DH parameters are used, because otherwise small subgroup
16719 attacks may become possible!
16723 * Avoid memory leak in i2d_DHparams.
16727 * Allow the -k option to be used more than once in the enc program:
16728 this allows the same encrypted message to be read by multiple recipients.
16732 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16733 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16734 it will always use the numerical form of the OID, even if it has a short
16739 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16740 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16741 otherwise bn_mod_exp was called. In the case of hardware keys for example
16742 no private key components need be present and it might store extra data
16743 in the RSA structure, which cannot be accessed from bn_mod_exp.
16744 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16745 private key operations.
16749 * Added support for SPARC Linux.
16753 * pem_password_cb function type incompatibly changed from
16754 typedef int pem_password_cb(char *buf, int size, int rwflag);
16756 ....(char *buf, int size, int rwflag, void *userdata);
16757 so that applications can pass data to their callbacks:
16758 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16759 additional void * argument, which is just handed through whenever
16760 the password callback is called.
16762 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16764 New function SSL_CTX_set_default_passwd_cb_userdata.
16766 Compatibility note: As many C implementations push function arguments
16767 onto the stack in reverse order, the new library version is likely to
16768 interoperate with programs that have been compiled with the old
16769 pem_password_cb definition (PEM_whatever takes some data that
16770 happens to be on the stack as its last argument, and the callback
16771 just ignores this garbage); but there is no guarantee whatsoever that
16774 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16775 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16776 problems not only on Windows, but also on some Unix platforms.
16777 To avoid problematic command lines, these definitions are now in an
16778 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16779 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16783 * MIPS III/IV assembler module is reimplemented.
16787 * More DES library cleanups: remove references to srand/rand and
16788 delete an unused file.
16792 * Add support for the free Netwide assembler (NASM) under Win32,
16793 since not many people have MASM (ml) and it can be hard to obtain.
16794 This is currently experimental but it seems to work OK and pass all
16795 the tests. Check out INSTALL.W32 for info.
16799 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16800 without temporary keys kept an extra copy of the server key,
16801 and connections with temporary keys did not free everything in case
16806 * New function RSA_check_key and new openssl rsa option -check
16807 for verifying the consistency of RSA keys.
16809 *Ulf Moeller, Bodo Moeller*
16811 * Various changes to make Win32 compile work:
16812 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16813 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16814 comparison" warnings.
16815 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16819 * Add a debugging option to PKCS#5 v2 key generation function: when
16820 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16821 derived keys are printed to stderr.
16825 * Copy the flags in ASN1_STRING_dup().
16827 *Roman E. Pavlov <pre@mo.msk.ru>*
16829 * The x509 application mishandled signing requests containing DSA
16830 keys when the signing key was also DSA and the parameters didn't match.
16832 It was supposed to omit the parameters when they matched the signing key:
16833 the verifying software was then supposed to automatically use the CA's
16834 parameters if they were absent from the end user certificate.
16836 Omitting parameters is no longer recommended. The test was also
16837 the wrong way round! This was probably due to unusual behaviour in
16838 EVP_cmp_parameters() which returns 1 if the parameters match.
16839 This meant that parameters were omitted when they *didn't* match and
16840 the certificate was useless. Certificates signed with 'ca' didn't have
16843 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16845 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16846 The interface is as follows:
16847 Applications can use
16848 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16849 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16850 "off" is now the default.
16851 The library internally uses
16852 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16853 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16854 to disable memory-checking temporarily.
16856 Some inconsistent states that previously were possible (and were
16857 even the default) are now avoided.
16859 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16860 with each memory chunk allocated; this is occasionally more helpful
16861 than just having a counter.
16863 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16865 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16870 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16871 which largely parallels "options", but is for changing API behaviour,
16872 whereas "options" are about protocol behaviour.
16873 Initial "mode" flags are:
16875 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16876 a single record has been written.
16877 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16878 retries use the same buffer location.
16879 (But all of the contents must be
16884 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16887 * Fix problems with no-hmac etc.
16889 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16891 * New functions RSA_get_default_method(), RSA_set_method() and
16892 RSA_get_method(). These allows replacement of RSA_METHODs without having
16893 to mess around with the internals of an RSA structure.
16897 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16898 Also really enable memory leak checks in openssl.c and in some
16901 *Chad C. Mulligan, Bodo Moeller*
16903 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16904 up the length of negative integers. This has now been simplified to just
16905 store the length when it is first determined and use it later, rather
16906 than trying to keep track of where data is copied and updating it to
16908 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16910 * Add a new function PKCS7_signatureVerify. This allows the verification
16911 of a PKCS#7 signature but with the signing certificate passed to the
16912 function itself. This contrasts with PKCS7_dataVerify which assumes the
16913 certificate is present in the PKCS#7 structure. This isn't always the
16914 case: certificates can be omitted from a PKCS#7 structure and be
16915 distributed by "out of band" means (such as a certificate database).
16919 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16920 function prototypes in pem.h, also change util/mkdef.pl to add the
16921 necessary function names.
16925 * mk1mf.pl (used by Windows builds) did not properly read the
16926 options set by Configure in the top level Makefile, and Configure
16927 was not even able to write more than one option correctly.
16928 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16932 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16933 file to be loaded from a BIO or FILE pointer. The BIO version will
16934 for example allow memory BIOs to contain config info.
16938 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16939 Whoever hopes to achieve shared-library compatibility across versions
16940 must use this, not the compile-time macro.
