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 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
27 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
28 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
29 OpenSSL 3.0 these are replaced by the more generic functions
30 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
31 The old versions have been converted to deprecated macros that just call the
36 * The security callback, which can be customised by application code, supports
37 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
38 in the "other" parameter. In most places this is what is passed. All these
39 places occur server side. However there was one client side call of this
40 security operation and it passed a DH object instead. This is incorrect
41 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
42 of the other locations. Therefore this client side call has been changed to
43 pass an EVP_PKEY instead.
47 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
48 interface. Their functionality remains unchanged.
52 * Added new option for 'openssl list', '-providers', which will display the
53 list of loaded providers, their names, version and status. It optionally
54 displays their gettable parameters.
58 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
59 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
60 type is internally recognised so the workaround is no longer needed.
62 Functionality is still retained as it is, but will only work with
63 EVP_PKEYs with a legacy internal key.
67 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
68 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
72 * Changed all "STACK" functions to be macros instead of inline functions. Macro
73 parameters are still checked for type safety at compile time via helper
78 * Remove the RAND_DRBG API
80 The RAND_DRBG API did not fit well into the new provider concept as
81 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
82 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
83 and some of its API calls are rather low-level. This holds in particular
84 for the callback mechanism (RAND_DRBG_set_callbacks()).
86 Adding a compatibility layer to continue supporting the RAND_DRBG API as
87 a legacy API for a regular deprecation period turned out to come at the
88 price of complicating the new provider API unnecessarily. Since the
89 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
92 *Paul Dale and Matthias St. Pierre*
94 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
95 as well as actual hostnames.
99 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
100 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
101 conversely, silently ignore DTLS protocol version bounds when configuring
102 TLS-based contexts. The commands can be repeated to set bounds of both
103 types. The same applies with the corresponding "min_protocol" and
104 "max_protocol" command-line switches, in case some application uses both TLS
107 SSL_CTX instances that are created for a fixed protocol version (e.g.
108 TLSv1_server_method()) also silently ignore version bounds. Previously
109 attempts to apply bounds to these protocol versions would result in an
110 error. Now only the "version-flexible" SSL_CTX instances are subject to
111 limits in configuration files in command-line options.
115 * Deprecated the `ENGINE` API. Engines should be replaced with providers
120 * Reworked the recorded ERR codes to make better space for system errors.
121 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
122 given code is a system error (true) or an OpenSSL error (false).
126 * Reworked the test perl framework to better allow parallel testing.
128 *Nicola Tuveri and David von Oheimb*
130 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
131 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
135 * 'Configure' has been changed to figure out the configuration target if
136 none is given on the command line. Consequently, the 'config' script is
137 now only a mere wrapper. All documentation is changed to only mention
140 *Rich Salz and Richard Levitte*
142 * Added a library context `OSSL_LIB_CTX` that applications as well as
143 other libraries can use to form a separate context within which
144 libcrypto operations are performed.
146 There are two ways this can be used:
148 - Directly, by passing a library context to functions that take
149 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
151 - Indirectly, by creating a new library context and then assigning
152 it as the new default, with `OSSL_LIB_CTX_set0_default`.
154 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
155 apart from the functions directly related to `OSSL_LIB_CTX`, accept
156 NULL to indicate that the default library context should be used.
158 Library code that changes the default library context using
159 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
160 second call before returning to the caller.
162 _(Note: the library context was initially called `OPENSSL_CTX` and
163 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
167 * Handshake now fails if Extended Master Secret extension is dropped
172 * Dropped interactive mode from the 'openssl' program. From now on,
173 the `openssl` command without arguments is equivalent to `openssl
178 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
179 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
180 While the old function names have been retained for backward compatibility
181 they should not be used in new developments
182 because their return values are confusing: Unlike other `_cmp()` functions
183 they do not return 0 in case their arguments are equal.
187 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
188 EC_GROUP_get_field_type().
192 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
193 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
194 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
195 Applications should rely on the library automatically assigning a suitable
196 EC_METHOD internally upon EC_GROUP construction.
200 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
201 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
202 assigned internally without application intervention.
203 Users of EC_GROUP_new() should switch to a different suitable constructor.
207 * Add CAdES-BES signature verification support, mostly derived
208 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
210 *Filipe Raimundo da Silva*
212 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
216 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
217 functions are not widely used and now OpenSSL automatically perform this
218 conversion when needed.
222 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
223 EC_KEY_precompute_mult(). These functions are not widely used and
224 applications should instead switch to named curves which OpenSSL has
225 hardcoded lookup tables for.
229 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
230 should instead use the L<EC_POINT_mul(3)> function.
234 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
235 that are not applicable to the new provider model. Applications should
236 instead use EVP_default_properties_is_fips_enabled() and
237 EVP_default_properties_enable_fips().
241 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
242 is set, an unexpected EOF is ignored, it pretends a close notify was received
243 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
247 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
248 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
249 used and applications should instead use the
250 L<EC_POINT_set_affine_coordinates(3)> and
251 L<EC_POINT_get_affine_coordinates(3)> functions.
255 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
256 arrays to be more easily constructed via a series of utility functions.
257 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
258 the various push functions and finally convert to a passable OSSL_PARAM
259 array using OSSL_PARAM_BLD_to_param().
263 * The security strength of SHA1 and MD5 based signatures in TLS has been
264 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
265 working at the default security level of 1 and instead requires security
266 level 0. The security level can be changed either using the cipher string
267 with @SECLEVEL, or calling SSL_CTX_set_security_level().
271 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
272 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
273 internal keys, if they correspond to one of those built in types.
277 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
278 contain a provider side internal key.
282 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
283 They are old functions that we don't use, and that you could disable with
284 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
288 * Project text documents not yet having a proper file name extension
289 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
290 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
294 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
295 have been converted to Markdown with the goal to produce documents
296 which not only look pretty when viewed online in the browser, but
297 remain well readable inside a plain text editor.
299 To achieve this goal, a 'minimalistic' Markdown style has been applied
300 which avoids formatting elements that interfere too much with the
301 reading flow in the text file. For example, it
303 * avoids [ATX headings][] and uses [setext headings][] instead
304 (which works for `<h1>` and `<h2>` headings only).
305 * avoids [inline links][] and uses [reference links][] instead.
306 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
308 [ATX headings]: https://github.github.com/gfm/#atx-headings
309 [setext headings]: https://github.github.com/gfm/#setext-headings
310 [inline links]: https://github.github.com/gfm/#inline-link
311 [reference links]: https://github.github.com/gfm/#reference-link
312 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
313 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
315 *Matthias St. Pierre*
317 * The test suite is changed to preserve results of each test recipe.
318 A new directory test-runs/ with subdirectories named like the
319 test recipes are created in the build tree for this purpose.
323 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
324 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
325 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
327 *David von Oheimb, Martin Peylo*
329 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
330 The legacy OCSP-focused and only partly documented API is retained.
331 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
335 * Added `util/check-format.pl`, a tool for checking adherence to the
336 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
337 The checks performed are incomplete and yield some false positives.
338 Still the tool should be useful for detecting most typical glitches.
342 * BIO_do_connect and BIO_do_handshake have been extended:
343 If domain name resolution yields multiple IP addresses all of them are tried
344 after connect() failures.
348 * All of the low level RSA functions have been deprecated including:
350 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
351 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
352 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
353 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
354 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
355 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
356 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
357 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
358 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
359 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
360 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
361 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
362 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
363 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
364 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
365 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
366 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
367 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
368 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
369 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
370 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
371 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
372 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
373 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
374 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
375 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
376 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
377 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
379 Use of these low level functions has been informally discouraged for a long
380 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
381 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
382 L<EVP_PKEY_decrypt(3)>.
386 * X509 certificates signed using SHA1 are no longer allowed at security
388 In TLS/SSL the default security level is 1. It can be set either
389 using the cipher string with @SECLEVEL, or calling
390 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
391 a call to SSL_CTX_use_certificate() will fail if the security level is not
393 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
394 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
399 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
400 modified to use PKEY APIs. These commands are now in maintenance mode
401 and no new features will be added to them.
405 * The command line utility rsautl has been deprecated.
406 Instead use the pkeyutl program.
410 * The command line utilities genrsa and rsa have been modified to use PKEY
411 APIs. They now write PKCS#8 keys by default. These commands are now in
412 maintenance mode and no new features will be added to them.
416 * All of the low level DH functions have been deprecated including:
418 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
419 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
420 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
421 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
422 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
423 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
424 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
425 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
426 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
427 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
428 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
429 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
430 DH_meth_get_generate_params and DH_meth_set_generate_params.
432 Use of these low level functions has been informally discouraged for a long
433 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
434 and L<EVP_PKEY_derive(3)>.
438 * All of the low level DSA functions have been deprecated including:
440 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
441 DSA_get_default_method, DSA_set_method, DSA_get_method,
442 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
443 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
444 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
445 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
446 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
447 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
448 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
449 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
450 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
451 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
452 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
455 Use of these low level functions has been informally discouraged for a long
456 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
457 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
461 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
462 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
463 This means that applications don't have to look at the curve NID and
464 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
465 However, they still can, that EVP_PKEY_set_alias_type() call acts as
466 a no-op when the EVP_PKEY is already of the given type.
468 Parameter and key generation is also reworked to make it possible
469 to generate EVP_PKEY_SM2 parameters and keys without having to go
470 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
471 However, code that does the latter will still work as before.
475 * Deprecated low level ECDH and ECDSA functions. These include:
477 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
478 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
481 Use of these low level functions has been informally discouraged for a long
482 time. Instead applications should use the EVP_PKEY_derive(3),
483 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
487 * Deprecated the EC_KEY_METHOD functions. These include:
489 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
490 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
491 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
492 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
493 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
494 EC_KEY_METHOD_get_verify.
496 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
500 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
501 and EVP_PKEY_decrypt() instead.
502 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
503 and EVP_PKEY_encrypt() instead.
507 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
508 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
509 a new formulation to include all the things it can be used for,
510 as well as words of caution.
514 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
515 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
519 * All of the low level HMAC functions have been deprecated including:
521 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
522 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
525 Use of these low level functions has been informally discouraged for a long
526 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
527 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
528 and L<EVP_MAC_final(3)>.
532 * Over two thousand fixes were made to the documentation, including:
533 - Common options (such as -rand/-writerand, TLS version control, etc)
534 were refactored and point to newly-enhanced descriptions in openssl.pod.
535 - Added style conformance for all options (with help from Richard Levitte),
536 documented all reported missing options, added a CI build to check
537 that all options are documented and that no unimplemented options
539 - Documented some internals, such as all use of environment variables.
540 - Addressed all internal broken L<> references.
544 * All of the low level CMAC functions have been deprecated including:
546 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
547 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
549 Use of these low level functions has been informally discouraged for a long
550 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
551 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
552 and L<EVP_MAC_final(3)>.
556 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
557 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
560 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
561 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
562 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
563 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
564 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
565 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
566 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
567 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
568 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
569 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
571 Use of these low level functions has been informally discouraged
572 for a long time. Applications should use the EVP_DigestInit_ex(3),
573 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
577 * Corrected the documentation of the return values from the `EVP_DigestSign*`
578 set of functions. The documentation mentioned negative values for some
579 errors, but this was never the case, so the mention of negative values
582 Code that followed the documentation and thereby check with something
583 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
587 * All of the low level cipher functions have been deprecated including:
589 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
590 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
591 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
592 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
593 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
594 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
595 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
596 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
597 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
598 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
599 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
600 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
601 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
602 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
603 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
604 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
605 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
606 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
607 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
608 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
609 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
610 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
611 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
612 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
613 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
614 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
615 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
616 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
617 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
619 Use of these low level functions has been informally discouraged for
620 a long time. Applications should use the high level EVP APIs, e.g.
621 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
622 equivalently named decrypt functions instead.
624 *Matt Caswell and Paul Dale*
626 * Removed include/openssl/opensslconf.h.in and replaced it with
627 include/openssl/configuration.h.in, which differs in not including
628 <openssl/macros.h>. A short header include/openssl/opensslconf.h
629 was added to include both.
631 This allows internal hacks where one might need to modify the set
632 of configured macros, for example this if deprecated symbols are
633 still supposed to be available internally:
635 #include <openssl/configuration.h>
637 #undef OPENSSL_NO_DEPRECATED
638 #define OPENSSL_SUPPRESS_DEPRECATED
640 #include <openssl/macros.h>
642 This should not be used by applications that use the exported
643 symbols, as that will lead to linking errors.
647 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
648 used in exponentiation with 512-bit moduli. No EC algorithms are
649 affected. Analysis suggests that attacks against 2-prime RSA1024,
650 3-prime RSA1536, and DSA1024 as a result of this defect would be very
651 difficult to perform and are not believed likely. Attacks against DH512
652 are considered just feasible. However, for an attack the target would
653 have to re-use the DH512 private key, which is not recommended anyway.
654 Also applications directly using the low level API BN_mod_exp may be
655 affected if they use BN_FLG_CONSTTIME.
660 * Most memory-debug features have been deprecated, and the functionality
661 replaced with no-ops.
665 * Added documentation for the STACK API.
669 * Introduced a new method type and API, OSSL_ENCODER, to
670 represent generic encoders. An implementation is expected to
671 be able to encode an object associated with a given name (such
672 as an algorithm name for an asymmetric key) into forms given by
673 implementation properties.
675 Encoders are primarily used from inside libcrypto, through
676 calls to functions like EVP_PKEY_print_private(),
677 PEM_write_bio_PrivateKey() and similar.
679 Encoders are specified in such a way that they can be made to
680 directly handle the provider side portion of an object, if this
681 provider side part comes from the same provider as the encoder
682 itself, but can also be made to handle objects in parametrized
683 form (as an OSSL_PARAM array of data). This allows a provider to
684 offer generic encoders as a service for any other provider.
688 * Added a .pragma directive to the syntax of configuration files, to
689 allow varying behavior in a supported and predictable manner.
690 Currently added pragma:
694 This allows dollar signs to be a keyword character unless it's
695 followed by a opening brace or parenthesis. This is useful for
696 platforms where dollar signs are commonly used in names, such as
697 volume names and system directory names on VMS.
701 * Added functionality to create an EVP_PKEY from user data. This
702 is effectively the same as creating a RSA, DH or DSA object and
703 then assigning them to an EVP_PKEY, but directly using algorithm
704 agnostic EVP functions. A benefit is that this should be future
705 proof for public key algorithms to come.
709 * Change the interpretation of the '--api' configuration option to
710 mean that this is a desired API compatibility level with no
711 further meaning. The previous interpretation, that this would
712 also mean to remove all deprecated symbols up to and including
713 the given version, no requires that 'no-deprecated' is also used
714 in the configuration.
716 When building applications, the desired API compatibility level
717 can be set with the OPENSSL_API_COMPAT macro like before. For
718 API compatibility version below 3.0, the old style numerical
719 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
720 For version 3.0 and on, the value is expected to be the decimal
721 value calculated from the major and minor version like this:
723 MAJOR * 10000 + MINOR * 100
727 -DOPENSSL_API_COMPAT=30000 For 3.0
728 -DOPENSSL_API_COMPAT=30200 For 3.2
730 To hide declarations that are deprecated up to and including the
731 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
732 given when building the application as well.
736 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
737 access to certificate and CRL stores via URIs and OSSL_STORE
740 This adds the following functions:
742 - X509_LOOKUP_store()
743 - X509_STORE_load_file()
744 - X509_STORE_load_path()
745 - X509_STORE_load_store()
746 - SSL_add_store_cert_subjects_to_stack()
747 - SSL_CTX_set_default_verify_store()
748 - SSL_CTX_load_verify_file()
749 - SSL_CTX_load_verify_dir()
750 - SSL_CTX_load_verify_store()
754 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
755 The presence of this system service is determined at run-time.
759 * Added functionality to create an EVP_PKEY context based on data
760 for methods from providers. This takes an algorithm name and a
761 property query string and simply stores them, with the intent
762 that any operation that uses this context will use those strings
763 to fetch the needed methods implicitly, thereby making the port
764 of application written for pre-3.0 OpenSSL easier.
768 * The undocumented function NCONF_WIN32() has been deprecated; for
769 conversion details see the HISTORY section of doc/man5/config.pod
773 * Introduced the new functions EVP_DigestSignInit_ex() and
774 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
775 EVP_DigestVerifyUpdate() have been converted to functions. See the man
776 pages for further details.
780 * Over two thousand fixes were made to the documentation, including:
781 adding missing command flags, better style conformance, documentation
784 *Rich Salz, Richard Levitte*
786 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
787 X25519, X448, Ed25519 and Ed448.
791 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
796 * Deprecated the public definition of ERR_STATE as well as the function
797 ERR_get_state(). This is done in preparation of making ERR_STATE an
802 * Added ERR functionality to give callers access to the stored function
803 names that have replaced the older function code based functions.
805 New functions are ERR_get_error_func(), ERR_peek_error_func(),
806 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
807 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
808 and ERR_peek_last_error_all().
810 These functions have become deprecated: ERR_get_error_line_data(),
811 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
812 ERR_func_error_string().
816 * Extended testing to be verbose for failing tests only. The make variables
817 VERBOSE_FAILURE or VF can be used to enable this:
819 $ make VF=1 test # Unix
820 $ mms /macro=(VF=1) test ! OpenVMS
821 $ nmake VF=1 test # Windows
825 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
826 used even when parsing explicit parameters, when loading a encoded key
827 or calling `EC_GROUP_new_from_ecpkparameters()`/
828 `EC_GROUP_new_from_ecparameters()`.
829 This prevents bypass of security hardening and performance gains,
830 especially for curves with specialized EC_METHODs.
831 By default, if a key encoded with explicit parameters is loaded and later
832 encoded, the output is still encoded with explicit parameters, even if
833 internally a "named" EC_GROUP is used for computation.
837 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
838 this change, EC_GROUP_set_generator would accept order and/or cofactor as
839 NULL. After this change, only the cofactor parameter can be NULL. It also
840 does some minimal sanity checks on the passed order.
845 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
846 An attack is simple, if the first CMS_recipientInfo is valid but the
847 second CMS_recipientInfo is chosen ciphertext. If the second
848 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
849 encryption key will be replaced by garbage, and the message cannot be
850 decoded, but if the RSA decryption fails, the correct encryption key is
851 used and the recipient will not notice the attack.
852 As a work around for this potential attack the length of the decrypted
853 key must be equal to the cipher default key length, in case the
854 certifiate is not given and all recipientInfo are tried out.
855 The old behaviour can be re-enabled in the CMS code by setting the
856 CMS_DEBUG_DECRYPT flag.
860 * Early start up entropy quality from the DEVRANDOM seed source has been
861 improved for older Linux systems. The RAND subsystem will wait for
862 /dev/random to be producing output before seeding from /dev/urandom.
863 The seeded state is stored for future library initialisations using
864 a system global shared memory segment. The shared memory identifier
865 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
866 the desired value. The default identifier is 114.
870 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
871 when primes for RSA keys are computed.
872 Since we previously always generated primes == 2 (mod 3) for RSA keys,
873 the 2-prime and 3-prime RSA modules were easy to distinguish, since
874 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
875 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
876 This avoids possible fingerprinting of newly generated RSA modules.
880 * Correct the extended master secret constant on EBCDIC systems. Without this
881 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
882 negotiate EMS will fail. Unfortunately this also means that TLS connections
883 between EBCDIC systems with this fix, and EBCDIC systems without this
884 fix will fail if they negotiate EMS.
888 * Changed the library initialisation so that the config file is now loaded
889 by default. This was already the case for libssl. It now occurs for both
890 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
891 OPENSSL_init_crypto() to suppress automatic loading of a config file.
895 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
896 where the former acts as a replacement for ERR_put_error(), and the
897 latter replaces the combination ERR_put_error()+ERR_add_error_data().
898 ERR_raise_data() adds more flexibility by taking a format string and
899 an arbitrary number of arguments following it, to be processed with
904 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
905 to check if a named provider is loaded and available. When called, it
906 will also activate all fallback providers if such are still present.
910 * Enforce a minimum DH modulus size of 512 bits.
914 * Changed DH parameters to generate the order q subgroup instead of 2q.
915 Previously generated DH parameters are still accepted by DH_check
916 but DH_generate_key works around that by clearing bit 0 of the
917 private key for those. This avoids leaking bit 0 of the private key.
921 * Significantly reduce secure memory usage by the randomness pools.
925 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
930 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
931 algorithms. An implementation of a key exchange algorithm can be obtained
932 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
933 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
934 the older EVP_PKEY_derive_init() function. See the man pages for the new
935 functions for further details.
939 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
943 * Removed the function names from error messages and deprecated the
946 * Removed NextStep support and the macro OPENSSL_UNISTD
950 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
951 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
952 Also removed "export var as function" capability; we do not export
953 variables, only functions.
957 * RC5_32_set_key has been changed to return an int type, with 0 indicating
958 an error and 1 indicating success. In previous versions of OpenSSL this
959 was a void type. If a key was set longer than the maximum possible this
964 * Support SM2 signing and verification schemes with X509 certificate.
968 * Use SHA256 as the default digest for TS query in the ts app.
972 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
973 This checks that the salt length is at least 128 bits, the derived key
974 length is at least 112 bits, and that the iteration count is at least 1000.
975 For backwards compatibility these checks are disabled by default in the
976 default provider, but are enabled by default in the fips provider.
977 To enable or disable these checks use the control
978 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
982 * Default cipher lists/suites are now available via a function, the
983 #defines are deprecated.
987 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
988 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
989 for Windows Store apps easier. Also, the "no-uplink" option has been added.
993 * Join the directories crypto/x509 and crypto/x509v3
997 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
998 This changes the size when using the genpkey app when no size is given. It
999 fixes an omission in earlier changes that changed all RSA, DSA and DH
1000 generation apps to use 2048 bits by default.
1004 * Added command 'openssl kdf' that uses the EVP_KDF API.
1008 * Added command 'openssl mac' that uses the EVP_MAC API.
1012 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1013 as default directories. Also added the command 'openssl info'
1014 for scripting purposes.
1018 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1019 deprecated. These undocumented functions were never integrated into the EVP
1020 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1021 Bi-directional IGE mode. These modes were never formally standardised and
1022 usage of these functions is believed to be very small. In particular
1023 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1024 is ever used. The security implications are believed to be minimal, but
1025 this issue was never fixed for backwards compatibility reasons. New code
1026 should not use these modes.
1030 * Add prediction resistance to the DRBG reseeding process.
1034 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1035 mandated by IEEE Std 1619-2018.
1039 * Added newline escaping functionality to a filename when using openssl dgst.
1040 This output format is to replicate the output format found in the `*sum`
1041 checksum programs. This aims to preserve backward compatibility.
1043 *Matt Eaton, Richard Levitte, and Paul Dale*
1045 * Removed the heartbeat message in DTLS feature, as it has very
1046 little usage and doesn't seem to fulfill a valuable purpose.
1047 The configuration option is now deprecated.
1051 * Changed the output of 'openssl {digestname} < file' to display the
1052 digest name in its output.
1056 * Added a new generic trace API which provides support for enabling
1057 instrumentation through trace output. This feature is mainly intended
1058 as an aid for developers and is disabled by default. To utilize it,
1059 OpenSSL needs to be configured with the `enable-trace` option.
1061 If the tracing API is enabled, the application can activate trace output
1062 by registering BIOs as trace channels for a number of tracing and debugging
1065 The 'openssl' application has been expanded to enable any of the types
1066 available via environment variables defined by the user, and serves as
1067 one possible example on how to use this functionality.
1069 *Richard Levitte & Matthias St. Pierre*
1071 * Added build tests for C++. These are generated files that only do one
1072 thing, to include one public OpenSSL head file each. This tests that
1073 the public header files can be usefully included in a C++ application.
1075 This test isn't enabled by default. It can be enabled with the option
1076 'enable-buildtest-c++'.
1080 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1084 * Add KMAC to EVP_MAC.
1088 * Added property based algorithm implementation selection framework to
1093 * Added SCA hardening for modular field inversion in EC_GROUP through
1094 a new dedicated field_inv() pointer in EC_METHOD.
1095 This also addresses a leakage affecting conversions from projective
1096 to affine coordinates.
1098 *Billy Bob Brumley, Nicola Tuveri*
1100 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1101 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1102 those algorithms that were already supported through the EVP_PKEY API
1103 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1104 and scrypt are now wrappers that call EVP_KDF.
1108 * Build devcrypto engine as a dynamic engine.
1110 *Eneas U de Queiroz*
1112 * Add keyed BLAKE2 to EVP_MAC.
1116 * Fix a bug in the computation of the endpoint-pair shared secret used
1117 by DTLS over SCTP. This breaks interoperability with older versions
1118 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1119 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1120 interoperability with such broken implementations. However, enabling
1121 this switch breaks interoperability with correct implementations.
1123 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1124 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1128 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1132 * Change the license to the Apache License v2.0.
1136 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1138 - Major releases (indicated by incrementing the MAJOR release number)
1139 may introduce incompatible API/ABI changes.
1140 - Minor releases (indicated by incrementing the MINOR release number)
1141 may introduce new features but retain API/ABI compatibility.
1142 - Patch releases (indicated by incrementing the PATCH number)
1143 are intended for bug fixes and other improvements of existing
1144 features only (like improving performance or adding documentation)
1145 and retain API/ABI compatibility.
1149 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1153 * Remove the 'dist' target and add a tarball building script. The
1154 'dist' target has fallen out of use, and it shouldn't be
1155 necessary to configure just to create a source distribution.
1159 * Recreate the OS390-Unix config target. It no longer relies on a
1160 special script like it did for OpenSSL pre-1.1.0.
1164 * Instead of having the source directories listed in Configure, add
1165 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1170 * Add GMAC to EVP_MAC.
1174 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1178 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1179 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1180 to facilitate the continued use of MACs through raw private keys in
1181 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1185 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1186 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1190 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1191 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1192 are retained for backwards compatibility.
1196 * AES-XTS mode now enforces that its two keys are different to mitigate
1197 the attacked described in "Efficient Instantiations of Tweakable
1198 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1199 Details of this attack can be obtained from:
1200 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1204 * Rename the object files, i.e. give them other names than in previous
1205 versions. Their names now include the name of the final product, as
1206 well as its type mnemonic (bin, lib, shlib).
1210 * Added new option for 'openssl list', '-objects', which will display the
1211 list of built in objects, i.e. OIDs with names.
1215 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1216 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1221 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1222 improves application performance by removing data copies and providing
1223 applications with zero-copy system calls such as sendfile and splice.
1227 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1228 option is set, openssl cleanses (zeroize) plaintext bytes from
1229 internal buffers after delivering them to the application. Note,
1230 the application is still responsible for cleansing other copies
1231 (e.g.: data received by SSL_read(3)).
1235 * `PKCS12_parse` now maintains the order of the parsed certificates
1236 when outputting them via `*ca` (rather than reversing it).
1243 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1247 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1249 * Certificates with explicit curve parameters are now disallowed in
1250 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1254 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1255 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1256 conversely, silently ignore DTLS protocol version bounds when configuring
1257 TLS-based contexts. The commands can be repeated to set bounds of both
1258 types. The same applies with the corresponding "min_protocol" and
1259 "max_protocol" command-line switches, in case some application uses both TLS
1262 SSL_CTX instances that are created for a fixed protocol version (e.g.
1263 TLSv1_server_method()) also silently ignore version bounds. Previously
1264 attempts to apply bounds to these protocol versions would result in an
1265 error. Now only the "version-flexible" SSL_CTX instances are subject to
1266 limits in configuration files in command-line options.
1270 * Handshake now fails if Extended Master Secret extension is dropped
1275 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1277 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1279 * Fixed segmentation fault in SSL_check_chain()
1280 Server or client applications that call the SSL_check_chain() function
1281 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1282 dereference as a result of incorrect handling of the
1283 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1284 or unrecognised signature algorithm is received from the peer. This could
1285 be exploited by a malicious peer in a Denial of Service attack.
1290 * Added AES consttime code for no-asm configurations
1291 an optional constant time support for AES was added
1292 when building openssl for no-asm.
1293 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1294 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1295 At this time this feature is by default disabled.
1296 It will be enabled by default in 3.0.
1300 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1302 * Revert the change of EOF detection while reading in libssl to avoid
1303 regressions in applications depending on the current way of reporting
1304 the EOF. As the existing method is not fully accurate the change to
1305 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1306 branch and will be present in the 3.0 release.
1310 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1311 when primes for RSA keys are computed.
1312 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1313 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1314 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1315 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1316 This avoids possible fingerprinting of newly generated RSA modules.
1320 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1322 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1323 while reading in libssl then we would report an error back to the
1324 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1325 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1326 therefore give a hint as to what went wrong.
1330 * Check that ed25519 and ed448 are allowed by the security level. Previously
1331 signature algorithms not using an MD were not being checked that they were
1332 allowed by the security level.
1336 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1337 was not quite right. The behaviour was not consistent between resumption
1338 and normal handshakes, and also not quite consistent with historical
1339 behaviour. The behaviour in various scenarios has been clarified and
1340 it has been updated to make it match historical behaviour as closely as
1345 * *[VMS only]* The header files that the VMS compilers include automatically,
1346 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1347 that the C++ compiler doesn't understand. This is a shortcoming in the
1348 compiler, but can be worked around with `__cplusplus` guards.
1350 C++ applications that use OpenSSL libraries must be compiled using the
1351 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1352 functions. Otherwise, only functions with symbols of less than 31
1353 characters can be used, as the linker will not be able to successfully
1354 resolve symbols with longer names.
1358 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1359 The presence of this system service is determined at run-time.
1363 * Added newline escaping functionality to a filename when using openssl dgst.
1364 This output format is to replicate the output format found in the `*sum`
1365 checksum programs. This aims to preserve backward compatibility.
1367 *Matt Eaton, Richard Levitte, and Paul Dale*
1369 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1374 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1376 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1377 number generator (RNG). This was intended to include protection in the
1378 event of a fork() system call in order to ensure that the parent and child
1379 processes did not share the same RNG state. However this protection was not
1380 being used in the default case.
1382 A partial mitigation for this issue is that the output from a high
1383 precision timer is mixed into the RNG state so the likelihood of a parent
1384 and child process sharing state is significantly reduced.
1386 If an application already calls OPENSSL_init_crypto() explicitly using
1387 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1390 *Matthias St. Pierre*
1392 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1393 used even when parsing explicit parameters, when loading a encoded key
1394 or calling `EC_GROUP_new_from_ecpkparameters()`/
1395 `EC_GROUP_new_from_ecparameters()`.
1396 This prevents bypass of security hardening and performance gains,
1397 especially for curves with specialized EC_METHODs.
1398 By default, if a key encoded with explicit parameters is loaded and later
1399 encoded, the output is still encoded with explicit parameters, even if
1400 internally a "named" EC_GROUP is used for computation.
1404 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1405 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1406 NULL. After this change, only the cofactor parameter can be NULL. It also
1407 does some minimal sanity checks on the passed order.
1412 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1413 An attack is simple, if the first CMS_recipientInfo is valid but the
1414 second CMS_recipientInfo is chosen ciphertext. If the second
1415 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1416 encryption key will be replaced by garbage, and the message cannot be
1417 decoded, but if the RSA decryption fails, the correct encryption key is
1418 used and the recipient will not notice the attack.
1419 As a work around for this potential attack the length of the decrypted
1420 key must be equal to the cipher default key length, in case the
1421 certifiate is not given and all recipientInfo are tried out.
1422 The old behaviour can be re-enabled in the CMS code by setting the
1423 CMS_DEBUG_DECRYPT flag.
1428 * Early start up entropy quality from the DEVRANDOM seed source has been
1429 improved for older Linux systems. The RAND subsystem will wait for
1430 /dev/random to be producing output before seeding from /dev/urandom.
1431 The seeded state is stored for future library initialisations using
1432 a system global shared memory segment. The shared memory identifier
1433 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1434 the desired value. The default identifier is 114.
1438 * Correct the extended master secret constant on EBCDIC systems. Without this
1439 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1440 negotiate EMS will fail. Unfortunately this also means that TLS connections
1441 between EBCDIC systems with this fix, and EBCDIC systems without this
1442 fix will fail if they negotiate EMS.
1446 * Use Windows installation paths in the mingw builds
1448 Mingw isn't a POSIX environment per se, which means that Windows
1449 paths should be used for installation.
1454 * Changed DH_check to accept parameters with order q and 2q subgroups.
1455 With order 2q subgroups the bit 0 of the private key is not secret
1456 but DH_generate_key works around that by clearing bit 0 of the
1457 private key for those. This avoids leaking bit 0 of the private key.
1461 * Significantly reduce secure memory usage by the randomness pools.
1465 * Revert the DEVRANDOM_WAIT feature for Linux systems
1467 The DEVRANDOM_WAIT feature added a select() call to wait for the
1468 /dev/random device to become readable before reading from the
1469 /dev/urandom device.
1471 It turned out that this change had negative side effects on
1472 performance which were not acceptable. After some discussion it
1473 was decided to revert this feature and leave it up to the OS
1474 resp. the platform maintainer to ensure a proper initialization
1475 during early boot time.
1477 *Matthias St. Pierre*
1479 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1481 * Add build tests for C++. These are generated files that only do one
1482 thing, to include one public OpenSSL head file each. This tests that
1483 the public header files can be usefully included in a C++ application.
1485 This test isn't enabled by default. It can be enabled with the option
1486 'enable-buildtest-c++'.
1490 * Enable SHA3 pre-hashing for ECDSA and DSA.
1494 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1495 This changes the size when using the genpkey app when no size is given. It
1496 fixes an omission in earlier changes that changed all RSA, DSA and DH
1497 generation apps to use 2048 bits by default.
1501 * Reorganize the manual pages to consistently have RETURN VALUES,
1502 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1503 util/fix-doc-nits accordingly.
1505 *Paul Yang, Joshua Lock*
1507 * Add the missing accessor EVP_PKEY_get0_engine()
1511 * Have apps like 's_client' and 's_server' output the signature scheme
1512 along with other cipher suite parameters when debugging.
1516 * Make OPENSSL_config() error agnostic again.
1520 * Do the error handling in RSA decryption constant time.
1524 * Prevent over long nonces in ChaCha20-Poly1305.
1526 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1527 for every encryption operation. RFC 7539 specifies that the nonce value
1528 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1529 and front pads the nonce with 0 bytes if it is less than 12
1530 bytes. However it also incorrectly allows a nonce to be set of up to 16
1531 bytes. In this case only the last 12 bytes are significant and any
1532 additional leading bytes are ignored.
1534 It is a requirement of using this cipher that nonce values are
1535 unique. Messages encrypted using a reused nonce value are susceptible to
1536 serious confidentiality and integrity attacks. If an application changes
1537 the default nonce length to be longer than 12 bytes and then makes a
1538 change to the leading bytes of the nonce expecting the new value to be a
1539 new unique nonce then such an application could inadvertently encrypt
1540 messages with a reused nonce.
1542 Additionally the ignored bytes in a long nonce are not covered by the
1543 integrity guarantee of this cipher. Any application that relies on the
1544 integrity of these ignored leading bytes of a long nonce may be further
1545 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1546 is safe because no such use sets such a long nonce value. However user
1547 applications that use this cipher directly and set a non-default nonce
1548 length to be longer than 12 bytes may be vulnerable.
1550 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1556 * Add DEVRANDOM_WAIT feature for Linux systems
1558 On older Linux systems where the getrandom() system call is not available,
1559 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1560 Contrary to getrandom(), the /dev/urandom device will not block during
1561 early boot when the kernel CSPRNG has not been seeded yet.
1563 To mitigate this known weakness, use select() to wait for /dev/random to
1564 become readable before reading from /dev/urandom.
1566 * Ensure that SM2 only uses SM3 as digest algorithm
1570 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1572 * Change the info callback signals for the start and end of a post-handshake
1573 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1574 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1575 confused by this and assume that a TLSv1.2 renegotiation has started. This
1576 can break KeyUpdate handling. Instead we no longer signal the start and end
1577 of a post handshake message exchange (although the messages themselves are
1578 still signalled). This could break some applications that were expecting
1579 the old signals. However without this KeyUpdate is not usable for many
1584 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1586 * Timing vulnerability in DSA signature generation
1588 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1589 timing side channel attack. An attacker could use variations in the signing
1590 algorithm to recover the private key.
1592 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1597 * Timing vulnerability in ECDSA signature generation
1599 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1600 timing side channel attack. An attacker could use variations in the signing
1601 algorithm to recover the private key.
1603 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1608 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1609 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1610 of two gigabytes and the error handling improved.
1612 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1613 categorized as a normal bug, not a security issue, because the DRBG reseeds
1614 automatically and is fully functional even without additional randomness
1615 provided by the application.
1617 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1619 * Add a new ClientHello callback. Provides a callback interface that gives
1620 the application the ability to adjust the nascent SSL object at the
1621 earliest stage of ClientHello processing, immediately after extensions have
1622 been collected but before they have been processed. In particular, this
1623 callback can adjust the supported TLS versions in response to the contents
1628 * Add SM2 base algorithm support.
1632 * s390x assembly pack: add (improved) hardware-support for the following
1633 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1634 aes-cfb/cfb8, aes-ecb.
1638 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1639 parameter is no longer accepted, as it leads to a corrupt table. NULL
1640 pem_str is reserved for alias entries only.
1644 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1645 step for prime curves. The new implementation is based on formulae from
1646 differential addition-and-doubling in homogeneous projective coordinates
1647 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1648 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1649 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1650 to work in projective coordinates.
1652 *Billy Bob Brumley, Nicola Tuveri*
1654 * Change generating and checking of primes so that the error rate of not
1655 being prime depends on the intended use based on the size of the input.
1656 For larger primes this will result in more rounds of Miller-Rabin.
1657 The maximal error rate for primes with more than 1080 bits is lowered
1660 *Kurt Roeckx, Annie Yousar*
1662 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1666 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1667 moving between systems, and to avoid confusion when a Windows build is
1668 done with mingw vs with MSVC. For POSIX installs, there's still a
1669 symlink or copy named 'tsget' to avoid that confusion as well.
1673 * Revert blinding in ECDSA sign and instead make problematic addition
1674 length-invariant. Switch even to fixed-length Montgomery multiplication.
1678 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1679 step for binary curves. The new implementation is based on formulae from
1680 differential addition-and-doubling in mixed Lopez-Dahab projective
1681 coordinates, modified to independently blind the operands.
1683 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1685 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1686 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1687 EC_METHODs to implement their own specialized "ladder step", to take
1688 advantage of more favorable coordinate systems or more efficient
1689 differential addition-and-doubling algorithms.
1691 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1693 * Modified the random device based seed sources to keep the relevant
1694 file descriptors open rather than reopening them on each access.
1695 This allows such sources to operate in a chroot() jail without
1696 the associated device nodes being available. This behaviour can be
1697 controlled using RAND_keep_random_devices_open().
1701 * Numerous side-channel attack mitigations have been applied. This may have
1702 performance impacts for some algorithms for the benefit of improved
1703 security. Specific changes are noted in this change log by their respective
1708 * AIX shared library support overhaul. Switch to AIX "natural" way of
1709 handling shared libraries, which means collecting shared objects of
1710 different versions and bitnesses in one common archive. This allows to
1711 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1712 doesn't affect the way 3rd party applications are linked, only how
1713 multi-version installation is managed.
1717 * Make ec_group_do_inverse_ord() more robust and available to other
1718 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1719 mitigations are applied to the fallback BN_mod_inverse().
1720 When using this function rather than BN_mod_inverse() directly, new
1721 EC cryptosystem implementations are then safer-by-default.
1725 * Add coordinate blinding for EC_POINT and implement projective
1726 coordinate blinding for generic prime curves as a countermeasure to
1727 chosen point SCA attacks.
1729 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1731 * Add blinding to ECDSA and DSA signatures to protect against side channel
1732 attacks discovered by Keegan Ryan (NCC Group).
1736 * Enforce checking in the pkeyutl command line app to ensure that the input
1737 length does not exceed the maximum supported digest length when performing
1738 a sign, verify or verifyrecover operation.
1742 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1743 I/O in combination with something like select() or poll() will hang. This
1744 can be turned off again using SSL_CTX_clear_mode().
1745 Many applications do not properly handle non-application data records, and
1746 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1747 around the problems in those applications, but can also break some.
1748 It's recommended to read the manpages about SSL_read(), SSL_write(),
1749 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1750 SSL_CTX_set_read_ahead() again.
1754 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1755 now allow empty (zero character) pass phrases.
1759 * Apply blinding to binary field modular inversion and remove patent
1760 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1764 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1765 binary and prime elliptic curves.
1769 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1770 constant time fixed point multiplication.
1774 * Revise elliptic curve scalar multiplication with timing attack
1775 defenses: ec_wNAF_mul redirects to a constant time implementation
1776 when computing fixed point and variable point multiplication (which
1777 in OpenSSL are mostly used with secret scalars in keygen, sign,
1778 ECDH derive operations).
1779 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1782 * Updated CONTRIBUTING
1786 * Updated DRBG / RAND to request nonce and additional low entropy
1787 randomness from the system.
1789 *Matthias St. Pierre*
1791 * Updated 'openssl rehash' to use OpenSSL consistent default.
1795 * Moved the load of the ssl_conf module to libcrypto, which helps
1796 loading engines that libssl uses before libssl is initialised.
1800 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1804 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1806 *Ingo Schwarze, Rich Salz*
1808 * Added output of accepting IP address and port for 'openssl s_server'
1812 * Added a new API for TLSv1.3 ciphersuites:
1813 SSL_CTX_set_ciphersuites()
1814 SSL_set_ciphersuites()
1818 * Memory allocation failures consistently add an error to the error
1823 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1824 in libcrypto when run as setuid/setgid.
1828 * Load any config file by default when libssl is used.
1832 * Added new public header file <openssl/rand_drbg.h> and documentation
1833 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1835 *Matthias St. Pierre*
1837 * QNX support removed (cannot find contributors to get their approval
1838 for the license change).
1842 * TLSv1.3 replay protection for early data has been implemented. See the
1843 SSL_read_early_data() man page for further details.
1847 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1848 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1849 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1850 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1851 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1852 configuration has been separated out. See the ciphers man page or the
1853 SSL_CTX_set_ciphersuites() man page for more information.
1857 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1858 in responder mode now supports the new "-multi" option, which
1859 spawns the specified number of child processes to handle OCSP
1860 requests. The "-timeout" option now also limits the OCSP
1861 responder's patience to wait to receive the full client request
1862 on a newly accepted connection. Child processes are respawned
1863 as needed, and the CA index file is automatically reloaded
1864 when changed. This makes it possible to run the "ocsp" responder
1865 as a long-running service, making the OpenSSL CA somewhat more
1866 feature-complete. In this mode, most diagnostic messages logged
1867 after entering the event loop are logged via syslog(3) rather than
1872 * Added support for X448 and Ed448. Heavily based on original work by
1877 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1878 objects loaded. This adds the functions OSSL_STORE_expect() and
1879 OSSL_STORE_find() as well as needed tools to construct searches and
1880 get the search data out of them.
1884 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1885 version of OpenSSL should review their configuration settings to ensure
1886 that they are still appropriate for TLSv1.3. For further information see:
1887 <https://wiki.openssl.org/index.php/TLS1.3>
1891 * Grand redesign of the OpenSSL random generator
1893 The default RAND method now utilizes an AES-CTR DRBG according to
1894 NIST standard SP 800-90Ar1. The new random generator is essentially
1895 a port of the default random generator from the OpenSSL FIPS 2.0
1896 object module. It is a hybrid deterministic random bit generator
1897 using an AES-CTR bit stream and which seeds and reseeds itself
1898 automatically using trusted system entropy sources.
1900 Some of its new features are:
1901 - Support for multiple DRBG instances with seed chaining.
1902 - The default RAND method makes use of a DRBG.
1903 - There is a public and private DRBG instance.
1904 - The DRBG instances are fork-safe.
1905 - Keep all global DRBG instances on the secure heap if it is enabled.
1906 - The public and private DRBG instance are per thread for lock free
1909 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1911 * Changed Configure so it only says what it does and doesn't dump
1912 so much data. Instead, ./configdata.pm should be used as a script
1913 to display all sorts of configuration data.
1917 * Added processing of "make variables" to Configure.
1921 * Added SHA512/224 and SHA512/256 algorithm support.
1925 * The last traces of Netware support, first removed in 1.1.0, have
1930 * Get rid of Makefile.shared, and in the process, make the processing
1931 of certain files (rc.obj, or the .def/.map/.opt files produced from
1932 the ordinal files) more visible and hopefully easier to trace and
1933 debug (or make silent).
1937 * Make it possible to have environment variable assignments as
1938 arguments to config / Configure.
1942 * Add multi-prime RSA (RFC 8017) support.
1946 * Add SM3 implemented according to GB/T 32905-2016
1947 *Jack Lloyd <jack.lloyd@ribose.com>,*
1948 *Ronald Tse <ronald.tse@ribose.com>,*
1949 *Erick Borsboom <erick.borsboom@ribose.com>*
1951 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1952 as documented in RFC6066.
1953 Based on a patch from Tomasz Moń
1955 *Filipe Raimundo da Silva*
1957 * Add SM4 implemented according to GB/T 32907-2016.
1958 *Jack Lloyd <jack.lloyd@ribose.com>,*
1959 *Ronald Tse <ronald.tse@ribose.com>,*
1960 *Erick Borsboom <erick.borsboom@ribose.com>*
1962 * Reimplement -newreq-nodes and ERR_error_string_n; the
1963 original author does not agree with the license change.
1967 * Add ARIA AEAD TLS support.
1971 * Some macro definitions to support VS6 have been removed. Visual
1972 Studio 6 has not worked since 1.1.0
1976 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1977 without clearing the errors.
1981 * Add "atfork" functions. If building on a system that without
1982 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1983 requirements. The RAND facility now uses/requires this.
1991 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1992 not possible to disable entirely. However, it's still possible to
1993 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1996 To disable, configure with 'no-ui-console'. 'no-ui' is still
1997 possible to use as an alias. Check at compile time with the
1998 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1999 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2003 * Add a STORE module, which implements a uniform and URI based reader of
2004 stores that can contain keys, certificates, CRLs and numerous other
2005 objects. The main API is loosely based on a few stdio functions,
2006 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2007 OSSL_STORE_error and OSSL_STORE_close.
2008 The implementation uses backends called "loaders" to implement arbitrary
2009 URI schemes. There is one built in "loader" for the 'file' scheme.
2013 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2014 then adjusted to work on FreeBSD 8.4 as well.
2015 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2016 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2020 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2021 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2022 error code calls like this:
2024 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2026 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2027 that can be encoded in C. For the foreseeable future, this will only
2030 *Richard Levitte and Tim Hudson*
2032 * Removed BSD cryptodev engine.
2036 * Add a build target 'build_all_generated', to build all generated files
2037 and only that. This can be used to prepare everything that requires
2038 things like perl for a system that lacks perl and then move everything
2039 to that system and do the rest of the build there.
2043 * In the UI interface, make it possible to duplicate the user data. This
2044 can be used by engines that need to retain the data for a longer time
2045 than just the call where this user data is passed.
2049 * Ignore the '-named_curve auto' value for compatibility of applications
2052 *Tomas Mraz <tmraz@fedoraproject.org>*
2054 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2055 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2056 alerts across multiple records (some of which could be empty). In practice
2057 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2058 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2059 support this at all. Supporting it adds significant complexity to the
2060 record layer, and its removal is unlikely to cause interoperability
2065 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2066 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2067 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2072 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2073 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2075 *Richard Levitte, Andy Polyakov*
2077 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2082 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2083 platform rather than 'mingw'.
2087 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2088 success if they are asked to add an object which already exists
2089 in the store. This change cascades to other functions which load
2090 certificates and CRLs.
2094 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2095 facilitate stack unwinding even from assembly subroutines.
2099 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2100 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2104 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2105 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2106 which is the minimum version we support.
2110 * Certificate time validation (X509_cmp_time) enforces stricter
2111 compliance with RFC 5280. Fractional seconds and timezone offsets
2112 are no longer allowed.
2116 * Add support for ARIA
2120 * s_client will now send the Server Name Indication (SNI) extension by
2121 default unless the new "-noservername" option is used. The server name is
2122 based on the host provided to the "-connect" option unless overridden by
2123 using "-servername".
2127 * Add support for SipHash
2131 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2132 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2133 prevent issues where no progress is being made and the peer continually
2134 sends unrecognised record types, using up resources processing them.
2138 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2139 using the algorithm defined in
2140 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2144 * Heartbeat support has been removed; the ABI is changed for now.
2146 *Richard Levitte, Rich Salz*
2148 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2152 * The RSA "null" method, which was partially supported to avoid patent
2153 issues, has been replaced to always returns NULL.
2160 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2162 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2163 used even when parsing explicit parameters, when loading a encoded key
2164 or calling `EC_GROUP_new_from_ecpkparameters()`/
2165 `EC_GROUP_new_from_ecparameters()`.
2166 This prevents bypass of security hardening and performance gains,
2167 especially for curves with specialized EC_METHODs.
2168 By default, if a key encoded with explicit parameters is loaded and later
2169 encoded, the output is still encoded with explicit parameters, even if
2170 internally a "named" EC_GROUP is used for computation.
2174 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2175 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2176 NULL. After this change, only the cofactor parameter can be NULL. It also
2177 does some minimal sanity checks on the passed order.
2182 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2183 An attack is simple, if the first CMS_recipientInfo is valid but the
2184 second CMS_recipientInfo is chosen ciphertext. If the second
2185 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2186 encryption key will be replaced by garbage, and the message cannot be
2187 decoded, but if the RSA decryption fails, the correct encryption key is
2188 used and the recipient will not notice the attack.
2189 As a work around for this potential attack the length of the decrypted
2190 key must be equal to the cipher default key length, in case the
2191 certifiate is not given and all recipientInfo are tried out.
2192 The old behaviour can be re-enabled in the CMS code by setting the
2193 CMS_DEBUG_DECRYPT flag.
2198 * Use Windows installation paths in the mingw builds
2200 Mingw isn't a POSIX environment per se, which means that Windows
2201 paths should be used for installation.
2206 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2208 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2209 This changes the size when using the genpkey app when no size is given. It
2210 fixes an omission in earlier changes that changed all RSA, DSA and DH
2211 generation apps to use 2048 bits by default.
2215 * Prevent over long nonces in ChaCha20-Poly1305.
2217 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2218 for every encryption operation. RFC 7539 specifies that the nonce value
2219 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2220 and front pads the nonce with 0 bytes if it is less than 12
2221 bytes. However it also incorrectly allows a nonce to be set of up to 16
2222 bytes. In this case only the last 12 bytes are significant and any
2223 additional leading bytes are ignored.
2225 It is a requirement of using this cipher that nonce values are
2226 unique. Messages encrypted using a reused nonce value are susceptible to
2227 serious confidentiality and integrity attacks. If an application changes
2228 the default nonce length to be longer than 12 bytes and then makes a
2229 change to the leading bytes of the nonce expecting the new value to be a
2230 new unique nonce then such an application could inadvertently encrypt
2231 messages with a reused nonce.