16941 (Exercise 0.9.4: Which is the minimum library version required by
16943 Note: All this applies only to multi-threaded programs, others don't
16948 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16949 through a BIO pair triggered the default case, i.e.
16950 SSLerr(...,SSL_R_UNKNOWN_STATE).
16954 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16955 can use the SSL library even if none of the specific BIOs is
16960 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16961 for the encoded length.
16963 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16965 * Add initial documentation of the X509V3 functions.
16969 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16970 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16971 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16972 secure PKCS#8 private key format with a high iteration count.
16976 * Fix determination of Perl interpreter: A perl or perl5
16977 *directory* in $PATH was also accepted as the interpreter.
16979 *Ralf S. Engelschall*
16981 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16982 wrong with it but it was very old and did things like calling
16983 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16984 unusual formatting.
16988 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16989 to use the new extension code.
16993 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16994 with macros. This should make it easier to change their form, add extra
16995 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17000 * Add to configuration table a new entry that can specify an alternative
17001 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17002 according to Mark Crispin <MRC@Panda.COM>.
17006 * DES CBC did not update the IV. Weird.
17010 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17011 Changing the behaviour of the former might break existing programs --
17012 where IV updating is needed, des_ncbc_encrypt can be used.
17015 * When bntest is run from "make test" it drives bc to check its
17016 calculations, as well as internally checking them. If an internal check
17017 fails, it needs to cause bc to give a non-zero result or make test carries
17018 on without noticing the failure. Fixed.
17022 * DES library cleanups.
17026 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17027 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17028 ciphers. NOTE: although the key derivation function has been verified
17029 against some published test vectors it has not been extensively tested
17030 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17035 * Instead of "mkdir -p", which is not fully portable, use new
17036 Perl script "util/mkdir-p.pl".
17040 * Rewrite the way password based encryption (PBE) is handled. It used to
17041 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17042 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17043 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17044 the 'parameter' field of the AlgorithmIdentifier is passed to the
17045 underlying key generation function so it must do its own ASN1 parsing.
17046 This has also changed the EVP_PBE_CipherInit() function which now has a
17047 'parameter' argument instead of literal salt and iteration count values
17048 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17052 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17053 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17054 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17055 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17056 value was just used as a "magic string" and not used directly its
17057 value doesn't matter.
17061 * Introduce some semblance of const correctness to BN. Shame C doesn't
17066 * "linux-sparc64" configuration (ultrapenguin).
17068 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17069 "linux-sparc" configuration.
17071 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17073 * config now generates no-xxx options for missing ciphers.
17077 * Support the EBCDIC character set (work in progress).
17078 File ebcdic.c not yet included because it has a different license.
17080 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17082 * Support BS2000/OSD-POSIX.
17084 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17086 * Make callbacks for key generation use `void *` instead of `char *`.
17090 * Make S/MIME samples compile (not yet tested).
17094 * Additional typesafe stacks.
17098 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17102 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17104 * New configuration variant "sco5-gcc".
17106 * Updated some demos.
17108 *Sean O Riordain, Wade Scholine*
17110 * Add missing BIO_free at exit of pkcs12 application.
17114 * Fix memory leak in conf.c.
17118 * Updates for Win32 to assembler version of MD5.
17122 * Set #! path to perl in apps/der_chop to where we found it
17123 instead of using a fixed path.
17127 * SHA library changes for irix64-mips4-cc.
17131 * Improvements for VMS support.
17135 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17137 * Bignum library bug fix. IRIX 6 passes "make test" now!
17138 This also avoids the problems with SC4.2 and unpatched SC5.
17140 *Andy Polyakov <appro@fy.chalmers.se>*
17142 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17143 These are required because of the typesafe stack would otherwise break
17144 existing code. If old code used a structure member which used to be STACK
17145 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17146 sk_num or sk_value it would produce an error because the num, data members
17147 are not present in STACK_OF. Now it just produces a warning. sk_set
17148 replaces the old method of assigning a value to sk_value
17149 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17150 that does this will no longer work (and should use sk_set instead) but
17151 this could be regarded as a "questionable" behaviour anyway.
17155 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17156 correctly handle encrypted S/MIME data.
17160 * Change type of various DES function arguments from des_cblock
17161 (which means, in function argument declarations, pointer to char)
17162 to des_cblock * (meaning pointer to array with 8 char elements),
17163 which allows the compiler to do more typechecking; it was like
17164 that back in SSLeay, but with lots of ugly casts.
17166 Introduce new type const_des_cblock.
17170 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17171 problems: find RecipientInfo structure that matches recipient certificate
17172 and initialise the ASN1 structures properly based on passed cipher.
17176 * Belatedly make the BN tests actually check the results.
17180 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17181 to and from BNs: it was completely broken. New compilation option
17182 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17183 key elements as negative integers.
17187 * Reorganize and speed up MD5.
17189 *Andy Polyakov <appro@fy.chalmers.se>*
17193 *Richard Levitte <richard@levitte.org>*
17195 * New option -out to asn1parse to allow the parsed structure to be
17196 output to a file. This is most useful when combined with the -strparse
17197 option to examine the output of things like OCTET STRINGS.
17201 * Make SSL library a little more fool-proof by not requiring any longer
17202 that `SSL_set_{accept,connect}_state` be called before
17203 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17204 in many applications because usually everything *appeared* to work as
17205 intended anyway -- now it really works as intended).
17209 * Move openssl.cnf out of lib/.