2233 Additionally the ignored bytes in a long nonce are not covered by the
2234 integrity guarantee of this cipher. Any application that relies on the
2235 integrity of these ignored leading bytes of a long nonce may be further
2236 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2237 is safe because no such use sets such a long nonce value. However user
2238 applications that use this cipher directly and set a non-default nonce
2239 length to be longer than 12 bytes may be vulnerable.
2241 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2247 * Added SCA hardening for modular field inversion in EC_GROUP through
2248 a new dedicated field_inv() pointer in EC_METHOD.
2249 This also addresses a leakage affecting conversions from projective
2250 to affine coordinates.
2252 *Billy Bob Brumley, Nicola Tuveri*
2254 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2255 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2259 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2263 * Remove the 'dist' target and add a tarball building script. The
2264 'dist' target has fallen out of use, and it shouldn't be
2265 necessary to configure just to create a source distribution.
2269 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2271 * Timing vulnerability in DSA signature generation
2273 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2274 timing side channel attack. An attacker could use variations in the signing
2275 algorithm to recover the private key.
2277 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2282 * Timing vulnerability in ECDSA signature generation
2284 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2285 timing side channel attack. An attacker could use variations in the signing
2286 algorithm to recover the private key.
2288 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2293 * Add coordinate blinding for EC_POINT and implement projective
2294 coordinate blinding for generic prime curves as a countermeasure to
2295 chosen point SCA attacks.
2297 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2299 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2301 * Client DoS due to large DH parameter
2303 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2304 malicious server can send a very large prime value to the client. This will
2305 cause the client to spend an unreasonably long period of time generating a
2306 key for this prime resulting in a hang until the client has finished. This
2307 could be exploited in a Denial Of Service attack.
2309 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2314 * Cache timing vulnerability in RSA Key Generation
2316 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2317 a cache timing side channel attack. An attacker with sufficient access to
2318 mount cache timing attacks during the RSA key generation process could
2319 recover the private key.
2321 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2322 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2327 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2328 parameter is no longer accepted, as it leads to a corrupt table. NULL
2329 pem_str is reserved for alias entries only.
2333 * Revert blinding in ECDSA sign and instead make problematic addition
2334 length-invariant. Switch even to fixed-length Montgomery multiplication.
2338 * Change generating and checking of primes so that the error rate of not
2339 being prime depends on the intended use based on the size of the input.
2340 For larger primes this will result in more rounds of Miller-Rabin.
2341 The maximal error rate for primes with more than 1080 bits is lowered
2344 *Kurt Roeckx, Annie Yousar*
2346 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2350 * Add blinding to ECDSA and DSA signatures to protect against side channel
2351 attacks discovered by Keegan Ryan (NCC Group).
2355 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2356 now allow empty (zero character) pass phrases.
2360 * Certificate time validation (X509_cmp_time) enforces stricter
2361 compliance with RFC 5280. Fractional seconds and timezone offsets
2362 are no longer allowed.
2366 * Fixed a text canonicalisation bug in CMS
2368 Where a CMS detached signature is used with text content the text goes
2369 through a canonicalisation process first prior to signing or verifying a
2370 signature. This process strips trailing space at the end of lines, converts
2371 line terminators to CRLF and removes additional trailing line terminators
2372 at the end of a file. A bug in the canonicalisation process meant that
2373 some characters, such as form-feed, were incorrectly treated as whitespace
2374 and removed. This is contrary to the specification (RFC5485). This fix
2375 could mean that detached text data signed with an earlier version of
2376 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2377 signed with a fixed OpenSSL may fail to verify with an earlier version of
2378 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2379 and use the "-binary" flag (for the "cms" command line application) or set
2380 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2384 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2386 * Constructed ASN.1 types with a recursive definition could exceed the stack
2388 Constructed ASN.1 types with a recursive definition (such as can be found
2389 in PKCS7) could eventually exceed the stack given malicious input with
2390 excessive recursion. This could result in a Denial Of Service attack. There
2391 are no such structures used within SSL/TLS that come from untrusted sources
2392 so this is considered safe.
2394 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2400 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2402 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2403 effectively reduced to only comparing the least significant bit of each
2404 byte. This allows an attacker to forge messages that would be considered as
2405 authenticated in an amount of tries lower than that guaranteed by the
2406 security claims of the scheme. The module can only be compiled by the
2407 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2409 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2415 * Add a build target 'build_all_generated', to build all generated files
2416 and only that. This can be used to prepare everything that requires
2417 things like perl for a system that lacks perl and then move everything
2418 to that system and do the rest of the build there.
2422 * Backport SSL_OP_NO_RENGOTIATION
2424 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2425 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2426 changes this is no longer possible in 1.1.0. Therefore the new
2427 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2428 1.1.0 to provide equivalent functionality.
2430 Note that if an application built against 1.1.0h headers (or above) is run
2431 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2432 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2436 * Removed the OS390-Unix config target. It relied on a script that doesn't
2441 * rsaz_1024_mul_avx2 overflow bug on x86_64
2443 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2444 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2445 Analysis suggests that attacks against RSA and DSA as a result of this
2446 defect would be very difficult to perform and are not believed likely.
2447 Attacks against DH1024 are considered just feasible, because most of the
2448 work necessary to deduce information about a private key may be performed
2449 offline. The amount of resources required for such an attack would be
2450 significant. However, for an attack on TLS to be meaningful, the server
2451 would have to share the DH1024 private key among multiple clients, which is
2452 no longer an option since CVE-2016-0701.
2454 This only affects processors that support the AVX2 but not ADX extensions
2455 like Intel Haswell (4th generation).
2457 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2458 was originally found via the OSS-Fuzz project.
2463 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2465 * bn_sqrx8x_internal carry bug on x86_64
2467 There is a carry propagating bug in the x86_64 Montgomery squaring
2468 procedure. No EC algorithms are affected. Analysis suggests that attacks
2469 against RSA and DSA as a result of this defect would be very difficult to
2470 perform and are not believed likely. Attacks against DH are considered just
2471 feasible (although very difficult) because most of the work necessary to
2472 deduce information about a private key may be performed offline. The amount
2473 of resources required for such an attack would be very significant and
2474 likely only accessible to a limited number of attackers. An attacker would
2475 additionally need online access to an unpatched system using the target
2476 private key in a scenario with persistent DH parameters and a private
2477 key that is shared between multiple clients.
2479 This only affects processors that support the BMI1, BMI2 and ADX extensions
2480 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2482 This issue was reported to OpenSSL by the OSS-Fuzz project.
2487 * Malformed X.509 IPAddressFamily could cause OOB read
2489 If an X.509 certificate has a malformed IPAddressFamily extension,
2490 OpenSSL could do a one-byte buffer overread. The most likely result
2491 would be an erroneous display of the certificate in text format.
2493 This issue was reported to OpenSSL by the OSS-Fuzz project.
2498 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2500 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2501 platform rather than 'mingw'.
2505 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2506 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2507 which is the minimum version we support.
2511 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2513 * Encrypt-Then-Mac renegotiation crash
2515 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2516 negotiated where it was not in the original handshake (or vice-versa) then
2517 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2518 and servers are affected.
2520 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2525 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2527 * Truncated packet could crash via OOB read
2529 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2530 cipher is being used, then a truncated packet can cause that host to
2531 perform an out-of-bounds read, usually resulting in a crash.
2533 This issue was reported to OpenSSL by Robert Święcki of Google.
2538 * Bad (EC)DHE parameters cause a client crash
2540 If a malicious server supplies bad parameters for a DHE or ECDHE key
2541 exchange then this can result in the client attempting to dereference a
2542 NULL pointer leading to a client crash. This could be exploited in a Denial
2545 This issue was reported to OpenSSL by Guido Vranken.
2550 * BN_mod_exp may produce incorrect results on x86_64
2552 There is a carry propagating bug in the x86_64 Montgomery squaring
2553 procedure. No EC algorithms are affected. Analysis suggests that attacks
2554 against RSA and DSA as a result of this defect would be very difficult to
2555 perform and are not believed likely. Attacks against DH are considered just
2556 feasible (although very difficult) because most of the work necessary to
2557 deduce information about a private key may be performed offline. The amount
2558 of resources required for such an attack would be very significant and
2559 likely only accessible to a limited number of attackers. An attacker would
2560 additionally need online access to an unpatched system using the target
2561 private key in a scenario with persistent DH parameters and a private
2562 key that is shared between multiple clients. For example this can occur by
2563 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2564 similar to CVE-2015-3193 but must be treated as a separate problem.
2566 This issue was reported to OpenSSL by the OSS-Fuzz project.
2571 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2573 * ChaCha20/Poly1305 heap-buffer-overflow
2575 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2576 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2577 crash. This issue is not considered to be exploitable beyond a DoS.
2579 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2584 * CMS Null dereference
2586 Applications parsing invalid CMS structures can crash with a NULL pointer
2587 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2588 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2589 structure callback if an attempt is made to free certain invalid encodings.
2590 Only CHOICE structures using a callback which do not handle NULL value are
2593 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2598 * Montgomery multiplication may produce incorrect results
2600 There is a carry propagating bug in the Broadwell-specific Montgomery
2601 multiplication procedure that handles input lengths divisible by, but
2602 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2603 and DH private keys are impossible. This is because the subroutine in
2604 question is not used in operations with the private key itself and an input
2605 of the attacker's direct choice. Otherwise the bug can manifest itself as
2606 transient authentication and key negotiation failures or reproducible
2607 erroneous outcome of public-key operations with specially crafted input.
2608 Among EC algorithms only Brainpool P-512 curves are affected and one
2609 presumably can attack ECDH key negotiation. Impact was not analyzed in
2610 detail, because pre-requisites for attack are considered unlikely. Namely
2611 multiple clients have to choose the curve in question and the server has to
2612 share the private key among them, neither of which is default behaviour.
2613 Even then only clients that chose the curve will be affected.
2615 This issue was publicly reported as transient failures and was not
2616 initially recognized as a security issue. Thanks to Richard Morgan for
2617 providing reproducible case.
2622 * Removed automatic addition of RPATH in shared libraries and executables,
2623 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2627 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2629 * Fix Use After Free for large message sizes
2631 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2632 message larger than approx 16k is received then the underlying buffer to
2633 store the incoming message is reallocated and moved. Unfortunately a
2634 dangling pointer to the old location is left which results in an attempt to
2635 write to the previously freed location. This is likely to result in a
2636 crash, however it could potentially lead to execution of arbitrary code.
2638 This issue only affects OpenSSL 1.1.0a.
2640 This issue was reported to OpenSSL by Robert Święcki.
2645 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2647 * OCSP Status Request extension unbounded memory growth
2649 A malicious client can send an excessively large OCSP Status Request
2650 extension. If that client continually requests renegotiation, sending a
2651 large OCSP Status Request extension each time, then there will be unbounded
2652 memory growth on the server. This will eventually lead to a Denial Of
2653 Service attack through memory exhaustion. Servers with a default
2654 configuration are vulnerable even if they do not support OCSP. Builds using
2655 the "no-ocsp" build time option are not affected.
2657 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2662 * SSL_peek() hang on empty record
2664 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2665 sends an empty record. This could be exploited by a malicious peer in a
2666 Denial Of Service attack.
2668 This issue was reported to OpenSSL by Alex Gaynor.
2673 * Excessive allocation of memory in tls_get_message_header() and
2674 dtls1_preprocess_fragment()
2676 A (D)TLS message includes 3 bytes for its length in the header for the
2677 message. This would allow for messages up to 16Mb in length. Messages of
2678 this length are excessive and OpenSSL includes a check to ensure that a
2679 peer is sending reasonably sized messages in order to avoid too much memory
2680 being consumed to service a connection. A flaw in the logic of version
2681 1.1.0 means that memory for the message is allocated too early, prior to
2682 the excessive message length check. Due to way memory is allocated in
2683 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2684 to service a connection. This could lead to a Denial of Service through
2685 memory exhaustion. However, the excessive message length check still takes
2686 place, and this would cause the connection to immediately fail. Assuming
2687 that the application calls SSL_free() on the failed connection in a timely
2688 manner then the 21Mb of allocated memory will then be immediately freed
2689 again. Therefore the excessive memory allocation will be transitory in
2690 nature. This then means that there is only a security impact if:
2692 1) The application does not call SSL_free() in a timely manner in the event
2693 that the connection fails
2695 2) The application is working in a constrained environment where there is
2696 very little free memory
2698 3) The attacker initiates multiple connection attempts such that there are
2699 multiple connections in a state where memory has been allocated for the
2700 connection; SSL_free() has not yet been called; and there is insufficient
2701 memory to service the multiple requests.
2703 Except in the instance of (1) above any Denial Of Service is likely to be
2704 transitory because as soon as the connection fails the memory is
2705 subsequently freed again in the SSL_free() call. However there is an
2706 increased risk during this period of application crashes due to the lack of
2707 memory - which would then mean a more serious Denial of Service.
2709 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2710 (CVE-2016-6307 and CVE-2016-6308)
2714 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2715 had to be removed. Primary reason is that vendor assembler can't
2716 assemble our modules with -KPIC flag. As result it, assembly
2717 support, was not even available as option. But its lack means
2718 lack of side-channel resistant code, which is incompatible with
2719 security by todays standards. Fortunately gcc is readily available
2720 prepackaged option, which we firmly point at...
2724 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2726 * Windows command-line tool supports UTF-8 opt-in option for arguments
2727 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2728 (to any value) allows Windows user to access PKCS#12 file generated
2729 with Windows CryptoAPI and protected with non-ASCII password, as well
2730 as files generated under UTF-8 locale on Linux also protected with
2735 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2736 have been disabled by default and removed from DEFAULT, just like RC4.
2737 See the RC4 item below to re-enable both.
2741 * The method for finding the storage location for the Windows RAND seed file
2742 has changed. First we check %RANDFILE%. If that is not set then we check
2743 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2744 all else fails we fall back to C:\.
2748 * The EVP_EncryptUpdate() function has had its return type changed from void
2749 to int. A return of 0 indicates and error while a return of 1 indicates
2754 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2755 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2756 off the constant time implementation for RSA, DSA and DH have been made
2757 no-ops and deprecated.
2761 * Windows RAND implementation was simplified to only get entropy by
2762 calling CryptGenRandom(). Various other RAND-related tickets
2765 *Joseph Wylie Yandle, Rich Salz*
2767 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2768 and `OPENSSL_LH_`, respectively. The old names are available
2769 with API compatibility. They new names are now completely documented.
2773 * Unify TYPE_up_ref(obj) methods signature.
2774 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2775 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2776 int (instead of void) like all others TYPE_up_ref() methods.
2777 So now these methods also check the return value of CRYPTO_atomic_add(),
2778 and the validity of object reference counter.
2780 *fdasilvayy@gmail.com*
2782 * With Windows Visual Studio builds, the .pdb files are installed
2783 alongside the installed libraries and executables. For a static
2784 library installation, ossl_static.pdb is the associate compiler
2785 generated .pdb file to be used when linking programs.
2789 * Remove openssl.spec. Packaging files belong with the packagers.
2793 * Automatic Darwin/OSX configuration has had a refresh, it will now
2794 recognise x86_64 architectures automatically. You can still decide
2795 to build for a different bitness with the environment variable
2796 KERNEL_BITS (can be 32 or 64), for example:
2798 KERNEL_BITS=32 ./config
2802 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2803 256 bit AES and HMAC with SHA256.
2807 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2811 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2815 * To enable users to have their own config files and build file templates,
2816 Configure looks in the directory indicated by the environment variable
2817 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2818 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2819 name and is used as is.
2823 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2824 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2825 X509_CERT_FILE_CTX was removed.
2829 * "shared" builds are now the default. To create only static libraries use
2830 the "no-shared" Configure option.
2834 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2835 All of these option have not worked for some while and are fundamental
2840 * Make various cleanup routines no-ops and mark them as deprecated. Most
2841 global cleanup functions are no longer required because they are handled
2842 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2843 Explicitly de-initing can cause problems (e.g. where a library that uses
2844 OpenSSL de-inits, but an application is still using it). The affected
2845 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2846 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2847 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2848 COMP_zlib_cleanup().
2852 * --strict-warnings no longer enables runtime debugging options
2853 such as REF_DEBUG. Instead, debug options are automatically
2854 enabled with '--debug' builds.
2856 *Andy Polyakov, Emilia Käsper*
2858 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2859 have been moved out of the public header files. New functions for managing
2860 these have been added.
2864 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2865 objects have been moved out of the public header files. New
2866 functions for managing these have been added.
2870 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2871 have been moved out of the public header files. New functions for managing
2872 these have been added.
2876 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2877 moved out of the public header files. New functions for managing these
2882 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2886 * Removed the mk1mf build scripts.
2890 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2891 it is always safe to #include a header now.
2895 * Removed the aged BC-32 config and all its supporting scripts
2899 * Removed support for Ultrix, Netware, and OS/2.
2903 * Add support for HKDF.
2905 *Alessandro Ghedini*
2907 * Add support for blake2b and blake2s
2911 * Added support for "pipelining". Ciphers that have the
2912 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2913 encryptions/decryptions simultaneously. There are currently no built-in
2914 ciphers with this property but the expectation is that engines will be able
2915 to offer it to significantly improve throughput. Support has been extended
2916 into libssl so that multiple records for a single connection can be
2917 processed in one go (for >=TLS 1.1).
2921 * Added the AFALG engine. This is an async capable engine which is able to
2922 offload work to the Linux kernel. In this initial version it only supports
2923 AES128-CBC. The kernel must be version 4.1.0 or greater.
2927 * OpenSSL now uses a new threading API. It is no longer necessary to
2928 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2929 are two supported threading models: pthreads and windows threads. It is
2930 also possible to configure OpenSSL at compile time for "no-threads". The
2931 old threading API should no longer be used. The functions have been
2932 replaced with "no-op" compatibility macros.
2934 *Alessandro Ghedini, Matt Caswell*
2936 * Modify behavior of ALPN to invoke callback after SNI/servername
2937 callback, such that updates to the SSL_CTX affect ALPN.
2941 * Add SSL_CIPHER queries for authentication and key-exchange.
2945 * Changes to the DEFAULT cipherlist:
2946 - Prefer (EC)DHE handshakes over plain RSA.
2947 - Prefer AEAD ciphers over legacy ciphers.
2948 - Prefer ECDSA over RSA when both certificates are available.
2949 - Prefer TLSv1.2 ciphers/PRF.
2950 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2955 * Change the ECC default curve list to be this, in order: x25519,
2956 secp256r1, secp521r1, secp384r1.
2960 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2961 disabled by default. They can be re-enabled using the
2962 enable-weak-ssl-ciphers option to Configure.
2966 * If the server has ALPN configured, but supports no protocols that the
2967 client advertises, send a fatal "no_application_protocol" alert.
2968 This behaviour is SHALL in RFC 7301, though it isn't universally
2969 implemented by other servers.
2973 * Add X25519 support.
2974 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2975 for public and private key encoding using the format documented in
2976 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2977 key generation and key derivation.
2979 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2984 * Deprecate SRP_VBASE_get_by_user.
2985 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2986 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
2987 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2988 seed, even if the seed is configured.
2990 Users should use SRP_VBASE_get1_by_user instead. Note that in
2991 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2992 also that even though configuring the SRP seed attempts to hide
2993 invalid usernames by continuing the handshake with fake
2994 credentials, this behaviour is not constant time and no strong
2995 guarantees are made that the handshake is indistinguishable from
2996 that of a valid user.
3000 * Configuration change; it's now possible to build dynamic engines
3001 without having to build shared libraries and vice versa. This
3002 only applies to the engines in engines/, those in crypto/engine/
3003 will always be built into libcrypto (i.e. "static").
3005 Building dynamic engines is enabled by default; to disable, use
3006 the configuration option "disable-dynamic-engine".
3008 The only requirements for building dynamic engines are the
3009 presence of the DSO module and building with position independent
3010 code, so they will also automatically be disabled if configuring
3011 with "disable-dso" or "disable-pic".
3013 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3014 are also taken away from openssl/opensslconf.h, as they are
3019 * Configuration change; if there is a known flag to compile
3020 position independent code, it will always be applied on the
3021 libcrypto and libssl object files, and never on the application
3022 object files. This means other libraries that use routines from
3023 libcrypto / libssl can be made into shared libraries regardless
3024 of how OpenSSL was configured.
3026 If this isn't desirable, the configuration options "disable-pic"
3027 or "no-pic" can be used to disable the use of PIC. This will
3028 also disable building shared libraries and dynamic engines.
3032 * Removed JPAKE code. It was experimental and has no wide use.
3036 * The INSTALL_PREFIX Makefile variable has been renamed to
3037 DESTDIR. That makes for less confusion on what this variable
3038 is for. Also, the configuration option --install_prefix is
3043 * Heartbeat for TLS has been removed and is disabled by default
3044 for DTLS; configure with enable-heartbeats. Code that uses the
3045 old #define's might need to be updated.
3047 *Emilia Käsper, Rich Salz*
3049 * Rename REF_CHECK to REF_DEBUG.
3053 * New "unified" build system
3055 The "unified" build system is aimed to be a common system for all
3056 platforms we support. With it comes new support for VMS.
3058 This system builds supports building in a different directory tree
3059 than the source tree. It produces one Makefile (for unix family
3060 or lookalikes), or one descrip.mms (for VMS).
3062 The source of information to make the Makefile / descrip.mms is
3063 small files called 'build.info', holding the necessary
3064 information for each directory with source to compile, and a
3065 template in Configurations, like unix-Makefile.tmpl or
3068 With this change, the library names were also renamed on Windows
3069 and on VMS. They now have names that are closer to the standard
3070 on Unix, and include the major version number, and in certain
3071 cases, the architecture they are built for. See "Notes on shared
3072 libraries" in INSTALL.
3074 We rely heavily on the perl module Text::Template.
3078 * Added support for auto-initialisation and de-initialisation of the library.
3079 OpenSSL no longer requires explicit init or deinit routines to be called,
3080 except in certain circumstances. See the OPENSSL_init_crypto() and
3081 OPENSSL_init_ssl() man pages for further information.
3085 * The arguments to the DTLSv1_listen function have changed. Specifically the
3086 "peer" argument is now expected to be a BIO_ADDR object.
3088 * Rewrite of BIO networking library. The BIO library lacked consistent
3089 support of IPv6, and adding it required some more extensive
3090 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3091 which hold all types of addresses and chains of address information.
3092 It also introduces a new API, with functions like BIO_socket,
3093 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3094 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3095 have been adapted accordingly.
3099 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3104 * CRIME protection: disable compression by default, even if OpenSSL is
3105 compiled with zlib enabled. Applications can still enable compression
3106 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3107 using the SSL_CONF library to configure compression.
3111 * The signature of the session callback configured with
3112 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3113 was explicitly marked as `const unsigned char*` instead of
3118 * Always DPURIFY. Remove the use of uninitialized memory in the
3119 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3123 * Removed many obsolete configuration items, including
3124 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3125 MD2_CHAR, MD2_INT, MD2_LONG
3127 IDEA_SHORT, IDEA_LONG
3128 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3130 *Rich Salz, with advice from Andy Polyakov*
3132 * Many BN internals have been moved to an internal header file.
3134 *Rich Salz with help from Andy Polyakov*
3136 * Configuration and writing out the results from it has changed.
3137 Files such as Makefile include/openssl/opensslconf.h and are now
3138 produced through general templates, such as Makefile.in and
3139 crypto/opensslconf.h.in and some help from the perl module
3142 Also, the center of configuration information is no longer
3143 Makefile. Instead, Configure produces a perl module in
3144 configdata.pm which holds most of the config data (in the hash
3145 table %config), the target data that comes from the target
3146 configuration in one of the `Configurations/*.conf` files (in
3151 * To clarify their intended purposes, the Configure options
3152 --prefix and --openssldir change their semantics, and become more
3153 straightforward and less interdependent.
3155 --prefix shall be used exclusively to give the location INSTALLTOP
3156 where programs, scripts, libraries, include files and manuals are
3157 going to be installed. The default is now /usr/local.
3159 --openssldir shall be used exclusively to give the default
3160 location OPENSSLDIR where certificates, private keys, CRLs are
3161 managed. This is also where the default openssl.cnf gets
3163 If the directory given with this option is a relative path, the
3164 values of both the --prefix value and the --openssldir value will
3165 be combined to become OPENSSLDIR.
3166 The default for --openssldir is INSTALLTOP/ssl.
3168 Anyone who uses --openssldir to specify where OpenSSL is to be
3169 installed MUST change to use --prefix instead.
3173 * The GOST engine was out of date and therefore it has been removed. An up
3174 to date GOST engine is now being maintained in an external repository.
3175 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3176 support for GOST ciphersuites (these are only activated if a GOST engine
3181 * EGD is no longer supported by default; use enable-egd when
3184 *Ben Kaduk and Rich Salz*
3186 * The distribution now has Makefile.in files, which are used to
3187 create Makefile's when Configure is run. *Configure must be run
3188 before trying to build now.*
3192 * The return value for SSL_CIPHER_description() for error conditions
3197 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3199 Obtaining and performing DNSSEC validation of TLSA records is
3200 the application's responsibility. The application provides
3201 the TLSA records of its choice to OpenSSL, and these are then
3202 used to authenticate the peer.
3204 The TLSA records need not even come from DNS. They can, for
3205 example, be used to implement local end-entity certificate or
3206 trust-anchor "pinning", where the "pin" data takes the form
3207 of TLSA records, which can augment or replace verification
3208 based on the usual WebPKI public certification authorities.
3212 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3213 continues to support deprecated interfaces in default builds.
3214 However, applications are strongly advised to compile their
3215 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3216 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3217 or the 1.1.0 releases.
3219 In environments in which all applications have been ported to
3220 not use any deprecated interfaces OpenSSL's Configure script
3221 should be used with the --api=1.1.0 option to entirely remove
3222 support for the deprecated features from the library and
3223 unconditionally disable them in the installed headers.
3224 Essentially the same effect can be achieved with the "no-deprecated"
3225 argument to Configure, except that this will always restrict
3226 the build to just the latest API, rather than a fixed API
3229 As applications are ported to future revisions of the API,
3230 they should update their compile-time OPENSSL_API_COMPAT define
3231 accordingly, but in most cases should be able to continue to
3232 compile with later releases.
3234 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3235 0x10000000L and 0x00908000L, respectively. However those
3236 versions did not support the OPENSSL_API_COMPAT feature, and
3237 so applications are not typically tested for explicit support
3238 of just the undeprecated features of either release.
3242 * Add support for setting the minimum and maximum supported protocol.
3243 It can bet set via the SSL_set_min_proto_version() and
3244 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3245 MaxProtocol. It's recommended to use the new APIs to disable
3246 protocols instead of disabling individual protocols using
3247 SSL_set_options() or SSL_CONF's Protocol. This change also
3248 removes support for disabling TLS 1.2 in the OpenSSL TLS
3249 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3253 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3257 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3258 and integrates ECDSA and ECDH functionality into EC. Implementations can
3259 now redirect key generation and no longer need to convert to or from
3262 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3263 include the ec.h header file instead.
3267 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3268 ciphers who are no longer supported and drops support the ephemeral RSA key
3269 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3273 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3274 opaque. For HMAC_CTX, the following constructors and destructors
3277 HMAC_CTX *HMAC_CTX_new(void);
3278 void HMAC_CTX_free(HMAC_CTX *ctx);
3280 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3281 destroy such methods has been added. See EVP_MD_meth_new(3) and
3282 EVP_CIPHER_meth_new(3) for documentation.
3285 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3286 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3287 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3288 an already created structure.
3289 2) For consistency with the majority of our object creators and
3290 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3291 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3292 for deprecated builds.
3296 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3297 cryptographic operations to be performed asynchronously as long as an
3298 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3299 further details. Libssl has also had this capability integrated with the
3300 introduction of the new mode SSL_MODE_ASYNC and associated error
3301 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3302 pages. This work was developed in partnership with Intel Corp.
3306 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3307 always enabled now. If you want to disable the support you should
3308 exclude it using the list of supported ciphers. This also means that the
3309 "-no_ecdhe" option has been removed from s_server.
3313 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3314 SSL_{CTX_}set1_curves() which can set a list.
3318 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3319 curve you want to support using SSL_{CTX_}set1_curves().
3323 * State machine rewrite. The state machine code has been significantly
3324 refactored in order to remove much duplication of code and solve issues
3325 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3326 further details). This change does have some associated API changes.
3327 Notably the SSL_state() function has been removed and replaced by
3328 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3329 SSL_set_state() has been removed altogether. The previous handshake states
3330 defined in ssl.h and ssl3.h have also been removed.
3334 * All instances of the string "ssleay" in the public API were replaced
3335 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3336 Some error codes related to internal RSA_eay API's were renamed.
3340 * The demo files in crypto/threads were moved to demo/threads.
3344 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3347 *Matt Caswell, Rich Salz*
3349 * New ASN.1 embed macro.
3351 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3352 structure is not allocated: it is part of the parent. That is instead of
3360 This reduces memory fragmentation and make it impossible to accidentally
3361 set a mandatory field to NULL.
3363 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3364 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3365 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3370 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3374 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3375 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3376 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3377 DES and RC4 ciphersuites.
3381 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3382 This changes the decoding behaviour for some invalid messages,
3383 though the change is mostly in the more lenient direction, and
3384 legacy behaviour is preserved as much as possible.
3388 * Fix no-stdio build.
3389 *David Woodhouse <David.Woodhouse@intel.com> and also*
3390 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3392 * New testing framework
3393 The testing framework has been largely rewritten and is now using
3394 perl and the perl modules Test::Harness and an extended variant of
3395 Test::More called OpenSSL::Test to do its work. All test scripts in
3396 test/ have been rewritten into test recipes, and all direct calls to
3397 executables in test/Makefile have become individual recipes using the
3398 simplified testing OpenSSL::Test::Simple.
3400 For documentation on our testing modules, do:
3402 perldoc test/testlib/OpenSSL/Test/Simple.pm
3403 perldoc test/testlib/OpenSSL/Test.pm
3407 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3408 are used; the latter aborts on memory leaks (usually checked on exit).
3409 Some undocumented "set malloc, etc., hooks" functions were removed
3410 and others were changed. All are now documented.
3414 * In DSA_generate_parameters_ex, if the provided seed is too short,
3417 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3419 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3420 from RFC4279, RFC4785, RFC5487, RFC5489.
3422 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3423 original RSA_PSK patch.
3427 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3428 era flag was never set throughout the codebase (only read). Also removed
3429 SSL3_FLAGS_POP_BUFFER which was only used if
3430 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3434 * Changed the default name options in the "ca", "crl", "req" and "x509"
3435 to be "oneline" instead of "compat".
3439 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3440 not aware of clients that still exhibit this bug, and the workaround
3441 hasn't been working properly for a while.
3445 * The return type of BIO_number_read() and BIO_number_written() as well as
3446 the corresponding num_read and num_write members in the BIO structure has
3447 changed from unsigned long to uint64_t. On platforms where an unsigned
3448 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3453 * Given the pervasive nature of TLS extensions it is inadvisable to run
3454 OpenSSL without support for them. It also means that maintaining
3455 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3456 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3460 * Removed support for the two export grade static DH ciphersuites
3461 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3462 were newly added (along with a number of other static DH ciphersuites) to
3463 1.0.2. However the two export ones have *never* worked since they were
3464 introduced. It seems strange in any case to be adding new export
3465 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3469 * Version negotiation has been rewritten. In particular SSLv23_method(),
3470 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3471 and turned into macros which simply call the new preferred function names
3472 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3473 should use the new names instead. Also as part of this change the ssl23.h
3474 header file has been removed.
3478 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3479 code and the associated standard is no longer considered fit-for-purpose.
3483 * RT2547 was closed. When generating a private key, try to make the
3484 output file readable only by the owner. This behavior change might
3485 be noticeable when interacting with other software.
3487 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3492 * Added HTTP GET support to the ocsp command.
3496 * Changed default digest for the dgst and enc commands from MD5 to
3501 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3505 * Added support for TLS extended master secret from
3506 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3507 initial patch which was a great help during development.
3511 * All libssl internal structures have been removed from the public header
3512 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3513 now redundant). Users should not attempt to access internal structures
3514 directly. Instead they should use the provided API functions.
3518 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3519 Access to deprecated functions can be re-enabled by running config with
3520 "enable-deprecated". In addition applications wishing to use deprecated
3521 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3522 will, by default, disable some transitive includes that previously existed
3523 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3527 * Added support for OCB mode. OpenSSL has been granted a patent license
3528 compatible with the OpenSSL license for use of OCB. Details are available
3529 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3530 for OCB can be removed by calling config with no-ocb.
3534 * SSLv2 support has been removed. It still supports receiving a SSLv2
3535 compatible client hello.
3539 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3540 done while fixing the error code for the key-too-small case.
3542 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3544 * CA.sh has been removed; use CA.pl instead.
3548 * Removed old DES API.
3552 * Remove various unsupported platforms:
3558 Sinix/ReliantUNIX RM400
3563 16-bit platforms such as WIN16
3567 * Clean up OPENSSL_NO_xxx #define's
3568 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3569 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3570 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3571 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3572 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3573 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3574 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3575 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3576 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3577 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3581 * Cleaned up dead code
3582 Remove all but one '#ifdef undef' which is to be looked at.
3586 * Clean up calling of xxx_free routines.
3587 Just like free(), fix most of the xxx_free routines to accept
3588 NULL. Remove the non-null checks from callers. Save much code.
3592 * Add secure heap for storage of private keys (when possible).
3593 Add BIO_s_secmem(), CBIGNUM, etc.
3594 Contributed by Akamai Technologies under our Corporate CLA.
3598 * Experimental support for a new, fast, unbiased prime candidate generator,
3599 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3601 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3603 * New output format NSS in the sess_id command line tool. This allows
3604 exporting the session id and the master key in NSS keylog format.
3606 *Martin Kaiser <martin@kaiser.cx>*
3608 * Harmonize version and its documentation. -f flag is used to display
3611 *mancha <mancha1@zoho.com>*
3613 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3614 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3616 *mancha <mancha1@zoho.com>*
3618 * Fix some double frees. These are not thought to be exploitable.
3620 *mancha <mancha1@zoho.com>*
3622 * A missing bounds check in the handling of the TLS heartbeat extension
3623 can be used to reveal up to 64k of memory to a connected client or
3626 Thanks for Neel Mehta of Google Security for discovering this bug and to
3627 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3628 preparing the fix ([CVE-2014-0160])
3630 *Adam Langley, Bodo Moeller*
3632 * Fix for the attack described in the paper "Recovering OpenSSL
3633 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3634 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3635 <http://eprint.iacr.org/2014/140>
3637 Thanks to Yuval Yarom and Naomi Benger for discovering this
3638 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3640 *Yuval Yarom and Naomi Benger*
3642 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3643 this fixes a limitation in previous versions of OpenSSL.
3647 * Experimental encrypt-then-mac support.
3649 Experimental support for encrypt then mac from
3650 draft-gutmann-tls-encrypt-then-mac-02.txt
3652 To enable it set the appropriate extension number (0x42 for the test
3653 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3655 For non-compliant peers (i.e. just about everything) this should have no
3658 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3662 * Add EVP support for key wrapping algorithms, to avoid problems with
3663 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3664 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3665 algorithms and include tests cases.
3669 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3674 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3675 MGF1 digest and OAEP label.
3679 * Make openssl verify return errors.
3681 *Chris Palmer <palmer@google.com> and Ben Laurie*
3683 * New function ASN1_TIME_diff to calculate the difference between two
3684 ASN1_TIME structures or one structure and the current time.
3688 * Update fips_test_suite to support multiple command line options. New
3689 test to induce all self test errors in sequence and check expected
3694 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3695 sign or verify all in one operation.
3699 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3700 test programs and fips_test_suite. Includes functionality to parse
3701 the minimal script output of fipsalgest.pl directly.
3705 * Add authorisation parameter to FIPS_module_mode_set().
3709 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3713 * Use separate DRBG fields for internal and external flags. New function
3714 FIPS_drbg_health_check() to perform on demand health checking. Add
3715 generation tests to fips_test_suite with reduced health check interval to
3716 demonstrate periodic health checking. Add "nodh" option to
3717 fips_test_suite to skip very slow DH test.
3721 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3726 * More extensive health check for DRBG checking many more failure modes.
3727 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3728 combination: call this in fips_test_suite.
3732 * Add support for canonical generation of DSA parameter 'g'. See
3735 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3736 POST to handle HMAC cases.
3740 * Add functions FIPS_module_version() and FIPS_module_version_text()
3741 to return numerical and string versions of the FIPS module number.
3745 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3746 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3747 outside the validated module in the FIPS capable OpenSSL.
3751 * Minor change to DRBG entropy callback semantics. In some cases
3752 there is no multiple of the block length between min_len and
3753 max_len. Allow the callback to return more than max_len bytes
3754 of entropy but discard any extra: it is the callback's responsibility
3755 to ensure that the extra data discarded does not impact the
3756 requested amount of entropy.
3760 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3761 information in FIPS186-3, SP800-57 and SP800-131A.
3765 * CCM support via EVP. Interface is very similar to GCM case except we
3766 must supply all data in one chunk (i.e. no update, final) and the
3767 message length must be supplied if AAD is used. Add algorithm test
3772 * Initial version of POST overhaul. Add POST callback to allow the status
3773 of POST to be monitored and/or failures induced. Modify fips_test_suite
3774 to use callback. Always run all selftests even if one fails.
3778 * XTS support including algorithm test driver in the fips_gcmtest program.
3779 Note: this does increase the maximum key length from 32 to 64 bytes but
3780 there should be no binary compatibility issues as existing applications
3781 will never use XTS mode.
3785 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3786 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3787 performs algorithm blocking for unapproved PRNG types. Also do not
3788 set PRNG type in FIPS_mode_set(): leave this to the application.
3789 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3790 the standard OpenSSL PRNG: set additional data to a date time vector.
3794 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3795 This shouldn't present any incompatibility problems because applications
3796 shouldn't be using these directly and any that are will need to rethink
3797 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3801 * Extensive self tests and health checking required by SP800-90 DRBG.
3802 Remove strength parameter from FIPS_drbg_instantiate and always
3803 instantiate at maximum supported strength.
3807 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3811 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3815 * New function DH_compute_key_padded() to compute a DH key and pad with
3816 leading zeroes if needed: this complies with SP800-56A et al.
3820 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3821 anything, incomplete, subject to change and largely untested at present.
3825 * Modify fipscanisteronly build option to only build the necessary object
3826 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3830 * Add experimental option FIPSSYMS to give all symbols in
3831 fipscanister.o and FIPS or fips prefix. This will avoid
3832 conflicts with future versions of OpenSSL. Add perl script
3833 util/fipsas.pl to preprocess assembly language source files
3834 and rename any affected symbols.
3838 * Add selftest checks and algorithm block of non-fips algorithms in
3839 FIPS mode. Remove DES2 from selftests.
3843 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3844 return internal method without any ENGINE dependencies. Add new
3845 tiny fips sign and verify functions.
3849 * New build option no-ec2m to disable characteristic 2 code.
3853 * New build option "fipscanisteronly". This only builds fipscanister.o
3854 and (currently) associated fips utilities. Uses the file Makefile.fips
3855 instead of Makefile.org as the prototype.
3859 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3860 Update fips_gcmtest to use IV generator.
3864 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3865 setting output buffer to NULL. The `*Final` function must be
3866 called although it will not retrieve any additional data. The tag
3867 can be set or retrieved with a ctrl. The IV length is by default 12
3868 bytes (96 bits) but can be set to an alternative value. If the IV
3869 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3874 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3875 underlying do_cipher function handles all cipher semantics itself
3876 including padding and finalisation. This is useful if (for example)
3877 an ENGINE cipher handles block padding itself. The behaviour of
3878 do_cipher is subtly changed if this flag is set: the return value
3879 is the number of characters written to the output buffer (zero is
3880 no longer an error code) or a negative error code. Also if the
3881 input buffer is NULL and length 0 finalisation should be performed.
3885 * If a candidate issuer certificate is already part of the constructed
3886 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3890 * Improve forward-security support: add functions
3892 void SSL_CTX_set_not_resumable_session_callback(
3893 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3894 void SSL_set_not_resumable_session_callback(
3895 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3897 for use by SSL/TLS servers; the callback function will be called whenever a
3898 new session is created, and gets to decide whether the session may be
3899 cached to make it resumable (return 0) or not (return 1). (As by the
3900 SSL/TLS protocol specifications, the session_id sent by the server will be
3901 empty to indicate that the session is not resumable; also, the server will
3902 not generate RFC 4507 (RFC 5077) session tickets.)
3904 A simple reasonable callback implementation is to return is_forward_secure.
3905 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3906 by the SSL/TLS server library, indicating whether it can provide forward
3909 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3911 * New -verify_name option in command line utilities to set verification
3916 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3917 Add CMAC pkey methods.
3921 * Experimental renegotiation in s_server -www mode. If the client
3922 browses /reneg connection is renegotiated. If /renegcert it is
3923 renegotiated requesting a certificate.
3927 * Add an "external" session cache for debugging purposes to s_server. This
3928 should help trace issues which normally are only apparent in deployed
3929 multi-process servers.
3933 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3934 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3935 BIO_set_cipher() and some obscure PEM functions were changed so they
3936 can now return an error. The RAND changes required a change to the
3937 RAND_METHOD structure.
3941 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3942 a gcc attribute to warn if the result of a function is ignored. This
3943 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3944 whose return value is often ignored.
3948 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3949 These allow SCTs (signed certificate timestamps) to be requested and
3950 validated when establishing a connection.
3952 *Rob Percival <robpercival@google.com>*
3957 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3959 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3960 used even when parsing explicit parameters, when loading a encoded key
3961 or calling `EC_GROUP_new_from_ecpkparameters()`/
3962 `EC_GROUP_new_from_ecparameters()`.
3963 This prevents bypass of security hardening and performance gains,
3964 especially for curves with specialized EC_METHODs.
3965 By default, if a key encoded with explicit parameters is loaded and later
3966 encoded, the output is still encoded with explicit parameters, even if
3967 internally a "named" EC_GROUP is used for computation.
3971 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3972 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3973 NULL. After this change, only the cofactor parameter can be NULL. It also
3974 does some minimal sanity checks on the passed order.
3979 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3980 An attack is simple, if the first CMS_recipientInfo is valid but the
3981 second CMS_recipientInfo is chosen ciphertext. If the second
3982 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3983 encryption key will be replaced by garbage, and the message cannot be
3984 decoded, but if the RSA decryption fails, the correct encryption key is
3985 used and the recipient will not notice the attack.
3986 As a work around for this potential attack the length of the decrypted
3987 key must be equal to the cipher default key length, in case the
3988 certifiate is not given and all recipientInfo are tried out.
3989 The old behaviour can be re-enabled in the CMS code by setting the
3990 CMS_DEBUG_DECRYPT flag.
3995 * Document issue with installation paths in diverse Windows builds
3997 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3998 binaries and run-time config file.
4003 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4005 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4006 This changes the size when using the genpkey app when no size is given. It
4007 fixes an omission in earlier changes that changed all RSA, DSA and DH
4008 generation apps to use 2048 bits by default.
4012 * Add FIPS support for Android Arm 64-bit
4014 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4015 Module in Version 2.0.10. For some reason, the corresponding target
4016 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4017 built with FIPS support on Android Arm 64-bit. This omission has been
4020 *Matthias St. Pierre*
4022 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4024 * 0-byte record padding oracle
4026 If an application encounters a fatal protocol error and then calls
4027 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4028 then OpenSSL can respond differently to the calling application if a 0 byte
4029 record is received with invalid padding compared to if a 0 byte record is
4030 received with an invalid MAC. If the application then behaves differently
4031 based on that in a way that is detectable to the remote peer, then this
4032 amounts to a padding oracle that could be used to decrypt data.
4034 In order for this to be exploitable "non-stitched" ciphersuites must be in
4035 use. Stitched ciphersuites are optimised implementations of certain
4036 commonly used ciphersuites. Also the application must call SSL_shutdown()
4037 twice even if a protocol error has occurred (applications should not do
4038 this but some do anyway).
4040 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4041 Aviram, with additional investigation by Steven Collison and Andrew
4042 Hourselt. It was reported to OpenSSL on 10th December 2018.
4047 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4051 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4053 * Microarchitecture timing vulnerability in ECC scalar multiplication
4055 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4056 shown to be vulnerable to a microarchitecture timing side channel attack.
4057 An attacker with sufficient access to mount local timing attacks during
4058 ECDSA signature generation could recover the private key.
4060 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4061 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4067 * Timing vulnerability in DSA signature generation
4069 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4070 timing side channel attack. An attacker could use variations in the signing
4071 algorithm to recover the private key.
4073 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4078 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4079 Module, accidentally introduced while backporting security fixes from the
4080 development branch and hindering the use of ECC in FIPS mode.
4084 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4086 * Client DoS due to large DH parameter
4088 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4089 malicious server can send a very large prime value to the client. This will
4090 cause the client to spend an unreasonably long period of time generating a
4091 key for this prime resulting in a hang until the client has finished. This
4092 could be exploited in a Denial Of Service attack.
4094 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4099 * Cache timing vulnerability in RSA Key Generation
4101 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4102 a cache timing side channel attack. An attacker with sufficient access to
4103 mount cache timing attacks during the RSA key generation process could
4104 recover the private key.
4106 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4107 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4112 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4113 parameter is no longer accepted, as it leads to a corrupt table. NULL
4114 pem_str is reserved for alias entries only.
4118 * Revert blinding in ECDSA sign and instead make problematic addition
4119 length-invariant. Switch even to fixed-length Montgomery multiplication.
4123 * Change generating and checking of primes so that the error rate of not
4124 being prime depends on the intended use based on the size of the input.
4125 For larger primes this will result in more rounds of Miller-Rabin.
4126 The maximal error rate for primes with more than 1080 bits is lowered
4129 *Kurt Roeckx, Annie Yousar*
4131 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4135 * Add blinding to ECDSA and DSA signatures to protect against side channel
4136 attacks discovered by Keegan Ryan (NCC Group).
4140 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4141 now allow empty (zero character) pass phrases.
4145 * Certificate time validation (X509_cmp_time) enforces stricter
4146 compliance with RFC 5280. Fractional seconds and timezone offsets
4147 are no longer allowed.
4151 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4153 * Constructed ASN.1 types with a recursive definition could exceed the stack
4155 Constructed ASN.1 types with a recursive definition (such as can be found
4156 in PKCS7) could eventually exceed the stack given malicious input with
4157 excessive recursion. This could result in a Denial Of Service attack. There
4158 are no such structures used within SSL/TLS that come from untrusted sources
4159 so this is considered safe.
4161 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4167 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4169 * Read/write after SSL object in error state
4171 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4172 mechanism. The intent was that if a fatal error occurred during a handshake
4173 then OpenSSL would move into the error state and would immediately fail if
4174 you attempted to continue the handshake. This works as designed for the
4175 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4176 SSL_connect()), however due to a bug it does not work correctly if
4177 SSL_read() or SSL_write() is called directly. In that scenario, if the
4178 handshake fails then a fatal error will be returned in the initial function
4179 call. If SSL_read()/SSL_write() is subsequently called by the application
4180 for the same SSL object then it will succeed and the data is passed without
4181 being decrypted/encrypted directly from the SSL/TLS record layer.
4183 In order to exploit this issue an application bug would have to be present
4184 that resulted in a call to SSL_read()/SSL_write() being issued after having
4185 already received a fatal error.
4187 This issue was reported to OpenSSL by David Benjamin (Google).
4192 * rsaz_1024_mul_avx2 overflow bug on x86_64
4194 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4195 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4196 Analysis suggests that attacks against RSA and DSA as a result of this
4197 defect would be very difficult to perform and are not believed likely.
4198 Attacks against DH1024 are considered just feasible, because most of the
4199 work necessary to deduce information about a private key may be performed
4200 offline. The amount of resources required for such an attack would be
4201 significant. However, for an attack on TLS to be meaningful, the server
4202 would have to share the DH1024 private key among multiple clients, which is
4203 no longer an option since CVE-2016-0701.
4205 This only affects processors that support the AVX2 but not ADX extensions
4206 like Intel Haswell (4th generation).
4208 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4209 was originally found via the OSS-Fuzz project.
4214 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4216 * bn_sqrx8x_internal carry bug on x86_64
4218 There is a carry propagating bug in the x86_64 Montgomery squaring
4219 procedure. No EC algorithms are affected. Analysis suggests that attacks
4220 against RSA and DSA as a result of this defect would be very difficult to
4221 perform and are not believed likely. Attacks against DH are considered just
4222 feasible (although very difficult) because most of the work necessary to
4223 deduce information about a private key may be performed offline. The amount
4224 of resources required for such an attack would be very significant and
4225 likely only accessible to a limited number of attackers. An attacker would
4226 additionally need online access to an unpatched system using the target
4227 private key in a scenario with persistent DH parameters and a private
4228 key that is shared between multiple clients.
4230 This only affects processors that support the BMI1, BMI2 and ADX extensions
4231 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4233 This issue was reported to OpenSSL by the OSS-Fuzz project.
4238 * Malformed X.509 IPAddressFamily could cause OOB read
4240 If an X.509 certificate has a malformed IPAddressFamily extension,
4241 OpenSSL could do a one-byte buffer overread. The most likely result
4242 would be an erroneous display of the certificate in text format.
4244 This issue was reported to OpenSSL by the OSS-Fuzz project.
4249 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4251 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4252 platform rather than 'mingw'.
4256 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4258 * Truncated packet could crash via OOB read
4260 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4261 cipher is being used, then a truncated packet can cause that host to
4262 perform an out-of-bounds read, usually resulting in a crash.
4264 This issue was reported to OpenSSL by Robert Święcki of Google.
4269 * BN_mod_exp may produce incorrect results on x86_64
4271 There is a carry propagating bug in the x86_64 Montgomery squaring
4272 procedure. No EC algorithms are affected. Analysis suggests that attacks
4273 against RSA and DSA as a result of this defect would be very difficult to
4274 perform and are not believed likely. Attacks against DH are considered just
4275 feasible (although very difficult) because most of the work necessary to
4276 deduce information about a private key may be performed offline. The amount
4277 of resources required for such an attack would be very significant and
4278 likely only accessible to a limited number of attackers. An attacker would
4279 additionally need online access to an unpatched system using the target
4280 private key in a scenario with persistent DH parameters and a private
4281 key that is shared between multiple clients. For example this can occur by
4282 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4283 similar to CVE-2015-3193 but must be treated as a separate problem.
4285 This issue was reported to OpenSSL by the OSS-Fuzz project.
4290 * Montgomery multiplication may produce incorrect results
4292 There is a carry propagating bug in the Broadwell-specific Montgomery
4293 multiplication procedure that handles input lengths divisible by, but
4294 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4295 and DH private keys are impossible. This is because the subroutine in
4296 question is not used in operations with the private key itself and an input
4297 of the attacker's direct choice. Otherwise the bug can manifest itself as
4298 transient authentication and key negotiation failures or reproducible
4299 erroneous outcome of public-key operations with specially crafted input.