17213 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17214 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17215 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17217 *Ralf S. Engelschall*
17219 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17220 handle PKCS#7 enveloped data properly.
17222 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17224 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17225 copying pointers. The cert_st handling is changed by this in
17226 various ways (and thus what used to be known as ctx->default_cert
17227 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17228 any longer when s->cert does not give us what we need).
17229 ssl_cert_instantiate becomes obsolete by this change.
17230 As soon as we've got the new code right (possibly it already is?),
17231 we have solved a couple of bugs of the earlier code where s->cert
17232 was used as if it could not have been shared with other SSL structures.
17234 Note that using the SSL API in certain dirty ways now will result
17235 in different behaviour than observed with earlier library versions:
17236 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17237 does not influence s as it used to.
17239 In order to clean up things more thoroughly, inside SSL_SESSION
17240 we don't use CERT any longer, but a new structure SESS_CERT
17241 that holds per-session data (if available); currently, this is
17242 the peer's certificate chain and, for clients, the server's certificate
17243 and temporary key. CERT holds only those values that can have
17244 meaningful defaults in an SSL_CTX.
17248 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17249 from the internal representation. Various PKCS#7 fixes: remove some
17250 evil casts and set the enc_dig_alg field properly based on the signing
17255 * Allow PKCS#12 password to be set from the command line or the
17256 environment. Let 'ca' get its config file name from the environment
17257 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17262 * Allow certificate policies extension to use an IA5STRING for the
17263 organization field. This is contrary to the PKIX definition but
17264 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17269 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17270 without disallowing inline assembler and the like for non-pedantic builds.
17274 * Support Borland C++ builder.
17276 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17282 * SHA-1 cleanups and performance enhancements.
17284 *Andy Polyakov <appro@fy.chalmers.se>*
17286 * Sparc v8plus assembler for the bignum library.
17288 *Andy Polyakov <appro@fy.chalmers.se>*
17290 * Accept any -xxx and +xxx compiler options in Configure.
17294 * Update HPUX configuration.
17298 * Add missing `sk_<type>_unshift()` function to safestack.h
17300 *Ralf S. Engelschall*
17302 * New function SSL_CTX_use_certificate_chain_file that sets the
17303 "extra_cert"s in addition to the certificate. (This makes sense
17304 only for "PEM" format files, as chains as a whole are not
17309 * Support verify_depth from the SSL API.
17310 x509_vfy.c had what can be considered an off-by-one-error:
17311 Its depth (which was not part of the external interface)
17312 was actually counting the number of certificates in a chain;
17313 now it really counts the depth.
17317 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17318 instead of X509err, which often resulted in confusing error
17319 messages since the error codes are not globally unique
17320 (e.g. an alleged error in ssl3_accept when a certificate
17321 didn't match the private key).
17323 * New function SSL_CTX_set_session_id_context that allows to set a default
17324 value (so that you don't need SSL_set_session_id_context for each
17325 connection using the SSL_CTX).
17329 * OAEP decoding bug fix.
17333 * Support INSTALL_PREFIX for package builders, as proposed by
17338 * New Configure options "threads" and "no-threads". For systems
17339 where the proper compiler options are known (currently Solaris
17340 and Linux), "threads" is the default.
17344 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17348 * Install various scripts to $(OPENSSLDIR)/misc, not to
17349 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17350 such as /usr/local/bin.
17354 * "make linux-shared" to build shared libraries.
17356 *Niels Poppe <niels@netbox.org>*
17358 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17362 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17363 extension adding in x509 utility.
17367 * Remove NOPROTO sections and error code comments.
17371 * Partial rewrite of the DEF file generator to now parse the ANSI
17376 * New Configure options --prefix=DIR and --openssldir=DIR.
17380 * Complete rewrite of the error code script(s). It is all now handled
17381 by one script at the top level which handles error code gathering,
17382 header rewriting and C source file generation. It should be much better
17383 than the old method: it now uses a modified version of Ulf's parser to
17384 read the ANSI prototypes in all header files (thus the old K&R definitions
17385 aren't needed for error creation any more) and do a better job of
17386 translating function codes into names. The old 'ASN1 error code embedded
17387 in a comment' is no longer necessary and it doesn't use .err files which
17388 have now been deleted. Also the error code call doesn't have to appear all
17389 on one line (which resulted in some large lines...).
17393 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17397 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17398 0 (which usually indicates a closed connection), but continue reading.
17402 * Fix some race conditions.
17406 * Add support for CRL distribution points extension. Add Certificate
17407 Policies and CRL distribution points documentation.
17411 * Move the autogenerated header file parts to crypto/opensslconf.h.
17415 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17416 8 of keying material. Merlin has also confirmed interop with this fix
17417 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17419 *Merlin Hughes <merlin@baltimore.ie>*
17421 * Fix lots of warnings.
17423 *Richard Levitte <levitte@stacken.kth.se>*
17425 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17426 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17428 *Richard Levitte <levitte@stacken.kth.se>*
17430 * Fix problems with sizeof(long) == 8.
17432 *Andy Polyakov <appro@fy.chalmers.se>*
17434 * Change functions to ANSI C.
17438 * Fix typos in error codes.
17440 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17442 * Remove defunct assembler files from Configure.
17446 * SPARC v8 assembler BIGNUM implementation.
17448 *Andy Polyakov <appro@fy.chalmers.se>*
17450 * Support for Certificate Policies extension: both print and set.
17451 Various additions to support the r2i method this uses.
17455 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17456 return a const string when you are expecting an allocated buffer.