4300 Among EC algorithms only Brainpool P-512 curves are affected and one
4301 presumably can attack ECDH key negotiation. Impact was not analyzed in
4302 detail, because pre-requisites for attack are considered unlikely. Namely
4303 multiple clients have to choose the curve in question and the server has to
4304 share the private key among them, neither of which is default behaviour.
4305 Even then only clients that chose the curve will be affected.
4307 This issue was publicly reported as transient failures and was not
4308 initially recognized as a security issue. Thanks to Richard Morgan for
4309 providing reproducible case.
4314 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4315 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4316 prevent issues where no progress is being made and the peer continually
4317 sends unrecognised record types, using up resources processing them.
4321 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4323 * Missing CRL sanity check
4325 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4326 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4327 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4329 This issue only affects the OpenSSL 1.0.2i
4334 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4336 * OCSP Status Request extension unbounded memory growth
4338 A malicious client can send an excessively large OCSP Status Request
4339 extension. If that client continually requests renegotiation, sending a
4340 large OCSP Status Request extension each time, then there will be unbounded
4341 memory growth on the server. This will eventually lead to a Denial Of
4342 Service attack through memory exhaustion. Servers with a default
4343 configuration are vulnerable even if they do not support OCSP. Builds using
4344 the "no-ocsp" build time option are not affected.
4346 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4351 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4354 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4360 * OOB write in MDC2_Update()
4362 An overflow can occur in MDC2_Update() either if called directly or
4363 through the EVP_DigestUpdate() function using MDC2. If an attacker
4364 is able to supply very large amounts of input data after a previous
4365 call to EVP_EncryptUpdate() with a partial block then a length check
4366 can overflow resulting in a heap corruption.
4368 The amount of data needed is comparable to SIZE_MAX which is impractical
4371 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4376 * Malformed SHA512 ticket DoS
4378 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4379 DoS attack where a malformed ticket will result in an OOB read which will
4382 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4383 a custom server callback and ticket lookup mechanism.
4385 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4390 * OOB write in BN_bn2dec()
4392 The function BN_bn2dec() does not check the return value of BN_div_word().
4393 This can cause an OOB write if an application uses this function with an
4394 overly large BIGNUM. This could be a problem if an overly large certificate
4395 or CRL is printed out from an untrusted source. TLS is not affected because
4396 record limits will reject an oversized certificate before it is parsed.
4398 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4403 * OOB read in TS_OBJ_print_bio()
4405 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4406 the total length the OID text representation would use and not the amount
4407 of data written. This will result in OOB reads when large OIDs are
4410 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4415 * Pointer arithmetic undefined behaviour
4417 Avoid some undefined pointer arithmetic
4419 A common idiom in the codebase is to check limits in the following manner:
4422 Where "p" points to some malloc'd data of SIZE bytes and
4425 "len" here could be from some externally supplied data (e.g. from a TLS
4428 The rules of C pointer arithmetic are such that "p + len" is only well
4429 defined where len <= SIZE. Therefore the above idiom is actually
4430 undefined behaviour.
4432 For example this could cause problems if some malloc implementation
4433 provides an address for "p" such that "p + len" actually overflows for
4434 values of len that are too big and therefore p + len < limit.
4436 This issue was reported to OpenSSL by Guido Vranken
4441 * Constant time flag not preserved in DSA signing
4443 Operations in the DSA signing algorithm should run in constant time in
4444 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4445 implementation means that a non-constant time codepath is followed for
4446 certain operations. This has been demonstrated through a cache-timing
4447 attack to be sufficient for an attacker to recover the private DSA key.
4449 This issue was reported by César Pereida (Aalto University), Billy Brumley
4450 (Tampere University of Technology), and Yuval Yarom (The University of
4451 Adelaide and NICTA).
4456 * DTLS buffered message DoS
4458 In a DTLS connection where handshake messages are delivered out-of-order
4459 those messages that OpenSSL is not yet ready to process will be buffered
4460 for later use. Under certain circumstances, a flaw in the logic means that
4461 those messages do not get removed from the buffer even though the handshake
4462 has been completed. An attacker could force up to approx. 15 messages to
4463 remain in the buffer when they are no longer required. These messages will
4464 be cleared when the DTLS connection is closed. The default maximum size for
4465 a message is 100k. Therefore the attacker could force an additional 1500k
4466 to be consumed per connection. By opening many simulataneous connections an
4467 attacker could cause a DoS attack through memory exhaustion.
4469 This issue was reported to OpenSSL by Quan Luo.
4474 * DTLS replay protection DoS
4476 A flaw in the DTLS replay attack protection mechanism means that records
4477 that arrive for future epochs update the replay protection "window" before
4478 the MAC for the record has been validated. This could be exploited by an
4479 attacker by sending a record for the next epoch (which does not have to
4480 decrypt or have a valid MAC), with a very large sequence number. This means
4481 that all subsequent legitimate packets are dropped causing a denial of
4482 service for a specific DTLS connection.
4484 This issue was reported to OpenSSL by the OCAP audit team.
4489 * Certificate message OOB reads
4491 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4492 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4493 theoretical DoS risk but this has not been observed in practice on common
4496 The messages affected are client certificate, client certificate request
4497 and server certificate. As a result the attack can only be performed
4498 against a client or a server which enables client authentication.
4500 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4505 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4507 * Prevent padding oracle in AES-NI CBC MAC check
4509 A MITM attacker can use a padding oracle attack to decrypt traffic
4510 when the connection uses an AES CBC cipher and the server support
4513 This issue was introduced as part of the fix for Lucky 13 padding
4514 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4515 constant time by making sure that always the same bytes are read and
4516 compared against either the MAC or padding bytes. But it no longer
4517 checked that there was enough data to have both the MAC and padding
4520 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4525 * Fix EVP_EncodeUpdate overflow
4527 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4528 Base64 encoding of binary data. If an attacker is able to supply very large
4529 amounts of input data then a length check can overflow resulting in a heap
4532 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4533 the `PEM_write_bio*` family of functions. These are mainly used within the
4534 OpenSSL command line applications, so any application which processes data
4535 from an untrusted source and outputs it as a PEM file should be considered
4536 vulnerable to this issue. User applications that call these APIs directly
4537 with large amounts of untrusted data may also be vulnerable.
4539 This issue was reported by Guido Vranken.
4544 * Fix EVP_EncryptUpdate overflow
4546 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4547 is able to supply very large amounts of input data after a previous call to
4548 EVP_EncryptUpdate() with a partial block then a length check can overflow
4549 resulting in a heap corruption. Following an analysis of all OpenSSL
4550 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4551 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4552 the first called function after an EVP_EncryptInit(), and therefore that
4553 specific call must be safe. The second form is where the length passed to
4554 EVP_EncryptUpdate() can be seen from the code to be some small value and
4555 therefore there is no possibility of an overflow. Since all instances are
4556 one of these two forms, it is believed that there can be no overflows in
4557 internal code due to this problem. It should be noted that
4558 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4559 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4560 of these calls have also been analysed too and it is believed there are no
4561 instances in internal usage where an overflow could occur.
4563 This issue was reported by Guido Vranken.
4568 * Prevent ASN.1 BIO excessive memory allocation
4570 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4571 a short invalid encoding can cause allocation of large amounts of memory
4572 potentially consuming excessive resources or exhausting memory.
4574 Any application parsing untrusted data through d2i BIO functions is
4575 affected. The memory based functions such as d2i_X509() are *not* affected.
4576 Since the memory based functions are used by the TLS library, TLS
4577 applications are not affected.
4579 This issue was reported by Brian Carpenter.
4586 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4587 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4588 in arbitrary stack data being returned in the buffer.
4590 This issue was reported by Guido Vranken.
4595 * Modify behavior of ALPN to invoke callback after SNI/servername
4596 callback, such that updates to the SSL_CTX affect ALPN.
4600 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4605 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4606 methods are enabled and ssl2 is disabled the methods return NULL.
4610 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4612 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4613 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4614 provide any "EXPORT" or "LOW" strength ciphers.
4618 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4619 is by default disabled at build-time. Builds that are not configured with
4620 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4621 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4622 will need to explicitly call either of:
4624 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4626 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4628 as appropriate. Even if either of those is used, or the application
4629 explicitly uses the version-specific SSLv2_method() or its client and
4630 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4631 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4632 ciphers, and SSLv2 56-bit DES are no longer available.
4637 * Fix a double-free in DSA code
4639 A double free bug was discovered when OpenSSL parses malformed DSA private
4640 keys and could lead to a DoS attack or memory corruption for applications
4641 that receive DSA private keys from untrusted sources. This scenario is
4644 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4650 * Disable SRP fake user seed to address a server memory leak.
4652 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4654 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4655 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4656 was changed to ignore the "fake user" SRP seed, even if the seed
4659 Users should use SRP_VBASE_get1_by_user instead. Note that in
4660 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4661 also that even though configuring the SRP seed attempts to hide
4662 invalid usernames by continuing the handshake with fake
4663 credentials, this behaviour is not constant time and no strong
4664 guarantees are made that the handshake is indistinguishable from
4665 that of a valid user.
4670 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4672 In the BN_hex2bn function the number of hex digits is calculated using an
4673 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4674 large values of `i` this can result in `bn_expand` not allocating any
4675 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4676 field as NULL leading to a subsequent NULL ptr deref. For very large values
4677 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4678 In this case memory is allocated to the internal BIGNUM data field, but it
4679 is insufficiently sized leading to heap corruption. A similar issue exists
4680 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4681 is ever called by user applications with very large untrusted hex/dec data.
4682 This is anticipated to be a rare occurrence.
4684 All OpenSSL internal usage of these functions use data that is not expected
4685 to be untrusted, e.g. config file data or application command line
4686 arguments. If user developed applications generate config file data based
4687 on untrusted data then it is possible that this could also lead to security
4688 consequences. This is also anticipated to be rare.
4690 This issue was reported to OpenSSL by Guido Vranken.
4695 * Fix memory issues in `BIO_*printf` functions
4697 The internal `fmtstr` function used in processing a "%s" format string in
4698 the `BIO_*printf` functions could overflow while calculating the length of a
4699 string and cause an OOB read when printing very long strings.
4701 Additionally the internal `doapr_outch` function can attempt to write to an
4702 OOB memory location (at an offset from the NULL pointer) in the event of a
4703 memory allocation failure. In 1.0.2 and below this could be caused where
4704 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4705 could be in processing a very long "%s" format string. Memory leaks can
4708 The first issue may mask the second issue dependent on compiler behaviour.
4709 These problems could enable attacks where large amounts of untrusted data
4710 is passed to the `BIO_*printf` functions. If applications use these functions
4711 in this way then they could be vulnerable. OpenSSL itself uses these
4712 functions when printing out human-readable dumps of ASN.1 data. Therefore
4713 applications that print this data could be vulnerable if the data is from
4714 untrusted sources. OpenSSL command line applications could also be
4715 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4716 as command line arguments.
4718 Libssl is not considered directly vulnerable. Additionally certificates etc
4719 received via remote connections via libssl are also unlikely to be able to
4720 trigger these issues because of message size limits enforced within libssl.
4722 This issue was reported to OpenSSL Guido Vranken.
4727 * Side channel attack on modular exponentiation
4729 A side-channel attack was found which makes use of cache-bank conflicts on
4730 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4731 of RSA keys. The ability to exploit this issue is limited as it relies on
4732 an attacker who has control of code in a thread running on the same
4733 hyper-threaded core as the victim thread which is performing decryptions.
4735 This issue was reported to OpenSSL by Yuval Yarom, The University of
4736 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4737 Nadia Heninger, University of Pennsylvania with more information at
4738 <http://cachebleed.info>.
4743 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4744 if no keysize is specified with default_bits. This fixes an
4745 omission in an earlier change that changed all RSA/DSA key generation
4746 apps to use 2048 bits by default.
4750 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4752 * DH small subgroups
4754 Historically OpenSSL only ever generated DH parameters based on "safe"
4755 primes. More recently (in version 1.0.2) support was provided for
4756 generating X9.42 style parameter files such as those required for RFC 5114
4757 support. The primes used in such files may not be "safe". Where an
4758 application is using DH configured with parameters based on primes that are
4759 not "safe" then an attacker could use this fact to find a peer's private
4760 DH exponent. This attack requires that the attacker complete multiple
4761 handshakes in which the peer uses the same private DH exponent. For example
4762 this could be used to discover a TLS server's private DH exponent if it's
4763 reusing the private DH exponent or it's using a static DH ciphersuite.
4765 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4766 TLS. It is not on by default. If the option is not set then the server
4767 reuses the same private DH exponent for the life of the server process and
4768 would be vulnerable to this attack. It is believed that many popular
4769 applications do set this option and would therefore not be at risk.
4771 The fix for this issue adds an additional check where a "q" parameter is
4772 available (as is the case in X9.42 based parameters). This detects the
4773 only known attack, and is the only possible defense for static DH
4774 ciphersuites. This could have some performance impact.
4776 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4777 default and cannot be disabled. This could have some performance impact.
4779 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4784 * SSLv2 doesn't block disabled ciphers
4786 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4787 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4788 been disabled, provided that the SSLv2 protocol was not also disabled via
4791 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4792 and Sebastian Schinzel.
4797 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4799 * BN_mod_exp may produce incorrect results on x86_64
4801 There is a carry propagating bug in the x86_64 Montgomery squaring
4802 procedure. No EC algorithms are affected. Analysis suggests that attacks
4803 against RSA and DSA as a result of this defect would be very difficult to
4804 perform and are not believed likely. Attacks against DH are considered just
4805 feasible (although very difficult) because most of the work necessary to
4806 deduce information about a private key may be performed offline. The amount
4807 of resources required for such an attack would be very significant and
4808 likely only accessible to a limited number of attackers. An attacker would
4809 additionally need online access to an unpatched system using the target
4810 private key in a scenario with persistent DH parameters and a private
4811 key that is shared between multiple clients. For example this can occur by
4812 default in OpenSSL DHE based SSL/TLS ciphersuites.
4814 This issue was reported to OpenSSL by Hanno Böck.
4819 * Certificate verify crash with missing PSS parameter
4821 The signature verification routines will crash with a NULL pointer
4822 dereference if presented with an ASN.1 signature using the RSA PSS
4823 algorithm and absent mask generation function parameter. Since these
4824 routines are used to verify certificate signature algorithms this can be
4825 used to crash any certificate verification operation and exploited in a
4826 DoS attack. Any application which performs certificate verification is
4827 vulnerable including OpenSSL clients and servers which enable client
4830 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4835 * X509_ATTRIBUTE memory leak
4837 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4838 memory. This structure is used by the PKCS#7 and CMS routines so any
4839 application which reads PKCS#7 or CMS data from untrusted sources is
4840 affected. SSL/TLS is not affected.
4842 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4848 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4849 This changes the decoding behaviour for some invalid messages,
4850 though the change is mostly in the more lenient direction, and
4851 legacy behaviour is preserved as much as possible.
4855 * In DSA_generate_parameters_ex, if the provided seed is too short,
4858 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4860 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4862 * Alternate chains certificate forgery
4864 During certificate verification, OpenSSL will attempt to find an
4865 alternative certificate chain if the first attempt to build such a chain
4866 fails. An error in the implementation of this logic can mean that an
4867 attacker could cause certain checks on untrusted certificates to be
4868 bypassed, such as the CA flag, enabling them to use a valid leaf
4869 certificate to act as a CA and "issue" an invalid certificate.
4871 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4876 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4878 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4879 incompatibility in the handling of HMAC. The previous ABI has now been
4884 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4886 * Malformed ECParameters causes infinite loop
4888 When processing an ECParameters structure OpenSSL enters an infinite loop
4889 if the curve specified is over a specially malformed binary polynomial
4892 This can be used to perform denial of service against any
4893 system which processes public keys, certificate requests or
4894 certificates. This includes TLS clients and TLS servers with
4895 client authentication enabled.
4897 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4902 * Exploitable out-of-bounds read in X509_cmp_time
4904 X509_cmp_time does not properly check the length of the ASN1_TIME
4905 string and can read a few bytes out of bounds. In addition,
4906 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4909 An attacker can use this to craft malformed certificates and CRLs of
4910 various sizes and potentially cause a segmentation fault, resulting in
4911 a DoS on applications that verify certificates or CRLs. TLS clients
4912 that verify CRLs are affected. TLS clients and servers with client
4913 authentication enabled may be affected if they use custom verification
4916 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4917 independently by Hanno Böck.
4922 * PKCS7 crash with missing EnvelopedContent
4924 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4925 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4926 with missing content and trigger a NULL pointer dereference on parsing.
4928 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4929 structures from untrusted sources are affected. OpenSSL clients and
4930 servers are not affected.
4932 This issue was reported to OpenSSL by Michal Zalewski (Google).
4937 * CMS verify infinite loop with unknown hash function
4939 When verifying a signedData message the CMS code can enter an infinite loop
4940 if presented with an unknown hash function OID. This can be used to perform
4941 denial of service against any system which verifies signedData messages using
4943 This issue was reported to OpenSSL by Johannes Bauer.
4948 * Race condition handling NewSessionTicket
4950 If a NewSessionTicket is received by a multi-threaded client when attempting to
4951 reuse a previous ticket then a race condition can occur potentially leading to
4952 a double free of the ticket data.
4957 * Only support 256-bit or stronger elliptic curves with the
4958 'ecdh_auto' setting (server) or by default (client). Of supported
4959 curves, prefer P-256 (both).
4963 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4965 * ClientHello sigalgs DoS fix
4967 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4968 invalid signature algorithms extension a NULL pointer dereference will
4969 occur. This can be exploited in a DoS attack against the server.
4971 This issue was was reported to OpenSSL by David Ramos of Stanford
4975 *Stephen Henson and Matt Caswell*
4977 * Multiblock corrupted pointer fix
4979 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4980 feature only applies on 64 bit x86 architecture platforms that support AES
4981 NI instructions. A defect in the implementation of "multiblock" can cause
4982 OpenSSL's internal write buffer to become incorrectly set to NULL when
4983 using non-blocking IO. Typically, when the user application is using a
4984 socket BIO for writing, this will only result in a failed connection.
4985 However if some other BIO is used then it is likely that a segmentation
4986 fault will be triggered, thus enabling a potential DoS attack.
4988 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4993 * Segmentation fault in DTLSv1_listen fix
4995 The DTLSv1_listen function is intended to be stateless and processes the
4996 initial ClientHello from many peers. It is common for user code to loop
4997 over the call to DTLSv1_listen until a valid ClientHello is received with
4998 an associated cookie. A defect in the implementation of DTLSv1_listen means
4999 that state is preserved in the SSL object from one invocation to the next
5000 that can lead to a segmentation fault. Errors processing the initial
5001 ClientHello can trigger this scenario. An example of such an error could be
5002 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5005 This issue was reported to OpenSSL by Per Allansson.
5010 * Segmentation fault in ASN1_TYPE_cmp fix
5012 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5013 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5014 certificate signature algorithm consistency this can be used to crash any
5015 certificate verification operation and exploited in a DoS attack. Any
5016 application which performs certificate verification is vulnerable including
5017 OpenSSL clients and servers which enable client authentication.
5022 * Segmentation fault for invalid PSS parameters fix
5024 The signature verification routines will crash with a NULL pointer
5025 dereference if presented with an ASN.1 signature using the RSA PSS
5026 algorithm and invalid parameters. Since these routines are used to verify
5027 certificate signature algorithms this can be used to crash any
5028 certificate verification operation and exploited in a DoS attack. Any
5029 application which performs certificate verification is vulnerable including
5030 OpenSSL clients and servers which enable client authentication.
5032 This issue was was reported to OpenSSL by Brian Carpenter.
5037 * ASN.1 structure reuse memory corruption fix
5039 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5040 memory corruption via an invalid write. Such reuse is and has been
5041 strongly discouraged and is believed to be rare.
5043 Applications that parse structures containing CHOICE or ANY DEFINED BY
5044 components may be affected. Certificate parsing (d2i_X509 and related
5045 functions) are however not affected. OpenSSL clients and servers are
5051 * PKCS7 NULL pointer dereferences fix
5053 The PKCS#7 parsing code does not handle missing outer ContentInfo
5054 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5055 missing content and trigger a NULL pointer dereference on parsing.
5057 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5058 otherwise parse PKCS#7 structures from untrusted sources are
5059 affected. OpenSSL clients and servers are not affected.
5061 This issue was reported to OpenSSL by Michal Zalewski (Google).
5066 * DoS via reachable assert in SSLv2 servers fix
5068 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5069 servers that both support SSLv2 and enable export cipher suites by sending
5070 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5072 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5073 (OpenSSL development team).
5078 * Empty CKE with client auth and DHE fix
5080 If client auth is used then a server can seg fault in the event of a DHE
5081 ciphersuite being selected and a zero length ClientKeyExchange message
5082 being sent by the client. This could be exploited in a DoS attack.
5087 * Handshake with unseeded PRNG fix
5089 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5090 with an unseeded PRNG. The conditions are:
5091 - The client is on a platform where the PRNG has not been seeded
5092 automatically, and the user has not seeded manually
5093 - A protocol specific client method version has been used (i.e. not
5094 SSL_client_methodv23)
5095 - A ciphersuite is used that does not require additional random data from
5096 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5098 If the handshake succeeds then the client random that has been used will
5099 have been generated from a PRNG with insufficient entropy and therefore the
5100 output may be predictable.
5102 For example using the following command with an unseeded openssl will
5103 succeed on an unpatched platform:
5105 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5110 * Use After Free following d2i_ECPrivatekey error fix
5112 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5113 could cause a use after free condition. This, in turn, could cause a double
5114 free in several private key parsing functions (such as d2i_PrivateKey
5115 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5116 for applications that receive EC private keys from untrusted
5117 sources. This scenario is considered rare.
5119 This issue was discovered by the BoringSSL project and fixed in their
5125 * X509_to_X509_REQ NULL pointer deref fix
5127 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5128 the certificate key is invalid. This function is rarely used in practice.
5130 This issue was discovered by Brian Carpenter.
5135 * Removed the export ciphers from the DEFAULT ciphers
5139 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5141 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5142 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5143 So far those who have to target multiple platforms would compromise
5144 and argue that binary targeting say ARMv5 would still execute on
5145 ARMv8. "Universal" build resolves this compromise by providing
5146 near-optimal performance even on newer platforms.
5150 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5151 (other platforms pending).
5153 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5155 * Add support for the SignedCertificateTimestampList certificate and
5156 OCSP response extensions from RFC6962.
5160 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5161 for corner cases. (Certain input points at infinity could lead to
5162 bogus results, with non-infinity inputs mapped to infinity too.)
5166 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5167 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5168 common cases are optimized and there still is room for further
5169 improvements. Vector Permutation AES for Altivec is also added.
5173 * Add support for little-endian ppc64 Linux target.
5175 *Marcelo Cerri (IBM)*
5177 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5178 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5179 are optimized and there still is room for further improvements.
5180 Both 32- and 64-bit modes are supported.
5182 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5184 * Improved ARMv7 NEON support.
5188 * Support for SPARC Architecture 2011 crypto extensions, first
5189 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5190 SHA256/512, MD5, GHASH and modular exponentiation.
5192 *Andy Polyakov, David Miller*
5194 * Accelerated modular exponentiation for Intel processors, a.k.a.
5197 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5199 * Support for new and upcoming Intel processors, including AVX2,
5200 BMI and SHA ISA extensions. This includes additional "stitched"
5201 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5204 This work was sponsored by Intel Corp.
5208 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5209 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5210 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5214 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5215 this fixes a limitation in previous versions of OpenSSL.
5219 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5220 MGF1 digest and OAEP label.
5224 * Add EVP support for key wrapping algorithms, to avoid problems with
5225 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5226 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5227 algorithms and include tests cases.
5231 * Add functions to allocate and set the fields of an ECDSA_METHOD
5234 *Douglas E. Engert, Steve Henson*
5236 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5237 difference in days and seconds between two tm or ASN1_TIME structures.
5241 * Add -rev test option to s_server to just reverse order of characters
5242 received by client and send back to server. Also prints an abbreviated
5243 summary of the connection parameters.
5247 * New option -brief for s_client and s_server to print out a brief summary
5248 of connection parameters.
5252 * Add callbacks for arbitrary TLS extensions.
5254 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5256 * New option -crl_download in several openssl utilities to download CRLs
5257 from CRLDP extension in certificates.
5261 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5265 * New function X509_CRL_diff to generate a delta CRL from the difference
5266 of two full CRLs. Add support to "crl" utility.
5270 * New functions to set lookup_crls function and to retrieve
5271 X509_STORE from X509_STORE_CTX.
5275 * Print out deprecated issuer and subject unique ID fields in
5280 * Extend OCSP I/O functions so they can be used for simple general purpose
5281 HTTP as well as OCSP. New wrapper function which can be used to download
5282 CRLs using the OCSP API.
5286 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5290 * `SSL_CONF*` functions. These provide a common framework for application
5291 configuration using configuration files or command lines.
5295 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5296 message callback and prints the results. Needs compile time option
5297 "enable-ssl-trace". New options to s_client and s_server to enable
5302 * New ctrl and macro to retrieve supported points extensions.
5303 Print out extension in s_server and s_client.
5307 * New functions to retrieve certificate signature and signature
5312 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5317 * New Suite B modes for TLS code. These use and enforce the requirements
5318 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5319 only use Suite B curves. The Suite B modes can be set by using the
5320 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5324 * New chain verification flags for Suite B levels of security. Check
5325 algorithms are acceptable when flags are set in X509_verify_cert.
5329 * Make tls1_check_chain return a set of flags indicating checks passed
5330 by a certificate chain. Add additional tests to handle client
5331 certificates: checks for matching certificate type and issuer name
5336 * If an attempt is made to use a signature algorithm not in the peer
5337 preference list abort the handshake. If client has no suitable
5338 signature algorithms in response to a certificate request do not
5339 use the certificate.