17460 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17461 types DirectoryString and DisplayText.
17465 * Add code to allow r2i extensions to access the configuration database,
17466 add an LHASH database driver and add several ctx helper functions.
17470 * Fix an evil bug in bn_expand2() which caused various BN functions to
17471 fail when they extended the size of a BIGNUM.
17475 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17476 support typesafe stack.
17480 * Fix typo in SSL_[gs]et_options().
17482 *Nils Frostberg <nils@medcom.se>*
17484 * Delete various functions and files that belonged to the (now obsolete)
17485 old X509V3 handling code.
17489 * New Configure option "rsaref".
17493 * Don't auto-generate pem.h.
17497 * Introduce type-safe ASN.1 SETs.
17501 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17503 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17505 * Introduce type-safe STACKs. This will almost certainly break lots of code
17506 that links with OpenSSL (well at least cause lots of warnings), but fear
17507 not: the conversion is trivial, and it eliminates loads of evil casts. A
17508 few STACKed things have been converted already. Feel free to convert more.
17509 In the fullness of time, I'll do away with the STACK type altogether.
17513 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17514 specified in `<certfile>` by updating the entry in the index.txt file.
17515 This way one no longer has to edit the index.txt file manually for
17516 revoking a certificate. The -revoke option does the gory details now.
17518 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17520 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17521 `-text` option at all and this way the `-noout -text` combination was
17522 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17524 *Ralf S. Engelschall*
17526 * Make sure a corresponding plain text error message exists for the
17527 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17528 verify callback function determined that a certificate was revoked.
17530 *Ralf S. Engelschall*
17532 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17533 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17534 all available ciphers including rc5, which was forgotten until now.
17535 In order to let the testing shell script know which algorithms
17536 are available, a new (up to now undocumented) command
17537 `openssl list-cipher-commands` is used.
17541 * Bugfix: s_client occasionally would sleep in select() when
17542 it should have checked SSL_pending() first.
17546 * New functions DSA_do_sign and DSA_do_verify to provide access to
17547 the raw DSA values prior to ASN.1 encoding.
17551 * Tweaks to Configure
17553 *Niels Poppe <niels@netbox.org>*
17555 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17560 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17564 * New config option to avoid instructions that are illegal on the 80386.
17565 The default code is faster, but requires at least a 486.
17569 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17570 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17571 same as SSL2_VERSION anyway.
17575 * New "-showcerts" option for s_client.
17579 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17580 application. Various cleanups and fixes.
17584 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17585 modify error routines to work internally. Add error codes and PBE init
17586 to library startup routines.
17590 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17591 packing functions to asn1 and evp. Changed function names and error
17592 codes along the way.
17596 * PKCS12 integration: and so it begins... First of several patches to
17597 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17598 objects to objects.h
17602 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17603 and display support for Thawte strong extranet extension.
17607 * Add LinuxPPC support.
17609 *Jeff Dubrule <igor@pobox.org>*
17611 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17612 bn_div_words in alpha.s.
17614 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17616 * Make sure the RSA OAEP test is skipped under -DRSAref because
17617 OAEP isn't supported when OpenSSL is built with RSAref.
17619 *Ulf Moeller <ulf@fitug.de>*
17621 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17622 so they no longer are missing under -DNOPROTO.
17624 *Soren S. Jorvang <soren@t.dk>*
17626 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17628 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17629 doesn't work when the session is reused. Coming soon!
17633 * Fix a security hole, that allows sessions to be reused in the wrong
17634 context thus bypassing client cert protection! All software that uses
17635 client certs and session caches in multiple contexts NEEDS PATCHING to
17636 allow session reuse! A fuller solution is in the works.
17638 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17640 * Some more source tree cleanups (removed obsolete files
17641 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17642 permission on "config" script to be executable) and a fix for the INSTALL
17645 *Ulf Moeller <ulf@fitug.de>*
17647 * Remove some legacy and erroneous uses of malloc, free instead of
17650 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17652 * Make rsa_oaep_test return non-zero on error.
17654 *Ulf Moeller <ulf@fitug.de>*
17656 * Add support for native Solaris shared libraries. Configure
17657 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17658 if someone would make that last step automatic.
17660 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17662 * ctx_size was not built with the right compiler during "make links". Fixed.
17666 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17667 except NULL ciphers". This means the default cipher list will no longer
17668 enable NULL ciphers. They need to be specifically enabled e.g. with
17669 the string "DEFAULT:eNULL".
17673 * Fix to RSA private encryption routines: if p < q then it would
17674 occasionally produce an invalid result. This will only happen with
17675 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17679 * Be less restrictive and allow also `perl util/perlpath.pl
17680 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17681 because this way one can also use an interpreter named `perl5` (which is
17682 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17683 installed as `perl`).
17685 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17687 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17689 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17691 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17692 advapi32.lib to Win32 build and change the pem test comparison
17693 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17694 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17695 and crypto/des/ede_cbcm_enc.c.
17699 * DES quad checksum was broken on big-endian architectures. Fixed.
17703 * Comment out two functions in bio.h that aren't implemented. Fix up the
17704 Win32 test batch file so it (might) work again. The Win32 test batch file
17705 is horrible: I feel ill....
17709 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17710 in e_os.h. Audit of header files to check ANSI and non ANSI
17711 sections: 10 functions were absent from non ANSI section and not exported
17712 from Windows DLLs. Fixed up libeay.num for new functions.
17716 * Make `openssl version` output lines consistent.
17718 *Ralf S. Engelschall*
17720 * Fix Win32 symbol export lists for BIO functions: Added
17721 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17722 to ms/libeay{16,32}.def.
17724 *Ralf S. Engelschall*
17726 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17727 fine under Unix and passes some trivial tests I've now added. But the
17728 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17729 added to make sure no one expects that this stuff really works in the
17730 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17731 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17734 *Ralf S. Engelschall*
17736 * Fix the generation of two part addresses in perl.
17738 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17740 * Add config entry for Linux on MIPS.
17742 *John Tobey <jtobey@channel1.com>*
17744 * Make links whenever Configure is run, unless we are on Windoze.
17748 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17749 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17754 * Add a useful kludge to allow package maintainers to specify compiler and
17755 other platforms details on the command line without having to patch the
17756 Configure script every time: One now can use
17757 `perl Configure <id>:<details>`,
17758 i.e. platform ids are allowed to have details appended
17759 to them (separated by colons). This is treated as there would be a static
17760 pre-configured entry in Configure's %table under key `<id>` with value
17761 `<details>` and `perl Configure <id>` is called. So, when you want to
17762 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17763 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17764 now, which overrides the FreeBSD-elf entry on-the-fly.
17766 *Ralf S. Engelschall*
17768 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17772 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17773 on the `perl Configure ...` command line. This way one can compile
17774 OpenSSL libraries with Position Independent Code (PIC) which is needed
17775 for linking it into DSOs.
17777 *Ralf S. Engelschall*
17779 * Remarkably, export ciphers were totally broken and no-one had noticed!
17784 * Cleaned up the LICENSE document: The official contact for any license
17785 questions now is the OpenSSL core team under openssl-core@openssl.org.
17786 And add a paragraph about the dual-license situation to make sure people
17787 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17788 to the OpenSSL toolkit.
17790 *Ralf S. Engelschall*
17792 * General source tree makefile cleanups: Made `making xxx in yyy...`
17793 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17794 Additionally cleaned up the `make links` target: Remove unnecessary
17795 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17796 to speed processing and no longer clutter the display with confusing
17797 stuff. Instead only the actually done links are displayed.
17799 *Ralf S. Engelschall*
17801 * Permit null encryption ciphersuites, used for authentication only. It used
17802 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17803 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17808 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17809 signed attributes when verifying signatures (this would break them),
17810 the detached data encoding was wrong and public keys obtained using
17811 X509_get_pubkey() weren't freed.
17815 * Add text documentation for the BUFFER functions. Also added a work around
17816 to a Win95 console bug. This was triggered by the password read stuff: the
17817 last character typed gets carried over to the next fread(). If you were
17818 generating a new cert request using 'req' for example then the last
17819 character of the passphrase would be CR which would then enter the first
17824 * Added the new 'Includes OpenSSL Cryptography Software' button as
17825 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17826 button and can be used by applications based on OpenSSL to show the
17827 relationship to the OpenSSL project.
17829 *Ralf S. Engelschall*
17831 * Remove confusing variables in function signatures in files
17832 ssl/ssl_lib.c and ssl/ssl.h.
17834 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17836 * Don't install bss_file.c under PREFIX/include/
17838 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17840 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17841 functions that return function pointers and has support for NT specific
17842 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17843 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17844 unsigned to signed types: this was killing the Win32 compile.
17848 * Add new certificate file to stack functions,
17849 SSL_add_dir_cert_subjects_to_stack() and
17850 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17851 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17852 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17853 This means that Apache-SSL and similar packages don't have to mess around
17854 to add as many CAs as they want to the preferred list.
17858 * Experiment with doxygen documentation. Currently only partially applied to
17860 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17861 openssl.doxy as the configuration file.
17865 * Get rid of remaining C++-style comments which strict C compilers hate.
17867 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17869 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17870 compiled in by default: it has problems with large keys.
17874 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17875 DH private keys and/or callback functions which directly correspond to
17876 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17877 is needed for applications which have to configure certificates on a
17878 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17880 For the RSA certificate situation is makes no difference, but
17881 for the DSA certificate situation this fixes the "no shared cipher"
17882 problem where the OpenSSL cipher selection procedure failed because the
17883 temporary keys were not overtaken from the context and the API provided
17884 no way to reconfigure them.
17885 The new functions now let applications reconfigure the stuff and they
17886 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17887 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17888 non-public-API function ssl_cert_instantiate() is used as a helper
17889 function and also to reduce code redundancy inside ssl_rsa.c.
17891 *Ralf S. Engelschall*
17893 * Move s_server -dcert and -dkey options out of the undocumented feature
17894 area because they are useful for the DSA situation and should be
17895 recognized by the users.
17897 *Ralf S. Engelschall*
17899 * Fix the cipher decision scheme for export ciphers: the export bits are
17900 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17901 SSL_EXP_MASK. So, the original variable has to be used instead of the
17902 already masked variable.
17904 *Richard Levitte <levitte@stacken.kth.se>*
17906 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17908 *Richard Levitte <levitte@stacken.kth.se>*
17910 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17911 from `int` to `unsigned int` because it is a length and initialized by
17912 EVP_DigestFinal() which expects an `unsigned int *`.
17914 *Richard Levitte <levitte@stacken.kth.se>*
17916 * Don't hard-code path to Perl interpreter on shebang line of Configure
17917 script. Instead use the usual Shell->Perl transition trick.