5343 * If server EC tmp key is not in client preference list abort handshake.
5347 * Add support for certificate stores in CERT structure. This makes it
5348 possible to have different stores per SSL structure or one store in
5349 the parent SSL_CTX. Include distinct stores for certificate chain
5350 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5351 to build and store a certificate chain in CERT structure: returning
5352 an error if the chain cannot be built: this will allow applications
5353 to test if a chain is correctly configured.
5355 Note: if the CERT based stores are not set then the parent SSL_CTX
5356 store is used to retain compatibility with existing behaviour.
5360 * New function ssl_set_client_disabled to set a ciphersuite disabled
5361 mask based on the current session, check mask when sending client
5362 hello and checking the requested ciphersuite.
5366 * New ctrls to retrieve and set certificate types in a certificate
5367 request message. Print out received values in s_client. If certificate
5368 types is not set with custom values set sensible values based on
5369 supported signature algorithms.
5373 * Support for distinct client and server supported signature algorithms.
5377 * Add certificate callback. If set this is called whenever a certificate
5378 is required by client or server. An application can decide which
5379 certificate chain to present based on arbitrary criteria: for example
5380 supported signature algorithms. Add very simple example to s_server.
5381 This fixes many of the problems and restrictions of the existing client
5382 certificate callback: for example you can now clear an existing
5383 certificate and specify the whole chain.
5387 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5388 the certificate can be used for (if anything). Set valid_flags field
5389 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5390 to have similar checks in it.
5392 Add new "cert_flags" field to CERT structure and include a "strict mode".
5393 This enforces some TLS certificate requirements (such as only permitting
5394 certificate signature algorithms contained in the supported algorithms
5395 extension) which some implementations ignore: this option should be used
5396 with caution as it could cause interoperability issues.
5400 * Update and tidy signature algorithm extension processing. Work out
5401 shared signature algorithms based on preferences and peer algorithms
5402 and print them out in s_client and s_server. Abort handshake if no
5403 shared signature algorithms.
5407 * Add new functions to allow customised supported signature algorithms
5408 for SSL and SSL_CTX structures. Add options to s_client and s_server
5413 * New function SSL_certs_clear() to delete all references to certificates
5414 from an SSL structure. Before this once a certificate had been added
5415 it couldn't be removed.
5419 * Integrate hostname, email address and IP address checking with certificate
5420 verification. New verify options supporting checking in openssl utility.
5424 * Fixes and wildcard matching support to hostname and email checking
5425 functions. Add manual page.
5427 *Florian Weimer (Red Hat Product Security Team)*
5429 * New functions to check a hostname email or IP address against a
5430 certificate. Add options x509 utility to print results of checks against
5435 * Fix OCSP checking.
5437 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5439 * Initial experimental support for explicitly trusted non-root CAs.
5440 OpenSSL still tries to build a complete chain to a root but if an
5441 intermediate CA has a trust setting included that is used. The first
5442 setting is used: whether to trust (e.g., -addtrust option to the x509
5447 * Add -trusted_first option which attempts to find certificates in the
5448 trusted store even if an untrusted chain is also supplied.
5452 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5453 platform support for Linux and Android.
5457 * Support for linux-x32, ILP32 environment in x86_64 framework.
5461 * Experimental multi-implementation support for FIPS capable OpenSSL.
5462 When in FIPS mode the approved implementations are used as normal,
5463 when not in FIPS mode the internal unapproved versions are used instead.
5464 This means that the FIPS capable OpenSSL isn't forced to use the
5465 (often lower performance) FIPS implementations outside FIPS mode.
5469 * Transparently support X9.42 DH parameters when calling
5470 PEM_read_bio_DHparameters. This means existing applications can handle
5471 the new parameter format automatically.
5475 * Initial experimental support for X9.42 DH parameter format: mainly
5476 to support use of 'q' parameter for RFC5114 parameters.
5480 * Add DH parameters from RFC5114 including test data to dhtest.
5484 * Support for automatic EC temporary key parameter selection. If enabled
5485 the most preferred EC parameters are automatically used instead of
5486 hardcoded fixed parameters. Now a server just has to call:
5487 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5488 support ECDH and use the most appropriate parameters.
5492 * Enhance and tidy EC curve and point format TLS extension code. Use
5493 static structures instead of allocation if default values are used.
5494 New ctrls to set curves we wish to support and to retrieve shared curves.
5495 Print out shared curves in s_server. New options to s_server and s_client
5496 to set list of supported curves.
5500 * New ctrls to retrieve supported signature algorithms and
5501 supported curve values as an array of NIDs. Extend openssl utility
5502 to print out received values.
5506 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5507 between NIDs and the more common NIST names such as "P-256". Enhance
5508 ecparam utility and ECC method to recognise the NIST names for curves.
5512 * Enhance SSL/TLS certificate chain handling to support different
5513 chains for each certificate instead of one chain in the parent SSL_CTX.
5517 * Support for fixed DH ciphersuite client authentication: where both
5518 server and client use DH certificates with common parameters.
5522 * Support for fixed DH ciphersuites: those requiring DH server
5527 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5529 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5530 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5531 X509_CINF_get_signature were reverted post internal team review.
5536 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5538 * OCSP Status Request extension unbounded memory growth
5540 A malicious client can send an excessively large OCSP Status Request
5541 extension. If that client continually requests renegotiation, sending a
5542 large OCSP Status Request extension each time, then there will be unbounded
5543 memory growth on the server. This will eventually lead to a Denial Of
5544 Service attack through memory exhaustion. Servers with a default
5545 configuration are vulnerable even if they do not support OCSP. Builds using
5546 the "no-ocsp" build time option are not affected.
5548 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5553 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5556 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5562 * OOB write in MDC2_Update()
5564 An overflow can occur in MDC2_Update() either if called directly or
5565 through the EVP_DigestUpdate() function using MDC2. If an attacker
5566 is able to supply very large amounts of input data after a previous
5567 call to EVP_EncryptUpdate() with a partial block then a length check
5568 can overflow resulting in a heap corruption.
5570 The amount of data needed is comparable to SIZE_MAX which is impractical
5573 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5578 * Malformed SHA512 ticket DoS
5580 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5581 DoS attack where a malformed ticket will result in an OOB read which will
5584 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5585 a custom server callback and ticket lookup mechanism.
5587 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5592 * OOB write in BN_bn2dec()
5594 The function BN_bn2dec() does not check the return value of BN_div_word().
5595 This can cause an OOB write if an application uses this function with an
5596 overly large BIGNUM. This could be a problem if an overly large certificate
5597 or CRL is printed out from an untrusted source. TLS is not affected because
5598 record limits will reject an oversized certificate before it is parsed.
5600 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5605 * OOB read in TS_OBJ_print_bio()
5607 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5608 the total length the OID text representation would use and not the amount
5609 of data written. This will result in OOB reads when large OIDs are
5612 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5617 * Pointer arithmetic undefined behaviour
5619 Avoid some undefined pointer arithmetic
5621 A common idiom in the codebase is to check limits in the following manner:
5624 Where "p" points to some malloc'd data of SIZE bytes and
5627 "len" here could be from some externally supplied data (e.g. from a TLS
5630 The rules of C pointer arithmetic are such that "p + len" is only well
5631 defined where len <= SIZE. Therefore the above idiom is actually
5632 undefined behaviour.
5634 For example this could cause problems if some malloc implementation
5635 provides an address for "p" such that "p + len" actually overflows for
5636 values of len that are too big and therefore p + len < limit.
5638 This issue was reported to OpenSSL by Guido Vranken
5643 * Constant time flag not preserved in DSA signing
5645 Operations in the DSA signing algorithm should run in constant time in
5646 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5647 implementation means that a non-constant time codepath is followed for
5648 certain operations. This has been demonstrated through a cache-timing
5649 attack to be sufficient for an attacker to recover the private DSA key.
5651 This issue was reported by César Pereida (Aalto University), Billy Brumley
5652 (Tampere University of Technology), and Yuval Yarom (The University of
5653 Adelaide and NICTA).
5658 * DTLS buffered message DoS
5660 In a DTLS connection where handshake messages are delivered out-of-order
5661 those messages that OpenSSL is not yet ready to process will be buffered
5662 for later use. Under certain circumstances, a flaw in the logic means that
5663 those messages do not get removed from the buffer even though the handshake
5664 has been completed. An attacker could force up to approx. 15 messages to
5665 remain in the buffer when they are no longer required. These messages will
5666 be cleared when the DTLS connection is closed. The default maximum size for
5667 a message is 100k. Therefore the attacker could force an additional 1500k
5668 to be consumed per connection. By opening many simulataneous connections an
5669 attacker could cause a DoS attack through memory exhaustion.
5671 This issue was reported to OpenSSL by Quan Luo.
5676 * DTLS replay protection DoS
5678 A flaw in the DTLS replay attack protection mechanism means that records
5679 that arrive for future epochs update the replay protection "window" before
5680 the MAC for the record has been validated. This could be exploited by an
5681 attacker by sending a record for the next epoch (which does not have to
5682 decrypt or have a valid MAC), with a very large sequence number. This means
5683 that all subsequent legitimate packets are dropped causing a denial of
5684 service for a specific DTLS connection.
5686 This issue was reported to OpenSSL by the OCAP audit team.
5691 * Certificate message OOB reads
5693 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5694 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5695 theoretical DoS risk but this has not been observed in practice on common
5698 The messages affected are client certificate, client certificate request
5699 and server certificate. As a result the attack can only be performed
5700 against a client or a server which enables client authentication.
5702 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5707 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5709 * Prevent padding oracle in AES-NI CBC MAC check
5711 A MITM attacker can use a padding oracle attack to decrypt traffic
5712 when the connection uses an AES CBC cipher and the server support
5715 This issue was introduced as part of the fix for Lucky 13 padding
5716 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5717 constant time by making sure that always the same bytes are read and
5718 compared against either the MAC or padding bytes. But it no longer
5719 checked that there was enough data to have both the MAC and padding
5722 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5727 * Fix EVP_EncodeUpdate overflow
5729 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5730 Base64 encoding of binary data. If an attacker is able to supply very large
5731 amounts of input data then a length check can overflow resulting in a heap
5734 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5735 the `PEM_write_bio*` family of functions. These are mainly used within the
5736 OpenSSL command line applications, so any application which processes data
5737 from an untrusted source and outputs it as a PEM file should be considered
5738 vulnerable to this issue. User applications that call these APIs directly
5739 with large amounts of untrusted data may also be vulnerable.
5741 This issue was reported by Guido Vranken.
5746 * Fix EVP_EncryptUpdate overflow
5748 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5749 is able to supply very large amounts of input data after a previous call to
5750 EVP_EncryptUpdate() with a partial block then a length check can overflow
5751 resulting in a heap corruption. Following an analysis of all OpenSSL
5752 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5753 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5754 the first called function after an EVP_EncryptInit(), and therefore that
5755 specific call must be safe. The second form is where the length passed to
5756 EVP_EncryptUpdate() can be seen from the code to be some small value and
5757 therefore there is no possibility of an overflow. Since all instances are
5758 one of these two forms, it is believed that there can be no overflows in
5759 internal code due to this problem. It should be noted that
5760 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5761 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5762 of these calls have also been analysed too and it is believed there are no
5763 instances in internal usage where an overflow could occur.
5765 This issue was reported by Guido Vranken.
5770 * Prevent ASN.1 BIO excessive memory allocation
5772 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5773 a short invalid encoding can casuse allocation of large amounts of memory
5774 potentially consuming excessive resources or exhausting memory.
5776 Any application parsing untrusted data through d2i BIO functions is
5777 affected. The memory based functions such as d2i_X509() are *not* affected.
5778 Since the memory based functions are used by the TLS library, TLS
5779 applications are not affected.
5781 This issue was reported by Brian Carpenter.
5788 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5789 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5790 in arbitrary stack data being returned in the buffer.
5792 This issue was reported by Guido Vranken.
5797 * Modify behavior of ALPN to invoke callback after SNI/servername
5798 callback, such that updates to the SSL_CTX affect ALPN.
5802 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5807 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5808 methods are enabled and ssl2 is disabled the methods return NULL.
5812 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5814 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5815 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5816 provide any "EXPORT" or "LOW" strength ciphers.
5820 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5821 is by default disabled at build-time. Builds that are not configured with
5822 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5823 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5824 will need to explicitly call either of:
5826 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5828 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5830 as appropriate. Even if either of those is used, or the application
5831 explicitly uses the version-specific SSLv2_method() or its client and
5832 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5833 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5834 ciphers, and SSLv2 56-bit DES are no longer available.
5839 * Fix a double-free in DSA code
5841 A double free bug was discovered when OpenSSL parses malformed DSA private
5842 keys and could lead to a DoS attack or memory corruption for applications
5843 that receive DSA private keys from untrusted sources. This scenario is
5846 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5852 * Disable SRP fake user seed to address a server memory leak.
5854 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5856 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5857 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5858 was changed to ignore the "fake user" SRP seed, even if the seed
5861 Users should use SRP_VBASE_get1_by_user instead. Note that in
5862 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5863 also that even though configuring the SRP seed attempts to hide
5864 invalid usernames by continuing the handshake with fake
5865 credentials, this behaviour is not constant time and no strong
5866 guarantees are made that the handshake is indistinguishable from
5867 that of a valid user.
5872 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5874 In the BN_hex2bn function the number of hex digits is calculated using an
5875 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5876 large values of `i` this can result in `bn_expand` not allocating any
5877 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5878 field as NULL leading to a subsequent NULL ptr deref. For very large values
5879 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5880 In this case memory is allocated to the internal BIGNUM data field, but it
5881 is insufficiently sized leading to heap corruption. A similar issue exists
5882 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5883 is ever called by user applications with very large untrusted hex/dec data.
5884 This is anticipated to be a rare occurrence.
5886 All OpenSSL internal usage of these functions use data that is not expected
5887 to be untrusted, e.g. config file data or application command line
5888 arguments. If user developed applications generate config file data based
5889 on untrusted data then it is possible that this could also lead to security
5890 consequences. This is also anticipated to be rare.
5892 This issue was reported to OpenSSL by Guido Vranken.
5897 * Fix memory issues in `BIO_*printf` functions
5899 The internal `fmtstr` function used in processing a "%s" format string in
5900 the `BIO_*printf` functions could overflow while calculating the length of a
5901 string and cause an OOB read when printing very long strings.
5903 Additionally the internal `doapr_outch` function can attempt to write to an
5904 OOB memory location (at an offset from the NULL pointer) in the event of a
5905 memory allocation failure. In 1.0.2 and below this could be caused where
5906 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5907 could be in processing a very long "%s" format string. Memory leaks can
5910 The first issue may mask the second issue dependent on compiler behaviour.
5911 These problems could enable attacks where large amounts of untrusted data
5912 is passed to the `BIO_*printf` functions. If applications use these functions
5913 in this way then they could be vulnerable. OpenSSL itself uses these
5914 functions when printing out human-readable dumps of ASN.1 data. Therefore
5915 applications that print this data could be vulnerable if the data is from
5916 untrusted sources. OpenSSL command line applications could also be
5917 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5918 as command line arguments.
5920 Libssl is not considered directly vulnerable. Additionally certificates etc
5921 received via remote connections via libssl are also unlikely to be able to
5922 trigger these issues because of message size limits enforced within libssl.
5924 This issue was reported to OpenSSL Guido Vranken.
5929 * Side channel attack on modular exponentiation
5931 A side-channel attack was found which makes use of cache-bank conflicts on
5932 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5933 of RSA keys. The ability to exploit this issue is limited as it relies on
5934 an attacker who has control of code in a thread running on the same
5935 hyper-threaded core as the victim thread which is performing decryptions.
5937 This issue was reported to OpenSSL by Yuval Yarom, The University of
5938 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5939 Nadia Heninger, University of Pennsylvania with more information at
5940 <http://cachebleed.info>.
5945 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5946 if no keysize is specified with default_bits. This fixes an
5947 omission in an earlier change that changed all RSA/DSA key generation
5948 apps to use 2048 bits by default.
5952 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5954 * Protection for DH small subgroup attacks
5956 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5957 switched on by default and cannot be disabled. This could have some
5962 * SSLv2 doesn't block disabled ciphers
5964 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5965 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5966 been disabled, provided that the SSLv2 protocol was not also disabled via
5969 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5970 and Sebastian Schinzel.
5975 * Reject DH handshakes with parameters shorter than 1024 bits.
5979 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5981 * Certificate verify crash with missing PSS parameter
5983 The signature verification routines will crash with a NULL pointer
5984 dereference if presented with an ASN.1 signature using the RSA PSS
5985 algorithm and absent mask generation function parameter. Since these
5986 routines are used to verify certificate signature algorithms this can be
5987 used to crash any certificate verification operation and exploited in a
5988 DoS attack. Any application which performs certificate verification is
5989 vulnerable including OpenSSL clients and servers which enable client
5992 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5997 * X509_ATTRIBUTE memory leak
5999 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6000 memory. This structure is used by the PKCS#7 and CMS routines so any
6001 application which reads PKCS#7 or CMS data from untrusted sources is
6002 affected. SSL/TLS is not affected.
6004 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6010 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6011 This changes the decoding behaviour for some invalid messages,
6012 though the change is mostly in the more lenient direction, and
6013 legacy behaviour is preserved as much as possible.
6017 * In DSA_generate_parameters_ex, if the provided seed is too short,
6018 use a random seed, as already documented.
6020 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6022 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6024 * Alternate chains certificate forgery
6026 During certificate verfification, OpenSSL will attempt to find an
6027 alternative certificate chain if the first attempt to build such a chain
6028 fails. An error in the implementation of this logic can mean that an
6029 attacker could cause certain checks on untrusted certificates to be
6030 bypassed, such as the CA flag, enabling them to use a valid leaf
6031 certificate to act as a CA and "issue" an invalid certificate.
6033 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6039 * Race condition handling PSK identify hint
6041 If PSK identity hints are received by a multi-threaded client then
6042 the values are wrongly updated in the parent SSL_CTX structure. This can
6043 result in a race condition potentially leading to a double free of the
6049 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6051 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6052 incompatibility in the handling of HMAC. The previous ABI has now been
6055 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6057 * Malformed ECParameters causes infinite loop
6059 When processing an ECParameters structure OpenSSL enters an infinite loop
6060 if the curve specified is over a specially malformed binary polynomial
6063 This can be used to perform denial of service against any
6064 system which processes public keys, certificate requests or
6065 certificates. This includes TLS clients and TLS servers with
6066 client authentication enabled.
6068 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6073 * Exploitable out-of-bounds read in X509_cmp_time
6075 X509_cmp_time does not properly check the length of the ASN1_TIME
6076 string and can read a few bytes out of bounds. In addition,
6077 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6080 An attacker can use this to craft malformed certificates and CRLs of
6081 various sizes and potentially cause a segmentation fault, resulting in
6082 a DoS on applications that verify certificates or CRLs. TLS clients
6083 that verify CRLs are affected. TLS clients and servers with client
6084 authentication enabled may be affected if they use custom verification
6087 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6088 independently by Hanno Böck.
6093 * PKCS7 crash with missing EnvelopedContent
6095 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6096 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6097 with missing content and trigger a NULL pointer dereference on parsing.
6099 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6100 structures from untrusted sources are affected. OpenSSL clients and
6101 servers are not affected.
6103 This issue was reported to OpenSSL by Michal Zalewski (Google).
6108 * CMS verify infinite loop with unknown hash function
6110 When verifying a signedData message the CMS code can enter an infinite loop
6111 if presented with an unknown hash function OID. This can be used to perform
6112 denial of service against any system which verifies signedData messages using
6114 This issue was reported to OpenSSL by Johannes Bauer.
6119 * Race condition handling NewSessionTicket
6121 If a NewSessionTicket is received by a multi-threaded client when attempting to
6122 reuse a previous ticket then a race condition can occur potentially leading to
6123 a double free of the ticket data.
6128 * Reject DH handshakes with parameters shorter than 768 bits.
6130 *Kurt Roeckx and Emilia Kasper*
6132 * dhparam: generate 2048-bit parameters by default.
6134 *Kurt Roeckx and Emilia Kasper*
6136 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6138 * Segmentation fault in ASN1_TYPE_cmp fix
6140 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6141 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6142 certificate signature algorithm consistency this can be used to crash any
6143 certificate verification operation and exploited in a DoS attack. Any
6144 application which performs certificate verification is vulnerable including
6145 OpenSSL clients and servers which enable client authentication.
6150 * ASN.1 structure reuse memory corruption fix
6152 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6153 memory corruption via an invalid write. Such reuse is and has been
6154 strongly discouraged and is believed to be rare.
6156 Applications that parse structures containing CHOICE or ANY DEFINED BY
6157 components may be affected. Certificate parsing (d2i_X509 and related
6158 functions) are however not affected. OpenSSL clients and servers are
6164 * PKCS7 NULL pointer dereferences fix
6166 The PKCS#7 parsing code does not handle missing outer ContentInfo
6167 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6168 missing content and trigger a NULL pointer dereference on parsing.
6170 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6171 otherwise parse PKCS#7 structures from untrusted sources are
6172 affected. OpenSSL clients and servers are not affected.
6174 This issue was reported to OpenSSL by Michal Zalewski (Google).
6179 * DoS via reachable assert in SSLv2 servers fix
6181 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6182 servers that both support SSLv2 and enable export cipher suites by sending
6183 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6185 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6186 (OpenSSL development team).
6191 * Use After Free following d2i_ECPrivatekey error fix
6193 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6194 could cause a use after free condition. This, in turn, could cause a double
6195 free in several private key parsing functions (such as d2i_PrivateKey
6196 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6197 for applications that receive EC private keys from untrusted
6198 sources. This scenario is considered rare.
6200 This issue was discovered by the BoringSSL project and fixed in their
6206 * X509_to_X509_REQ NULL pointer deref fix
6208 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6209 the certificate key is invalid. This function is rarely used in practice.
6211 This issue was discovered by Brian Carpenter.
6216 * Removed the export ciphers from the DEFAULT ciphers
6220 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6222 * Build fixes for the Windows and OpenVMS platforms
6224 *Matt Caswell and Richard Levitte*
6226 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6228 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6229 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6230 dereference. This could lead to a Denial Of Service attack. Thanks to
6231 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6236 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6237 dtls1_buffer_record function under certain conditions. In particular this
6238 could occur if an attacker sent repeated DTLS records with the same
6239 sequence number but for the next epoch. The memory leak could be exploited
6240 by an attacker in a Denial of Service attack through memory exhaustion.
6241 Thanks to Chris Mueller for reporting this issue.
6246 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6247 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6248 method would be set to NULL which could later result in a NULL pointer
6249 dereference. Thanks to Frank Schmirler for reporting this issue.
6254 * Abort handshake if server key exchange message is omitted for ephemeral
6257 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6258 reporting this issue.
6263 * Remove non-export ephemeral RSA code on client and server. This code
6264 violated the TLS standard by allowing the use of temporary RSA keys in
6265 non-export ciphersuites and could be used by a server to effectively
6266 downgrade the RSA key length used to a value smaller than the server
6267 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6268 INRIA or reporting this issue.
6273 * Fixed issue where DH client certificates are accepted without verification.
6274 An OpenSSL server will accept a DH certificate for client authentication
6275 without the certificate verify message. This effectively allows a client to
6276 authenticate without the use of a private key. This only affects servers
6277 which trust a client certificate authority which issues certificates
6278 containing DH keys: these are extremely rare and hardly ever encountered.
6279 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6285 * Ensure that the session ID context of an SSL is updated when its
6286 SSL_CTX is updated via SSL_set_SSL_CTX.
6288 The session ID context is typically set from the parent SSL_CTX,
6289 and can vary with the CTX.
6293 * Fix various certificate fingerprint issues.
6295 By using non-DER or invalid encodings outside the signed portion of a
6296 certificate the fingerprint can be changed without breaking the signature.
6297 Although no details of the signed portion of the certificate can be changed
6298 this can cause problems with some applications: e.g. those using the
6299 certificate fingerprint for blacklists.
6301 1. Reject signatures with non zero unused bits.
6303 If the BIT STRING containing the signature has non zero unused bits reject
6304 the signature. All current signature algorithms require zero unused bits.
6306 2. Check certificate algorithm consistency.
6308 Check the AlgorithmIdentifier inside TBS matches the one in the
6309 certificate signature. NB: this will result in signature failure
6310 errors for some broken certificates.
6312 Thanks to Konrad Kraszewski from Google for reporting this issue.
6314 3. Check DSA/ECDSA signatures use DER.
6316 Re-encode DSA/ECDSA signatures and compare with the original received
6317 signature. Return an error if there is a mismatch.
6319 This will reject various cases including garbage after signature
6320 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6321 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6322 (negative or with leading zeroes).
6324 Further analysis was conducted and fixes were developed by Stephen Henson
6325 of the OpenSSL core team.
6331 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6332 results on some platforms, including x86_64. This bug occurs at random
6333 with a very low probability, and is not known to be exploitable in any
6334 way, though its exact impact is difficult to determine. Thanks to Pieter
6335 Wuille (Blockstream) who reported this issue and also suggested an initial
6336 fix. Further analysis was conducted by the OpenSSL development team and
6337 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6338 the OpenSSL core team.
6343 * Do not resume sessions on the server if the negotiated protocol
6344 version does not match the session's version. Resuming with a different
6345 version, while not strictly forbidden by the RFC, is of questionable
6346 sanity and breaks all known clients.
6348 *David Benjamin, Emilia Käsper*
6350 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6351 early CCS messages during renegotiation. (Note that because
6352 renegotiation is encrypted, this early CCS was not exploitable.)