17919 *Ralf S. Engelschall*
17921 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17922 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17923 -noout -modulus` as it's already the case for `openssl rsa -noout
17924 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17925 currently the public key is printed (a decision which was already done by
17926 `openssl dsa -modulus` in the past) which serves a similar purpose.
17927 Additionally the NO_RSA no longer completely removes the whole -modulus
17928 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17931 *Ralf S. Engelschall*
17933 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17934 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17936 *Arne Ansper <arne@ats.cyber.ee>*
17938 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17939 to be added. Now both 'req' and 'ca' can use new objects defined in the
17944 * Add cool BIO that does syslog (or event log on NT).
17946 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17948 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17949 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17950 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17951 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17955 * Add preliminary config info for new extension code.
17959 * Make RSA_NO_PADDING really use no padding.
17961 *Ulf Moeller <ulf@fitug.de>*
17963 * Generate errors when private/public key check is done.
17967 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17968 for some CRL extensions and new objects added.
17972 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17973 key usage extension and fuller support for authority key id.
17977 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17978 padding method for RSA, which is recommended for new applications in PKCS
17979 #1 v2.0 (RFC 2437, October 1998).
17980 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17981 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17982 against Bleichbacher's attack on RSA.
17983 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17986 * Updates to the new SSL compression code
17988 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17990 * Fix so that the version number in the master secret, when passed
17991 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17992 (because the server will not accept higher), that the version number
17993 is 0x03,0x01, not 0x03,0x00
17995 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17997 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17998 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17999 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18003 * Support for RAW extensions where an arbitrary extension can be
18004 created by including its DER encoding. See apps/openssl.cnf for
18009 * Make sure latest Perl versions don't interpret some generated C array
18010 code as Perl array code in the crypto/err/err_genc.pl script.
18012 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18014 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18015 not many people have the assembler. Various Win32 compilation fixes and
18016 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18017 build instructions.
18021 * Modify configure script 'Configure' to automatically create crypto/date.h
18022 file under Win32 and also build pem.h from pem.org. New script
18023 util/mkfiles.pl to create the MINFO file on environments that can't do a
18024 'make files': perl util/mkfiles.pl >MINFO should work.
18028 * Major rework of DES function declarations, in the pursuit of correctness
18029 and purity. As a result, many evil casts evaporated, and some weirdness,
18030 too. You may find this causes warnings in your code. Zapping your evil
18031 casts will probably fix them. Mostly.
18035 * Fix for a typo in asn1.h. Bug fix to object creation script
18036 obj_dat.pl. It considered a zero in an object definition to mean
18037 "end of object": none of the objects in objects.h have any zeros
18038 so it wasn't spotted.
18040 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18042 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18043 Masking (CBCM). In the absence of test vectors, the best I have been able
18044 to do is check that the decrypt undoes the encrypt, so far. Send me test
18045 vectors if you have them.
18049 * Correct calculation of key length for export ciphers (too much space was
18050 allocated for null ciphers). This has not been tested!
18054 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18055 message is now correct (it understands "crypto" and "ssl" on its
18056 command line). There is also now an "update" option. This will update
18057 the util/ssleay.num and util/libeay.num files with any new functions.
18059 perl util/mkdef.pl crypto ssl update
18060 it will update them.
18064 * Overhauled the Perl interface:
18065 - ported BN stuff to OpenSSL's different BN library
18066 - made the perl/ source tree CVS-aware
18067 - renamed the package from SSLeay to OpenSSL (the files still contain
18068 their history because I've copied them in the repository)
18069 - removed obsolete files (the test scripts will be replaced
18070 by better Test::Harness variants in the future)
18072 *Ralf S. Engelschall*
18074 * First cut for a very conservative source tree cleanup:
18075 1. merge various obsolete readme texts into doc/ssleay.txt
18076 where we collect the old documents and readme texts.
18077 2. remove the first part of files where I'm already sure that we no
18078 longer need them because of three reasons: either they are just temporary
18079 files which were left by Eric or they are preserved original files where
18080 I've verified that the diff is also available in the CVS via "cvs diff
18081 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18082 the crypto/md/ stuff).
18084 *Ralf S. Engelschall*
18086 * More extension code. Incomplete support for subject and issuer alt
18087 name, issuer and authority key id. Change the i2v function parameters
18088 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18089 what that's for :-) Fix to ASN1 macro which messed up
18090 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18094 * Preliminary support for ENUMERATED type. This is largely copied from the
18099 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18101 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18103 * Make sure `make rehash` target really finds the `openssl` program.
18105 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18107 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18108 like to hear about it if this slows down other processors.
18112 * Add CygWin32 platform information to Configure script.
18114 *Alan Batie <batie@aahz.jf.intel.com>*
18116 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18118 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18120 * New program nseq to manipulate netscape certificate sequences
18124 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18129 * Fixes to BN code. Previously the default was to define BN_RECURSION
18130 but the BN code had some problems that would cause failures when
18131 doing certificate verification and some other functions.
18133 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18135 * Add ASN1 and PEM code to support netscape certificate sequences.
18139 * Add ASN1 and PEM code to support netscape certificate sequences.
18143 * Add several PKIX and private extended key usage OIDs.
18147 * Modify the 'ca' program to handle the new extension code. Modify
18148 openssl.cnf for new extension format, add comments.
18152 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18153 and add a sample to openssl.cnf so req -x509 now adds appropriate
18158 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18159 error code, add initial support to X509_print() and x509 application.
18163 * Takes a deep breath and start adding X509 V3 extension support code. Add
18164 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18165 stuff is currently isolated and isn't even compiled yet.
18169 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18170 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18171 Removed the versions check from X509 routines when loading extensions:
18172 this allows certain broken certificates that don't set the version
18173 properly to be processed.
18177 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18178 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18179 can still be regenerated with "make depend".
18183 * Spelling mistake in C version of CAST-128.
18185 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18187 * Changes to the error generation code. The perl script err-code.pl
18188 now reads in the old error codes and retains the old numbers, only
18189 adding new ones if necessary. It also only changes the .err files if new
18190 codes are added. The makefiles have been modified to only insert errors
18191 when needed (to avoid needlessly modifying header files). This is done
18192 by only inserting errors if the .err file is newer than the auto generated
18193 C file. To rebuild all the error codes from scratch (the old behaviour)
18194 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18195 or delete all the .err files.
18199 * CAST-128 was incorrectly implemented for short keys. The C version has
18200 been fixed, but is untested. The assembler versions are also fixed, but
18201 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18202 to regenerate it if needed.
18203 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18204 Hagino <itojun@kame.net>*
18206 * File was opened incorrectly in randfile.c.
18208 *Ulf Möller <ulf@fitug.de>*
18210 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18211 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18212 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18213 al: it's just almost always a UTCTime. Note this patch adds new error
18214 codes so do a "make errors" if there are problems.
18218 * Correct Linux 1 recognition in config.
18220 *Ulf Möller <ulf@fitug.de>*
18222 * Remove pointless MD5 hash when using DSA keys in ca.
18224 *Anonymous <nobody@replay.com>*
18226 * Generate an error if given an empty string as a cert directory. Also
18227 generate an error if handed NULL (previously returned 0 to indicate an
18228 error, but didn't set one).
18230 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18232 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18236 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18237 parameters. This was causing a warning which killed off the Win32 compile.
18241 * Remove C++ style comments from crypto/bn/bn_local.h.
18243 *Neil Costigan <neil.costigan@celocom.com>*
18245 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18246 based on a text string, looking up short and long names and finally
18247 "dot" format. The "dot" format stuff didn't work. Added new function
18248 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18249 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18250 OID is not part of the table.
18254 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18255 X509_LOOKUP_by_alias().
18259 * Sort openssl functions by name.
18263 * Get the gendsa program working (hopefully) and add it to app list. Remove
18264 encryption from sample DSA keys (in case anyone is interested the password
18269 * Make *all* `*_free` functions accept a NULL pointer.
18271 *Frans Heymans <fheymans@isaserver.be>*
18273 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18276 *Anonymous <nobody@replay.com>*
18278 * s_server should send the CAfile as acceptable CAs, not its own cert.
18280 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18282 * Don't blow it for numeric -newkey arguments to apps/req.
18284 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18286 * Temp key "for export" tests were wrong in s3_srvr.c.
18288 *Anonymous <nobody@replay.com>*
18290 * Add prototype for temp key callback functions
18291 SSL_CTX_set_tmp_{rsa,dh}_callback().
18295 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18296 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18300 * X509_name_add_entry() freed the wrong thing after an error.
18302 *Arne Ansper <arne@ats.cyber.ee>*
18304 * rsa_eay.c would attempt to free a NULL context.
18306 *Arne Ansper <arne@ats.cyber.ee>*
18308 * BIO_s_socket() had a broken should_retry() on Windoze.
18310 *Arne Ansper <arne@ats.cyber.ee>*
18312 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18314 *Arne Ansper <arne@ats.cyber.ee>*
18316 * Make sure the already existing X509_STORE->depth variable is initialized
18317 in X509_STORE_new(), but document the fact that this variable is still
18318 unused in the certificate verification process.
18320 *Ralf S. Engelschall*
18322 * Fix the various library and apps files to free up pkeys obtained from
18323 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18327 * Fix reference counting in X509_PUBKEY_get(). This makes
18328 demos/maurice/example2.c work, amongst others, probably.
18330 *Steve Henson and Ben Laurie*
18332 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18333 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18334 are no longer created. This way we have a single and consistent command
18335 line interface `openssl <command>`, similar to `cvs <command>`.
18337 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18339 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18340 BIT STRING wrapper always have zero unused bits.
18344 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18348 * Make the top-level INSTALL documentation easier to understand.
18352 * Makefiles updated to exit if an error occurs in a sub-directory
18353 make (including if user presses ^C) [Paul Sutton]
18355 * Make Montgomery context stuff explicit in RSA data structure.
18359 * Fix build order of pem and err to allow for generated pem.h.
18363 * Fix renumbering bug in X509_NAME_delete_entry().
18367 * Enhanced the err-ins.pl script so it makes the error library number
18368 global and can add a library name. This is needed for external ASN1 and
18369 other error libraries.
18373 * Fixed sk_insert which never worked properly.
18377 * Fix ASN1 macros so they can handle indefinite length constructed
18378 EXPLICIT tags. Some non standard certificates use these: they can now
18383 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18384 into a single doc/ssleay.txt bundle. This way the information is still
18385 preserved but no longer messes up this directory. Now it's new room for
18386 the new set of documentation files.
18388 *Ralf S. Engelschall*
18390 * SETs were incorrectly DER encoded. This was a major pain, because they
18391 shared code with SEQUENCEs, which aren't coded the same. This means that
18392 almost everything to do with SETs or SEQUENCEs has either changed name or
18393 number of arguments.