6356 * Tighten client-side session ticket handling during renegotiation:
6357 ensure that the client only accepts a session ticket if the server sends
6358 the extension anew in the ServerHello. Previously, a TLS client would
6359 reuse the old extension state and thus accept a session ticket if one was
6360 announced in the initial ServerHello.
6362 Similarly, ensure that the client requires a session ticket if one
6363 was advertised in the ServerHello. Previously, a TLS client would
6364 ignore a missing NewSessionTicket message.
6368 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6372 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6373 sends a carefully crafted handshake message, to cause OpenSSL to fail
6374 to free up to 64k of memory causing a memory leak. This could be
6375 exploited in a Denial Of Service attack. This issue affects OpenSSL
6376 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6377 whether SRTP is used or configured. Implementations of OpenSSL that
6378 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6380 The fix was developed by the OpenSSL team.
6385 * Session Ticket Memory Leak.
6387 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6388 integrity of that ticket is first verified. In the event of a session
6389 ticket integrity check failing, OpenSSL will fail to free memory
6390 causing a memory leak. By sending a large number of invalid session
6391 tickets an attacker could exploit this issue in a Denial Of Service
6397 * Build option no-ssl3 is incomplete.
6399 When OpenSSL is configured with "no-ssl3" as a build option, servers
6400 could accept and complete a SSL 3.0 handshake, and clients could be
6401 configured to send them.
6404 *Akamai and the OpenSSL team*
6406 * Add support for TLS_FALLBACK_SCSV.
6407 Client applications doing fallback retries should call
6408 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6411 *Adam Langley, Bodo Moeller*
6413 * Add additional DigestInfo checks.
6415 Re-encode DigestInto in DER and check against the original when
6416 verifying RSA signature: this will reject any improperly encoded
6417 DigestInfo structures.
6419 Note: this is a precautionary measure and no attacks are currently known.
6423 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6425 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6426 SRP code can be overrun an internal buffer. Add sanity check that
6427 g, A, B < N to SRP code.
6429 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6430 Group for discovering this issue.
6435 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6436 TLS 1.0 instead of higher protocol versions when the ClientHello message
6437 is badly fragmented. This allows a man-in-the-middle attacker to force a
6438 downgrade to TLS 1.0 even if both the server and the client support a
6439 higher protocol version, by modifying the client's TLS records.
6441 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6442 researching this issue.
6447 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6448 to a denial of service attack. A malicious server can crash the client
6449 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6450 ciphersuite and sending carefully crafted handshake messages.
6452 Thanks to Felix Gröbert (Google) for discovering and researching this
6458 * By sending carefully crafted DTLS packets an attacker could cause openssl
6459 to leak memory. This can be exploited through a Denial of Service attack.
6460 Thanks to Adam Langley for discovering and researching this issue.
6465 * An attacker can force openssl to consume large amounts of memory whilst
6466 processing DTLS handshake messages. This can be exploited through a
6467 Denial of Service attack.
6468 Thanks to Adam Langley for discovering and researching this issue.
6473 * An attacker can force an error condition which causes openssl to crash
6474 whilst processing DTLS packets due to memory being freed twice. This
6475 can be exploited through a Denial of Service attack.
6476 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6482 * If a multithreaded client connects to a malicious server using a resumed
6483 session and the server sends an ec point format extension it could write
6484 up to 255 bytes to freed memory.
6486 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6492 * A malicious server can crash an OpenSSL client with a null pointer
6493 dereference (read) by specifying an SRP ciphersuite even though it was not
6494 properly negotiated with the client. This can be exploited through a
6495 Denial of Service attack.
6497 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6498 discovering and researching this issue.
6503 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6504 X509_name_oneline, X509_name_print_ex et al. to leak some information
6505 from the stack. Applications may be affected if they echo pretty printing
6506 output to the attacker.
6508 Thanks to Ivan Fratric (Google) for discovering this issue.
6511 *Emilia Käsper, and Steve Henson*
6513 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6514 for corner cases. (Certain input points at infinity could lead to
6515 bogus results, with non-infinity inputs mapped to infinity too.)
6519 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6521 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6522 handshake can force the use of weak keying material in OpenSSL
6523 SSL/TLS clients and servers.
6525 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6526 researching this issue. ([CVE-2014-0224])
6528 *KIKUCHI Masashi, Steve Henson*
6530 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6531 OpenSSL DTLS client the code can be made to recurse eventually crashing
6534 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6537 *Imre Rad, Steve Henson*
6539 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6540 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6541 client or server. This is potentially exploitable to run arbitrary
6542 code on a vulnerable client or server.
6544 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6546 *Jüri Aedla, Steve Henson*
6548 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6549 are subject to a denial of service attack.
6551 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6552 this issue. ([CVE-2014-3470])
6554 *Felix Gröbert, Ivan Fratric, Steve Henson*
6556 * Harmonize version and its documentation. -f flag is used to display
6559 *mancha <mancha1@zoho.com>*
6561 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6562 in i2d_ECPrivateKey.
6564 *mancha <mancha1@zoho.com>*
6566 * Fix some double frees. These are not thought to be exploitable.
6568 *mancha <mancha1@zoho.com>*
6570 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6572 * A missing bounds check in the handling of the TLS heartbeat extension
6573 can be used to reveal up to 64k of memory to a connected client or
6576 Thanks for Neel Mehta of Google Security for discovering this bug and to
6577 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6578 preparing the fix ([CVE-2014-0160])
6580 *Adam Langley, Bodo Moeller*
6582 * Fix for the attack described in the paper "Recovering OpenSSL
6583 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6584 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6585 <http://eprint.iacr.org/2014/140>
6587 Thanks to Yuval Yarom and Naomi Benger for discovering this
6588 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6590 *Yuval Yarom and Naomi Benger*
6592 * TLS pad extension: draft-agl-tls-padding-03
6594 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6595 TLS client Hello record length value would otherwise be > 255 and
6596 less that 512 pad with a dummy extension containing zeroes so it
6597 is at least 512 bytes long.
6599 *Adam Langley, Steve Henson*
6601 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6603 * Fix for TLS record tampering bug. A carefully crafted invalid
6604 handshake could crash OpenSSL with a NULL pointer exception.
6605 Thanks to Anton Johansson for reporting this issues.
6608 * Keep original DTLS digest and encryption contexts in retransmission
6609 structures so we can use the previous session parameters if they need
6610 to be resent. ([CVE-2013-6450])
6614 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6615 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6616 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6617 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6618 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6619 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6621 *Rob Stradling, Adam Langley*
6623 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6625 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6626 supporting platforms or when small records were transferred.
6628 *Andy Polyakov, Steve Henson*
6630 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6632 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6634 This addresses the flaw in CBC record processing discovered by
6635 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6636 at: <http://www.isg.rhul.ac.uk/tls/>
6638 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6639 Security Group at Royal Holloway, University of London
6640 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6641 Emilia Käsper for the initial patch.
6644 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6646 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6647 ciphersuites which can be exploited in a denial of service attack.
6648 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6649 and detecting this bug and to Wolfgang Ettlinger
6650 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6655 * Return an error when checking OCSP signatures when key is NULL.
6656 This fixes a DoS attack. ([CVE-2013-0166])
6660 * Make openssl verify return errors.
6662 *Chris Palmer <palmer@google.com> and Ben Laurie*
6664 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6665 the right response is stapled. Also change SSL_get_certificate()
6666 so it returns the certificate actually sent.
6667 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6669 *Rob Stradling <rob.stradling@comodo.com>*
6671 * Fix possible deadlock when decoding public keys.
6675 * Don't use TLS 1.0 record version number in initial client hello
6680 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6682 * Sanity check record length before skipping explicit IV in TLS
6683 1.2, 1.1 and DTLS to fix DoS attack.
6685 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6686 fuzzing as a service testing platform.
6691 * Initialise tkeylen properly when encrypting CMS messages.
6692 Thanks to Solar Designer of Openwall for reporting this issue.
6696 * In FIPS mode don't try to use composite ciphers as they are not
6701 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6703 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6704 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6705 mean any application compiled against OpenSSL 1.0.0 headers setting
6706 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6707 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6708 0x10000000L Any application which was previously compiled against
6709 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6710 will need to be recompiled as a result. Letting be results in
6711 inability to disable specifically TLS 1.1 and in client context,
6712 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6716 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6717 disable just protocol X, but all protocols above X *if* there are
6718 protocols *below* X still enabled. In more practical terms it means
6719 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6720 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6721 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6726 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6728 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6729 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6730 in CRYPTO_realloc_clean.
6732 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6733 issue and to Adam Langley <agl@chromium.org> for fixing it.
6736 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6738 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6742 * Workarounds for some broken servers that "hang" if a client hello
6743 record length exceeds 255 bytes.
6745 1. Do not use record version number > TLS 1.0 in initial client
6746 hello: some (but not all) hanging servers will now work.
6747 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6748 the number of ciphers sent in the client hello. This should be
6749 set to an even number, such as 50, for example by passing:
6750 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6751 Most broken servers should now work.
6752 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6753 TLS 1.2 client support entirely.
6757 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6761 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6763 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6764 STRING form instead of a DigestInfo.
6768 * The format used for MDC2 RSA signatures is inconsistent between EVP
6769 and the RSA_sign/RSA_verify functions. This was made more apparent when
6770 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6771 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6772 the correct format in RSA_verify so both forms transparently work.
6776 * Some servers which support TLS 1.0 can choke if we initially indicate
6777 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6778 encrypted premaster secret. As a workaround use the maximum permitted
6779 client version in client hello, this should keep such servers happy
6780 and still work with previous versions of OpenSSL.
6784 * Add support for TLS/DTLS heartbeats.
6786 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6788 * Add support for SCTP.
6790 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6792 * Improved PRNG seeding for VOS.
6794 *Paul Green <Paul.Green@stratus.com>*
6796 * Extensive assembler packs updates, most notably:
6798 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6799 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6800 - x86_64: bit-sliced AES implementation;
6801 - ARM: NEON support, contemporary platforms optimizations;
6802 - s390x: z196 support;
6803 - `*`: GHASH and GF(2^m) multiplication implementations;
6807 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6808 (removal of unnecessary code)
6810 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6812 * Add TLS key material exporter from RFC 5705.
6816 * Add DTLS-SRTP negotiation from RFC 5764.
6820 * Add Next Protocol Negotiation,
6821 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6822 disabled with a no-npn flag to config or Configure. Code donated
6825 *Adam Langley <agl@google.com> and Ben Laurie*
6827 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6828 NIST-P256, NIST-P521, with constant-time single point multiplication on
6829 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6830 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6831 Code made available under Apache License version 2.0.
6833 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6834 line to include this in your build of OpenSSL, and run "make depend" (or
6835 "make update"). This enables the following EC_METHODs:
6837 EC_GFp_nistp224_method()
6838 EC_GFp_nistp256_method()
6839 EC_GFp_nistp521_method()
6841 EC_GROUP_new_by_curve_name() will automatically use these (while
6842 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6845 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6847 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6848 all platforms. Move ssize_t definition from e_os.h to the public
6849 header file e_os2.h as it now appears in public header file cms.h
6853 * New -sigopt option to the ca, req and x509 utilities. Additional
6854 signature parameters can be passed using this option and in
6859 * Add RSA PSS signing function. This will generate and set the
6860 appropriate AlgorithmIdentifiers for PSS based on those in the
6861 corresponding EVP_MD_CTX structure. No application support yet.
6865 * Support for companion algorithm specific ASN1 signing routines.
6866 New function ASN1_item_sign_ctx() signs a pre-initialised
6867 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6868 the appropriate parameters.
6872 * Add new algorithm specific ASN1 verification initialisation function
6873 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6874 handling will be the same no matter what EVP_PKEY_METHOD is used.
6875 Add a PSS handler to support verification of PSS signatures: checked
6876 against a number of sample certificates.
6880 * Add signature printing for PSS. Add PSS OIDs.
6882 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6884 * Add algorithm specific signature printing. An individual ASN1 method
6885 can now print out signatures instead of the standard hex dump.
6887 More complex signatures (e.g. PSS) can print out more meaningful
6888 information. Include DSA version that prints out the signature
6893 * Password based recipient info support for CMS library: implementing
6898 * Split password based encryption into PBES2 and PBKDF2 functions. This
6899 neatly separates the code into cipher and PBE sections and is required
6900 for some algorithms that split PBES2 into separate pieces (such as
6901 password based CMS).
6905 * Session-handling fixes:
6906 - Fix handling of connections that are resuming with a session ID,
6907 but also support Session Tickets.
6908 - Fix a bug that suppressed issuing of a new ticket if the client
6909 presented a ticket with an expired session.
6910 - Try to set the ticket lifetime hint to something reasonable.
6911 - Make tickets shorter by excluding irrelevant information.
6912 - On the client side, don't ignore renewed tickets.
6914 *Adam Langley, Bodo Moeller (Google)*
6916 * Fix PSK session representation.
6920 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6922 This work was sponsored by Intel.
6926 * Add GCM support to TLS library. Some custom code is needed to split
6927 the IV between the fixed (from PRF) and explicit (from TLS record)
6928 portions. This adds all GCM ciphersuites supported by RFC5288 and
6929 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6930 add a special AESGCM string for GCM only.
6934 * Expand range of ctrls for AES GCM. Permit setting invocation
6935 field on decrypt and retrieval of invocation field only on encrypt.
6939 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6940 As required by RFC5289 these ciphersuites cannot be used if for
6941 versions of TLS earlier than 1.2.
6945 * For FIPS capable OpenSSL interpret a NULL default public key method
6946 as unset and return the appropriate default but do *not* set the default.
6947 This means we can return the appropriate method in applications that
6948 switch between FIPS and non-FIPS modes.
6952 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6953 ENGINE is used then we cannot handle that in the FIPS module so we
6954 keep original code iff non-FIPS operations are allowed.
6958 * Add -attime option to openssl utilities.
6960 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6962 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6966 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6967 FIPS EC methods unconditionally for now.
6971 * New build option no-ec2m to disable characteristic 2 code.
6975 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6976 all cases can be covered as some introduce binary incompatibilities.
6980 * Redirect RSA operations to FIPS module including keygen,
6981 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6985 * Add similar low level API blocking to ciphers.
6989 * Low level digest APIs are not approved in FIPS mode: any attempt
6990 to use these will cause a fatal error. Applications that *really* want
6991 to use them can use the `private_*` version instead.
6995 * Redirect cipher operations to FIPS module for FIPS builds.
6999 * Redirect digest operations to FIPS module for FIPS builds.
7003 * Update build system to add "fips" flag which will link in fipscanister.o
7004 for static and shared library builds embedding a signature if needed.
7008 * Output TLS supported curves in preference order instead of numerical
7009 order. This is currently hardcoded for the highest order curves first.
7010 This should be configurable so applications can judge speed vs strength.
7014 * Add TLS v1.2 server support for client authentication.
7018 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7023 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7024 FIPS modules versions.
7028 * Add TLS v1.2 client side support for client authentication. Keep cache
7029 of handshake records longer as we don't know the hash algorithm to use
7030 until after the certificate request message is received.
7034 * Initial TLS v1.2 client support. Add a default signature algorithms
7035 extension including all the algorithms we support. Parse new signature
7036 format in client key exchange. Relax some ECC signing restrictions for
7037 TLS v1.2 as indicated in RFC5246.
7041 * Add server support for TLS v1.2 signature algorithms extension. Switch
7042 to new signature format when needed using client digest preference.
7043 All server ciphersuites should now work correctly in TLS v1.2. No client
7044 support yet and no support for client certificates.
7048 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7049 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7050 ciphersuites. At present only RSA key exchange ciphersuites work with
7051 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7052 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7053 and version checking.
7057 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7058 with this defined it will not be affected by any changes to ssl internal
7059 structures. Add several utility functions to allow openssl application
7060 to work with OPENSSL_NO_SSL_INTERN defined.
7064 * A long standing patch to add support for SRP from EdelWeb (Peter
7065 Sylvester and Christophe Renou) was integrated.
7066 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7067 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7070 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7074 * Permit abbreviated handshakes when renegotiating using the function
7075 SSL_renegotiate_abbreviated().
7077 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7079 * Add call to ENGINE_register_all_complete() to
7080 ENGINE_load_builtin_engines(), so some implementations get used
7081 automatically instead of needing explicit application support.
7085 * Add support for TLS key exporter as described in RFC5705.
7087 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7089 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7090 a few changes are required:
7092 Add SSL_OP_NO_TLSv1_1 flag.
7093 Add TLSv1_1 methods.
7094 Update version checking logic to handle version 1.1.
7095 Add explicit IV handling (ported from DTLS code).
7096 Add command line options to s_client/s_server.
7103 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7105 * X509_ATTRIBUTE memory leak
7107 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7108 memory. This structure is used by the PKCS#7 and CMS routines so any
7109 application which reads PKCS#7 or CMS data from untrusted sources is
7110 affected. SSL/TLS is not affected.
7112 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7118 * Race condition handling PSK identify hint
7120 If PSK identity hints are received by a multi-threaded client then
7121 the values are wrongly updated in the parent SSL_CTX structure. This can
7122 result in a race condition potentially leading to a double free of the
7128 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7130 * Malformed ECParameters causes infinite loop
7132 When processing an ECParameters structure OpenSSL enters an infinite loop
7133 if the curve specified is over a specially malformed binary polynomial
7136 This can be used to perform denial of service against any
7137 system which processes public keys, certificate requests or
7138 certificates. This includes TLS clients and TLS servers with
7139 client authentication enabled.
7141 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7146 * Exploitable out-of-bounds read in X509_cmp_time
7148 X509_cmp_time does not properly check the length of the ASN1_TIME
7149 string and can read a few bytes out of bounds. In addition,
7150 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7153 An attacker can use this to craft malformed certificates and CRLs of
7154 various sizes and potentially cause a segmentation fault, resulting in
7155 a DoS on applications that verify certificates or CRLs. TLS clients
7156 that verify CRLs are affected. TLS clients and servers with client
7157 authentication enabled may be affected if they use custom verification
7160 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7161 independently by Hanno Böck.
7166 * PKCS7 crash with missing EnvelopedContent
7168 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7169 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7170 with missing content and trigger a NULL pointer dereference on parsing.
7172 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7173 structures from untrusted sources are affected. OpenSSL clients and
7174 servers are not affected.
7176 This issue was reported to OpenSSL by Michal Zalewski (Google).
7181 * CMS verify infinite loop with unknown hash function
7183 When verifying a signedData message the CMS code can enter an infinite loop
7184 if presented with an unknown hash function OID. This can be used to perform
7185 denial of service against any system which verifies signedData messages using
7187 This issue was reported to OpenSSL by Johannes Bauer.
7192 * Race condition handling NewSessionTicket
7194 If a NewSessionTicket is received by a multi-threaded client when attempting to
7195 reuse a previous ticket then a race condition can occur potentially leading to
7196 a double free of the ticket data.
7201 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7203 * Segmentation fault in ASN1_TYPE_cmp fix
7205 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7206 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7207 certificate signature algorithm consistency this can be used to crash any
7208 certificate verification operation and exploited in a DoS attack. Any
7209 application which performs certificate verification is vulnerable including
7210 OpenSSL clients and servers which enable client authentication.
7215 * ASN.1 structure reuse memory corruption fix
7217 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7218 memory corruption via an invalid write. Such reuse is and has been
7219 strongly discouraged and is believed to be rare.
7221 Applications that parse structures containing CHOICE or ANY DEFINED BY
7222 components may be affected. Certificate parsing (d2i_X509 and related
7223 functions) are however not affected. OpenSSL clients and servers are
7229 * PKCS7 NULL pointer dereferences fix
7231 The PKCS#7 parsing code does not handle missing outer ContentInfo
7232 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7233 missing content and trigger a NULL pointer dereference on parsing.
7235 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7236 otherwise parse PKCS#7 structures from untrusted sources are
7237 affected. OpenSSL clients and servers are not affected.
7239 This issue was reported to OpenSSL by Michal Zalewski (Google).
7244 * DoS via reachable assert in SSLv2 servers fix
7246 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7247 servers that both support SSLv2 and enable export cipher suites by sending
7248 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7250 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7251 (OpenSSL development team).
7256 * Use After Free following d2i_ECPrivatekey error fix
7258 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7259 could cause a use after free condition. This, in turn, could cause a double
7260 free in several private key parsing functions (such as d2i_PrivateKey
7261 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7262 for applications that receive EC private keys from untrusted
7263 sources. This scenario is considered rare.
7265 This issue was discovered by the BoringSSL project and fixed in their
7271 * X509_to_X509_REQ NULL pointer deref fix
7273 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7274 the certificate key is invalid. This function is rarely used in practice.
7276 This issue was discovered by Brian Carpenter.
7281 * Removed the export ciphers from the DEFAULT ciphers
7285 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7287 * Build fixes for the Windows and OpenVMS platforms
7289 *Matt Caswell and Richard Levitte*
7291 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7293 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7294 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7295 dereference. This could lead to a Denial Of Service attack. Thanks to
7296 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7301 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7302 dtls1_buffer_record function under certain conditions. In particular this
7303 could occur if an attacker sent repeated DTLS records with the same
7304 sequence number but for the next epoch. The memory leak could be exploited
7305 by an attacker in a Denial of Service attack through memory exhaustion.
7306 Thanks to Chris Mueller for reporting this issue.
7311 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7312 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7313 method would be set to NULL which could later result in a NULL pointer
7314 dereference. Thanks to Frank Schmirler for reporting this issue.
7319 * Abort handshake if server key exchange message is omitted for ephemeral
7322 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7323 reporting this issue.
7328 * Remove non-export ephemeral RSA code on client and server. This code
7329 violated the TLS standard by allowing the use of temporary RSA keys in
7330 non-export ciphersuites and could be used by a server to effectively
7331 downgrade the RSA key length used to a value smaller than the server
7332 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7333 INRIA or reporting this issue.
7338 * Fixed issue where DH client certificates are accepted without verification.
7339 An OpenSSL server will accept a DH certificate for client authentication
7340 without the certificate verify message. This effectively allows a client to
7341 authenticate without the use of a private key. This only affects servers
7342 which trust a client certificate authority which issues certificates
7343 containing DH keys: these are extremely rare and hardly ever encountered.
7344 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7350 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7351 results on some platforms, including x86_64. This bug occurs at random
7352 with a very low probability, and is not known to be exploitable in any
7353 way, though its exact impact is difficult to determine. Thanks to Pieter
7354 Wuille (Blockstream) who reported this issue and also suggested an initial
7355 fix. Further analysis was conducted by the OpenSSL development team and
7356 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7357 the OpenSSL core team.
7362 * Fix various certificate fingerprint issues.
7364 By using non-DER or invalid encodings outside the signed portion of a
7365 certificate the fingerprint can be changed without breaking the signature.
7366 Although no details of the signed portion of the certificate can be changed
7367 this can cause problems with some applications: e.g. those using the
7368 certificate fingerprint for blacklists.
7370 1. Reject signatures with non zero unused bits.
7372 If the BIT STRING containing the signature has non zero unused bits reject
7373 the signature. All current signature algorithms require zero unused bits.
7375 2. Check certificate algorithm consistency.
7377 Check the AlgorithmIdentifier inside TBS matches the one in the
7378 certificate signature. NB: this will result in signature failure
7379 errors for some broken certificates.
7381 Thanks to Konrad Kraszewski from Google for reporting this issue.
7383 3. Check DSA/ECDSA signatures use DER.
7385 Reencode DSA/ECDSA signatures and compare with the original received
7386 signature. Return an error if there is a mismatch.
7388 This will reject various cases including garbage after signature
7389 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7390 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7391 (negative or with leading zeroes).
7393 Further analysis was conducted and fixes were developed by Stephen Henson
7394 of the OpenSSL core team.
7400 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7402 * Session Ticket Memory Leak.
7404 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7405 integrity of that ticket is first verified. In the event of a session
7406 ticket integrity check failing, OpenSSL will fail to free memory
7407 causing a memory leak. By sending a large number of invalid session
7408 tickets an attacker could exploit this issue in a Denial Of Service
7414 * Build option no-ssl3 is incomplete.
7416 When OpenSSL is configured with "no-ssl3" as a build option, servers
7417 could accept and complete a SSL 3.0 handshake, and clients could be
7418 configured to send them.
7421 *Akamai and the OpenSSL team*
7423 * Add support for TLS_FALLBACK_SCSV.
7424 Client applications doing fallback retries should call
7425 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7428 *Adam Langley, Bodo Moeller*
7430 * Add additional DigestInfo checks.
7432 Reencode DigestInto in DER and check against the original when
7433 verifying RSA signature: this will reject any improperly encoded
7434 DigestInfo structures.
7436 Note: this is a precautionary measure and no attacks are currently known.
7440 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7442 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7443 to a denial of service attack. A malicious server can crash the client
7444 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7445 ciphersuite and sending carefully crafted handshake messages.
7447 Thanks to Felix Gröbert (Google) for discovering and researching this
7453 * By sending carefully crafted DTLS packets an attacker could cause openssl
7454 to leak memory. This can be exploited through a Denial of Service attack.
7455 Thanks to Adam Langley for discovering and researching this issue.
7460 * An attacker can force openssl to consume large amounts of memory whilst
7461 processing DTLS handshake messages. This can be exploited through a
7462 Denial of Service attack.
7463 Thanks to Adam Langley for discovering and researching this issue.
7468 * An attacker can force an error condition which causes openssl to crash
7469 whilst processing DTLS packets due to memory being freed twice. This
7470 can be exploited through a Denial of Service attack.
7471 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7477 * If a multithreaded client connects to a malicious server using a resumed
7478 session and the server sends an ec point format extension it could write
7479 up to 255 bytes to freed memory.