18395 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18397 * Fix test data to work with the above.
18401 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18402 was already fixed by Eric for 0.9.1 it seems.
18404 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18406 * Autodetect FreeBSD3.
18410 * Fix various bugs in Configure. This affects the following platforms:
18414 unixware-2.0-pentium
18419 * Eliminate generated files from CVS. Reorder tests to regenerate files
18420 before they are needed.
18424 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18428 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18430 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18431 changed SSLeay to OpenSSL in version strings.
18433 *Ralf S. Engelschall*
18435 * Some fixups to the top-level documents.
18439 * Fixed the nasty bug where rsaref.h was not found under compile-time
18440 because the symlink to include/ was missing.
18442 *Ralf S. Engelschall*
18444 * Incorporated the popular no-RSA/DSA-only patches
18445 which allow to compile a RSA-free SSLeay.
18447 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18449 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18450 when "ssleay" is still not found.
18452 *Ralf S. Engelschall*
18454 * Added more platforms to Configure: Cray T3E, HPUX 11,
18456 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18458 * Updated the README file.
18460 *Ralf S. Engelschall*
18462 * Added various .cvsignore files in the CVS repository subdirs
18463 to make a "cvs update" really silent.
18465 *Ralf S. Engelschall*
18467 * Recompiled the error-definition header files and added
18468 missing symbols to the Win32 linker tables.
18470 *Ralf S. Engelschall*
18472 * Cleaned up the top-level documents;
18473 o new files: CHANGES and LICENSE
18474 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18475 o merged COPYRIGHT into LICENSE
18476 o removed obsolete TODO file
18477 o renamed MICROSOFT to INSTALL.W32
18479 *Ralf S. Engelschall*
18481 * Removed dummy files from the 0.9.1b source tree:
18482 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18483 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18484 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18485 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18486 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18488 *Ralf S. Engelschall*
18490 * Added various platform portability fixes.
18494 * The Genesis of the OpenSSL rpject:
18495 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18496 Young and Tim J. Hudson created while they were working for C2Net until
18499 *The OpenSSL Project*
18501 ### Changes between 0.9.0b and 0.9.1b [not released]
18503 * Updated a few CA certificates under certs/
18507 * Changed some BIGNUM api stuff.
18511 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18512 DGUX x86, Linux Alpha, etc.
18516 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18517 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18522 * Add -strparse option to asn1pars program which parses nested
18525 *Dr Stephen Henson <shenson@bigfoot.com>*
18527 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18531 * DSA fix for "ca" program.
18535 * Added "-genkey" option to "dsaparam" program.
18539 * Added RIPE MD160 (rmd160) message digest.
18543 * Added -a (all) option to "ssleay version" command.
18547 * Added PLATFORM define which is the id given to Configure.
18551 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18555 * Extended the ASN.1 parser routines.
18559 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18563 * Added a BN_CTX to the BN library.
18567 * Fixed the weak key values in DES library
18571 * Changed API in EVP library for cipher aliases.
18575 * Added support for RC2/64bit cipher.
18579 * Converted the lhash library to the crypto/mem.c functions.
18583 * Added more recognized ASN.1 object ids.
18587 * Added more RSA padding checks for SSL/TLS.
18591 * Added BIO proxy/filter functionality.
18595 * Added extra_certs to SSL_CTX which can be used
18596 send extra CA certificates to the client in the CA cert chain sending
18597 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18601 * Now Fortezza is denied in the authentication phase because
18602 this is key exchange mechanism is not supported by SSLeay at all.
18606 * Additional PKCS1 checks.
18610 * Support the string "TLSv1" for all TLS v1 ciphers.
18614 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18615 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18619 * Fixed a few memory leaks.
18623 * Fixed various code and comment typos.
18627 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18628 bytes sent in the client random.
18630 *Edward Bishop <ebishop@spyglass.com>*
18634 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18635 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18636 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18637 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18638 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18639 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18640 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18641 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18642 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18643 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18644 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18645 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18646 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18647 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18648 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18649 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18650 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18651 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18652 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18653 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18654 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18655 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18656 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18657 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18658 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18659 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18660 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18661 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18662 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18663 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18664 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18665 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18666 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18667 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18668 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18669 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18670 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18671 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18672 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18673 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18674 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18675 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18676 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18677 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18678 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18679 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18680 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18681 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18682 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18683 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18684 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18685 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18686 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18687 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18688 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18689 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18690 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18691 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18692 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18693 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18694 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18695 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18696 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18697 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18698 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18699 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18700 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18701 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18702 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18703 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18704 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18705 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18706 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18707 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18708 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18709 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18710 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18711 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18712 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18713 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18714 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18715 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18716 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18717 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18718 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18719 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18720 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18721 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18722 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18723 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18724 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18725 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18726 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18727 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18728 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18729 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18730 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18731 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18732 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18733 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18734 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18735 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18736 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18737 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18738 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18739 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18740 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18741 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18742 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18743 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18744 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18745 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18746 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18747 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18748 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18749 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18750 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18751 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18752 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18753 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18754 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18755 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18756 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18757 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18758 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18759 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18760 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18761 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18762 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18763 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18764 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18765 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18766 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18767 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18768 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18769 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18770 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18771 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18772 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18773 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18774 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18775 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18776 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18777 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18778 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18779 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18780 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18781 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18782 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18783 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18784 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18785 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18786 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18787 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18788 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18789 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18790 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18791 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18792 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18793 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18794 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655