7481 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7487 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7488 X509_name_oneline, X509_name_print_ex et al. to leak some information
7489 from the stack. Applications may be affected if they echo pretty printing
7490 output to the attacker.
7492 Thanks to Ivan Fratric (Google) for discovering this issue.
7495 *Emilia Käsper, and Steve Henson*
7497 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7498 for corner cases. (Certain input points at infinity could lead to
7499 bogus results, with non-infinity inputs mapped to infinity too.)
7503 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7505 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7506 handshake can force the use of weak keying material in OpenSSL
7507 SSL/TLS clients and servers.
7509 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7510 researching this issue. ([CVE-2014-0224])
7512 *KIKUCHI Masashi, Steve Henson*
7514 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7515 OpenSSL DTLS client the code can be made to recurse eventually crashing
7518 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7521 *Imre Rad, Steve Henson*
7523 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7524 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7525 client or server. This is potentially exploitable to run arbitrary
7526 code on a vulnerable client or server.
7528 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7530 *Jüri Aedla, Steve Henson*
7532 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7533 are subject to a denial of service attack.
7535 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7536 this issue. ([CVE-2014-3470])
7538 *Felix Gröbert, Ivan Fratric, Steve Henson*
7540 * Harmonize version and its documentation. -f flag is used to display
7543 *mancha <mancha1@zoho.com>*
7545 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7546 in i2d_ECPrivateKey.
7548 *mancha <mancha1@zoho.com>*
7550 * Fix some double frees. These are not thought to be exploitable.
7552 *mancha <mancha1@zoho.com>*
7554 * Fix for the attack described in the paper "Recovering OpenSSL
7555 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7556 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7557 <http://eprint.iacr.org/2014/140>
7559 Thanks to Yuval Yarom and Naomi Benger for discovering this
7560 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7562 *Yuval Yarom and Naomi Benger*
7564 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7566 * Keep original DTLS digest and encryption contexts in retransmission
7567 structures so we can use the previous session parameters if they need
7568 to be resent. ([CVE-2013-6450])
7572 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7573 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7574 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7575 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7576 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7577 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7579 *Rob Stradling, Adam Langley*
7581 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7583 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7585 This addresses the flaw in CBC record processing discovered by
7586 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7587 at: <http://www.isg.rhul.ac.uk/tls/>
7589 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7590 Security Group at Royal Holloway, University of London
7591 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7592 Emilia Käsper for the initial patch.
7595 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7597 * Return an error when checking OCSP signatures when key is NULL.
7598 This fixes a DoS attack. ([CVE-2013-0166])
7602 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7603 the right response is stapled. Also change SSL_get_certificate()
7604 so it returns the certificate actually sent.
7605 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7606 (This is a backport)
7608 *Rob Stradling <rob.stradling@comodo.com>*
7610 * Fix possible deadlock when decoding public keys.
7614 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7616 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7619 * Sanity check record length before skipping explicit IV in DTLS
7622 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7623 fuzzing as a service testing platform.
7628 * Initialise tkeylen properly when encrypting CMS messages.
7629 Thanks to Solar Designer of Openwall for reporting this issue.
7633 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7635 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7636 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7637 in CRYPTO_realloc_clean.
7639 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7640 issue and to Adam Langley <agl@chromium.org> for fixing it.
7643 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7645 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7647 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7648 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7649 content decryption and always return the same error. Note: this attack
7650 needs on average 2^20 messages so it only affects automated senders. The
7651 old behaviour can be re-enabled in the CMS code by setting the
7652 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7653 an MMA defence is not necessary.
7654 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7655 this issue. ([CVE-2012-0884])
7659 * Fix CVE-2011-4619: make sure we really are receiving a
7660 client hello before rejecting multiple SGC restarts. Thanks to
7661 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7665 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7667 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7668 Thanks to Antonio Martin, Enterprise Secure Access Research and
7669 Development, Cisco Systems, Inc. for discovering this bug and
7670 preparing a fix. ([CVE-2012-0050])
7674 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7676 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7677 of the Vaudenay padding oracle attack on CBC mode encryption
7678 which enables an efficient plaintext recovery attack against
7679 the OpenSSL implementation of DTLS. Their attack exploits timing
7680 differences arising during decryption processing. A research
7681 paper describing this attack can be found at:
7682 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7683 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7684 Security Group at Royal Holloway, University of London
7685 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7686 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7687 for preparing the fix. ([CVE-2011-4108])
7689 *Robin Seggelmann, Michael Tuexen*
7691 * Clear bytes used for block padding of SSL 3.0 records.
7694 *Adam Langley (Google)*
7696 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7697 Kadianakis <desnacked@gmail.com> for discovering this issue and
7698 Adam Langley for preparing the fix. ([CVE-2011-4619])
7700 *Adam Langley (Google)*
7702 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7704 *Andrey Kulikov <amdeich@gmail.com>*
7706 * Prevent malformed RFC3779 data triggering an assertion failure.
7707 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7708 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7710 *Rob Austein <sra@hactrn.net>*
7712 * Improved PRNG seeding for VOS.
7714 *Paul Green <Paul.Green@stratus.com>*
7716 * Fix ssl_ciph.c set-up race.
7718 *Adam Langley (Google)*
7720 * Fix spurious failures in ecdsatest.c.
7722 *Emilia Käsper (Google)*
7724 * Fix the BIO_f_buffer() implementation (which was mixing different
7725 interpretations of the `..._len` fields).
7727 *Adam Langley (Google)*
7729 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7730 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7731 threads won't reuse the same blinding coefficients.
7733 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7734 lock to call BN_BLINDING_invert_ex, and avoids one use of
7735 BN_BLINDING_update for each BN_BLINDING structure (previously,
7736 the last update always remained unused).
7738 *Emilia Käsper (Google)*
7740 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7742 *Bob Buckholz (Google)*
7744 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7746 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7747 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7749 *Kaspar Brand <ossl@velox.ch>*
7751 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7752 for multi-threaded use of ECDH. ([CVE-2011-3210])
7754 *Adam Langley (Google)*
7756 * Fix x509_name_ex_d2i memory leak on bad inputs.
7760 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7761 signature public key algorithm by using OID xref utilities instead.
7762 Before this you could only use some ECC ciphersuites with SHA1 only.
7766 * Add protection against ECDSA timing attacks as mentioned in the paper
7767 by Billy Bob Brumley and Nicola Tuveri, see:
7768 <http://eprint.iacr.org/2011/232.pdf>
7770 *Billy Bob Brumley and Nicola Tuveri*
7772 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7774 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7776 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7778 * Fix bug in string printing code: if *any* escaping is enabled we must
7779 escape the escape character (backslash) or the resulting string is
7784 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7786 * Disable code workaround for ancient and obsolete Netscape browsers
7787 and servers: an attacker can use it in a ciphersuite downgrade attack.
7788 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7792 * Fixed J-PAKE implementation error, originally discovered by
7793 Sebastien Martini, further info and confirmation from Stefan
7794 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7798 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7800 * Fix extension code to avoid race conditions which can result in a buffer
7801 overrun vulnerability: resumed sessions must not be modified as they can
7802 be shared by multiple threads. CVE-2010-3864
7806 * Fix WIN32 build system to correctly link an ENGINE directory into
7811 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7813 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7816 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7818 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7820 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7821 context. The operation can be customised via the ctrl mechanism in
7822 case ENGINEs want to include additional functionality.
7826 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7830 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7831 output hashes compatible with older versions of OpenSSL.
7833 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7835 * Fix compression algorithm handling: if resuming a session use the
7836 compression algorithm of the resumed session instead of determining
7837 it from client hello again. Don't allow server to change algorithm.
7841 * Add load_crls() function to apps tidying load_certs() too. Add option
7842 to verify utility to allow additional CRLs to be included.
7846 * Update OCSP request code to permit adding custom headers to the request:
7847 some responders need this.
7851 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7854 *Julia Lawall <julia@diku.dk>*
7856 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7857 needlessly dereferenced structures, used obsolete functions and
7858 didn't handle all updated verify codes correctly.
7862 * Disable MD2 in the default configuration.
7866 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7867 indicate the initial BIO being pushed or popped. This makes it possible
7868 to determine whether the BIO is the one explicitly called or as a result
7869 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7870 it handles reference counts correctly and doesn't zero out the I/O bio
7871 when it is not being explicitly popped. WARNING: applications which
7872 included workarounds for the old buggy behaviour will need to be modified
7873 or they could free up already freed BIOs.
7877 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7878 renaming to all platforms (within the 0.9.8 branch, this was
7879 done conditionally on Netware platforms to avoid a name clash).
7881 *Guenter <lists@gknw.net>*
7883 * Add ECDHE and PSK support to DTLS.
7885 *Michael Tuexen <tuexen@fh-muenster.de>*
7887 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7892 * Add "missing" function EVP_MD_flags() (without this the only way to
7893 retrieve a digest flags is by accessing the structure directly. Update
7894 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7895 or cipher is registered as in the "from" argument. Print out all
7896 registered digests in the dgst usage message instead of manually
7897 attempting to work them out.
7901 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7902 this allows the use of compression and extensions. Change default cipher
7903 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7904 by default unless an application cipher string requests it.
7908 * Alter match criteria in PKCS12_parse(). It used to try to use local
7909 key ids to find matching certificates and keys but some PKCS#12 files
7910 don't follow the (somewhat unwritten) rules and this strategy fails.
7911 Now just gather all certificates together and the first private key
7912 then look for the first certificate that matches the key.
7916 * Support use of registered digest and cipher names for dgst and cipher
7917 commands instead of having to add each one as a special case. So now
7924 openssl dgst -sha256 foo
7926 and this works for ENGINE based algorithms too.
7930 * Update Gost ENGINE to support parameter files.
7932 *Victor B. Wagner <vitus@cryptocom.ru>*
7934 * Support GeneralizedTime in ca utility.
7936 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7938 * Enhance the hash format used for certificate directory links. The new
7939 form uses the canonical encoding (meaning equivalent names will work
7940 even if they aren't identical) and uses SHA1 instead of MD5. This form
7941 is incompatible with the older format and as a result c_rehash should
7942 be used to rebuild symbolic links.
7946 * Make PKCS#8 the default write format for private keys, replacing the
7947 traditional format. This form is standardised, more secure and doesn't
7948 include an implicit MD5 dependency.
7952 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7953 committed to OpenSSL should pass this lot as a minimum.
7957 * Add session ticket override functionality for use by EAP-FAST.
7959 *Jouni Malinen <j@w1.fi>*
7961 * Modify HMAC functions to return a value. Since these can be implemented
7962 in an ENGINE errors can occur.
7966 * Type-checked OBJ_bsearch_ex.
7970 * Type-checked OBJ_bsearch. Also some constification necessitated
7971 by type-checking. Still to come: TXT_DB, bsearch(?),
7972 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7977 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7978 seconds to a tm structure directly, instead of going through OS
7979 specific date routines. This avoids any issues with OS routines such
7980 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7981 and X509_time_adj_ex() to cover the extended range. The existing
7982 X509_time_adj() is still usable and will no longer have any date issues.
7986 * Delta CRL support. New use deltas option which will attempt to locate
7987 and search any appropriate delta CRLs available.
7989 This work was sponsored by Google.
7993 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7994 code and add additional score elements. Validate alternate CRL paths
7995 as part of the CRL checking and indicate a new error "CRL path validation
7996 error" in this case. Applications wanting additional details can use
7997 the verify callback and check the new "parent" field. If this is not
7998 NULL CRL path validation is taking place. Existing applications won't
7999 see this because it requires extended CRL support which is off by
8002 This work was sponsored by Google.
8006 * Support for freshest CRL extension.
8008 This work was sponsored by Google.
8012 * Initial indirect CRL support. Currently only supported in the CRLs
8013 passed directly and not via lookup. Process certificate issuer
8014 CRL entry extension and lookup CRL entries by bother issuer name
8015 and serial number. Check and process CRL issuer entry in IDP extension.
8017 This work was sponsored by Google.
8021 * Add support for distinct certificate and CRL paths. The CRL issuer
8022 certificate is validated separately in this case. Only enabled if
8023 an extended CRL support flag is set: this flag will enable additional
8024 CRL functionality in future.
8026 This work was sponsored by Google.
8030 * Add support for policy mappings extension.
8032 This work was sponsored by Google.
8036 * Fixes to pathlength constraint, self issued certificate handling,
8037 policy processing to align with RFC3280 and PKITS tests.
8039 This work was sponsored by Google.
8043 * Support for name constraints certificate extension. DN, email, DNS
8044 and URI types are currently supported.
8046 This work was sponsored by Google.
8050 * To cater for systems that provide a pointer-based thread ID rather
8051 than numeric, deprecate the current numeric thread ID mechanism and
8052 replace it with a structure and associated callback type. This
8053 mechanism allows a numeric "hash" to be extracted from a thread ID in
8054 either case, and on platforms where pointers are larger than 'long',
8055 mixing is done to help ensure the numeric 'hash' is usable even if it
8056 can't be guaranteed unique. The default mechanism is to use "&errno"
8057 as a pointer-based thread ID to distinguish between threads.
8059 Applications that want to provide their own thread IDs should now use
8060 CRYPTO_THREADID_set_callback() to register a callback that will call
8061 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8063 Note that ERR_remove_state() is now deprecated, because it is tied
8064 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8065 to free the current thread's error state should be replaced by
8066 ERR_remove_thread_state(NULL).
8068 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8069 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8070 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8071 application was previously providing a numeric thread callback that
8072 was inappropriate for distinguishing threads, then uniqueness might
8073 have been obtained with &errno that happened immediately in the
8074 intermediate development versions of OpenSSL; this is no longer the
8075 case, the numeric thread callback will now override the automatic use
8078 *Geoff Thorpe, with help from Bodo Moeller*
8080 * Initial support for different CRL issuing certificates. This covers a
8081 simple case where the self issued certificates in the chain exist and
8082 the real CRL issuer is higher in the existing chain.
8084 This work was sponsored by Google.
8088 * Removed effectively defunct crypto/store from the build.
8092 * Revamp of STACK to provide stronger type-checking. Still to come:
8093 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8094 ASN1_STRING, CONF_VALUE.
8098 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8099 RAM on SSL connections. This option can save about 34k per idle SSL.
8103 * Revamp of LHASH to provide stronger type-checking. Still to come:
8104 STACK, TXT_DB, bsearch, qsort.
8108 * Initial support for Cryptographic Message Syntax (aka CMS) based
8109 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8110 support for data, signedData, compressedData, digestedData and
8111 encryptedData, envelopedData types included. Scripts to check against
8112 RFC4134 examples draft and interop and consistency checks of many
8113 content types and variants.
8117 * Add options to enc utility to support use of zlib compression BIO.
8121 * Extend mk1mf to support importing of options and assembly language
8122 files from Configure script, currently only included in VC-WIN32.
8123 The assembly language rules can now optionally generate the source
8124 files from the associated perl scripts.
8128 * Implement remaining functionality needed to support GOST ciphersuites.
8129 Interop testing has been performed using CryptoPro implementations.
8131 *Victor B. Wagner <vitus@cryptocom.ru>*
8133 * s390x assembler pack.
8137 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8142 * Implement Opaque PRF Input TLS extension as specified in
8143 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8144 official specification yet and no extension type assignment by
8145 IANA exists, this extension (for now) will have to be explicitly
8146 enabled when building OpenSSL by providing the extension number
8147 to use. For example, specify an option
8149 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8151 to the "config" or "Configure" script to enable the extension,
8152 assuming extension number 0x9527 (which is a completely arbitrary
8153 and unofficial assignment based on the MD5 hash of the Internet
8154 Draft). Note that by doing so, you potentially lose
8155 interoperability with other TLS implementations since these might
8156 be using the same extension number for other purposes.
8158 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8159 opaque PRF input value to use in the handshake. This will create
8160 an internal copy of the length-'len' string at 'src', and will
8161 return non-zero for success.
8163 To get more control and flexibility, provide a callback function
8166 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8167 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8171 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8174 Callback function 'cb' will be called in handshakes, and is
8175 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8176 Argument 'arg' is for application purposes (the value as given to
8177 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8178 be provided to the callback function). The callback function
8179 has to return non-zero to report success: usually 1 to use opaque
8180 PRF input just if possible, or 2 to enforce use of the opaque PRF
8181 input. In the latter case, the library will abort the handshake
8182 if opaque PRF input is not successfully negotiated.
8184 Arguments 'peerinput' and 'len' given to the callback function
8185 will always be NULL and 0 in the case of a client. A server will
8186 see the client's opaque PRF input through these variables if
8187 available (NULL and 0 otherwise). Note that if the server
8188 provides an opaque PRF input, the length must be the same as the
8189 length of the client's opaque PRF input.
8191 Note that the callback function will only be called when creating
8192 a new session (session resumption can resume whatever was
8193 previously negotiated), and will not be called in SSL 2.0
8194 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8195 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8196 for applications that need to enforce opaque PRF input.
8200 * Update ssl code to support digests other than SHA1+MD5 for handshake
8203 *Victor B. Wagner <vitus@cryptocom.ru>*
8205 * Add RFC4507 support to OpenSSL. This includes the corrections in
8206 RFC4507bis. The encrypted ticket format is an encrypted encoded
8207 SSL_SESSION structure, that way new session features are automatically
8210 If a client application caches session in an SSL_SESSION structure
8211 support is transparent because tickets are now stored in the encoded
8214 The SSL_CTX structure automatically generates keys for ticket
8215 protection in servers so again support should be possible
8216 with no application modification.
8218 If a client or server wishes to disable RFC4507 support then the option
8219 SSL_OP_NO_TICKET can be set.
8221 Add a TLS extension debugging callback to allow the contents of any client
8222 or server extensions to be examined.
8224 This work was sponsored by Google.
8228 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8229 OpenSSL should now compile cleanly on gcc 4.2
8231 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8233 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8234 support including streaming MAC support: this is required for GOST
8235 ciphersuite support.
8237 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8239 * Add option -stream to use PKCS#7 streaming in smime utility. New
8240 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8241 to output in BER and PEM format.
8245 * Experimental support for use of HMAC via EVP_PKEY interface. This
8246 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8247 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8248 ENGINE support for HMAC keys which are unextractable. New -mac and
8249 -macopt options to dgst utility.
8253 * New option -sigopt to dgst utility. Update dgst to use
8254 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8255 alternative signing parameters such as X9.31 or PSS in the dgst
8260 * Change ssl_cipher_apply_rule(), the internal function that does
8261 the work each time a ciphersuite string requests enabling
8262 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8263 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8264 the order of disabled ciphersuites such that those ciphersuites
8265 that most recently went from enabled to disabled not only stay
8266 in order with respect to each other, but also have higher priority
8267 than other disabled ciphersuites the next time ciphersuites are
8270 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8271 the same ciphersuites as with "HIGH" alone, but in a specific
8272 order where the PSK ciphersuites come first (since they are the
8273 most recently disabled ciphersuites when "HIGH" is parsed).
8275 Also, change ssl_create_cipher_list() (using this new
8276 functionality) such that between otherwise identical
8277 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8282 * Change ssl_create_cipher_list() so that it automatically
8283 arranges the ciphersuites in reasonable order before starting
8284 to process the rule string. Thus, the definition for "DEFAULT"
8285 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8286 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8287 This makes it much easier to arrive at a reasonable default order
8288 in applications for which anonymous ciphers are OK (meaning
8289 that you can't actually use DEFAULT).
8291 *Bodo Moeller; suggested by Victor Duchovni*
8293 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8294 processing) into multiple integers instead of setting
8295 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8296 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8297 (These masks as well as the individual bit definitions are hidden
8298 away into the non-exported interface ssl/ssl_locl.h, so this
8299 change to the definition of the SSL_CIPHER structure shouldn't
8300 affect applications.) This give us more bits for each of these
8301 categories, so there is no longer a need to coagulate AES128 and
8302 AES256 into a single algorithm bit, and to coagulate Camellia128
8303 and Camellia256 into a single algorithm bit, which has led to all
8306 Thus, among other things, the kludge introduced in 0.9.7m and
8307 0.9.8e for masking out AES256 independently of AES128 or masking
8308 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8310 With the change, we also introduce new ciphersuite aliases that
8311 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8316 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8317 Use the leftmost N bytes of the signature input if the input is
8318 larger than the prime q (with N being the size in bytes of q).
8322 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8323 it yet and it is largely untested.
8327 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8331 * Initial incomplete changes to avoid need for function casts in OpenSSL
8332 some compilers (gcc 4.2 and later) reject their use. Safestack is
8333 reimplemented. Update ASN1 to avoid use of legacy functions.
8337 * Win32/64 targets are linked with Winsock2.
8341 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8342 to external functions. This can be used to increase CRL handling
8343 efficiency especially when CRLs are very large by (for example) storing
8344 the CRL revoked certificates in a database.
8348 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8349 new CRLs added to a directory can be used. New command line option
8350 -verify_return_error to s_client and s_server. This causes real errors
8351 to be returned by the verify callback instead of carrying on no matter
8352 what. This reflects the way a "real world" verify callback would behave.
8356 * GOST engine, supporting several GOST algorithms and public key formats.
8357 Kindly donated by Cryptocom.
8361 * Partial support for Issuing Distribution Point CRL extension. CRLs
8362 partitioned by DP are handled but no indirect CRL or reason partitioning
8363 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8364 selected via a scoring technique which handles IDP and AKID in CRLs.
8368 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8369 will ultimately be used for all verify operations: this will remove the
8370 X509_STORE dependency on certificate verification and allow alternative
8371 lookup methods. X509_STORE based implementations of these two callbacks.
8375 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8376 Modify get_crl() to find a valid (unexpired) CRL if possible.
8380 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8381 this would be called X509_CRL_cmp() but that name is already used by
8382 a function that just compares CRL issuer names. Cache several CRL
8383 extensions in X509_CRL structure and cache CRLDP in X509.
8387 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8388 this maps equivalent X509_NAME structures into a consistent structure.
8389 Name comparison can then be performed rapidly using memcmp().
8393 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8398 * Allow digests to supply their own micalg string for S/MIME type using
8399 the ctrl EVP_MD_CTRL_MICALG.
8403 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8404 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8405 ctrl. It can then customise the structure before and/or after signing
8410 * New function OBJ_add_sigid() to allow application defined signature OIDs
8411 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8412 to free up any added signature OIDs.
8416 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8417 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8418 digest and cipher tables. New options added to openssl utility:
8419 list-message-digest-algorithms and list-cipher-algorithms.
8423 * Change the array representation of binary polynomials: the list
8424 of degrees of non-zero coefficients is now terminated with -1.
8425 Previously it was terminated with 0, which was also part of the
8426 value; thus, the array representation was not applicable to
8427 polynomials where t^0 has coefficient zero. This change makes
8428 the array representation useful in a more general context.
8432 * Various modifications and fixes to SSL/TLS cipher string
8433 handling. For ECC, the code now distinguishes between fixed ECDH
8434 with RSA certificates on the one hand and with ECDSA certificates
8435 on the other hand, since these are separate ciphersuites. The
8436 unused code for Fortezza ciphersuites has been removed.
8438 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8439 (not "ECDHE"). For consistency with the code for DH
8440 certificates, use of ECDH certificates is now considered ECDH
8441 authentication, not RSA or ECDSA authentication (the latter is
8442 merely the CA's signing algorithm and not actively used in the
8445 The temporary ciphersuite alias "ECCdraft" is no longer
8446 available, and ECC ciphersuites are no longer excluded from "ALL"
8447 and "DEFAULT". The following aliases now exist for RFC 4492
8448 ciphersuites, most of these by analogy with the DH case:
8450 kECDHr - ECDH cert, signed with RSA
8451 kECDHe - ECDH cert, signed with ECDSA
8452 kECDH - ECDH cert (signed with either RSA or ECDSA)
8453 kEECDH - ephemeral ECDH
8454 ECDH - ECDH cert or ephemeral ECDH
8460 AECDH - anonymous ECDH
8461 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8465 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8466 Use correct micalg parameters depending on digest(s) in signed message.
8470 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8471 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8475 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8476 an engine to register a method. Add ENGINE lookups for methods and
8477 functional reference processing.
8481 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8482 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8487 * New -resign option to smime utility. This adds one or more signers
8488 to an existing PKCS#7 signedData structure. Also -md option to use an
8489 alternative message digest algorithm for signing.
8493 * Tidy up PKCS#7 routines and add new functions to make it easier to
8494 create PKCS7 structures containing multiple signers. Update smime
8495 application to support multiple signers.
8499 * New -macalg option to pkcs12 utility to allow setting of an alternative
8504 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8505 Reorganize PBE internals to lookup from a static table using NIDs,
8506 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8507 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8508 PRF which will be automatically used with PBES2.
8512 * Replace the algorithm specific calls to generate keys in "req" with the
8517 * Update PKCS#7 enveloped data routines to use new API. This is now
8518 supported by any public key method supporting the encrypt operation. A
8519 ctrl is added to allow the public key algorithm to examine or modify
8520 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8525 * Add a ctrl to asn1 method to allow a public key algorithm to express
8526 a default digest type to use. In most cases this will be SHA1 but some
8527 algorithms (such as GOST) need to specify an alternative digest. The
8528 return value indicates how strong the preference is 1 means optional and
8529 2 is mandatory (that is it is the only supported type). Modify
8530 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8531 use the default md. Update openssl utilities to use the default digest
8532 type for signing if it is not explicitly indicated.
8536 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8537 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8538 signing method from the key type. This effectively removes the link
8539 between digests and public key types.
8543 * Add an OID cross reference table and utility functions. Its purpose is to
8544 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8545 rsaEncryption. This will allow some of the algorithm specific hackery
8546 needed to use the correct OID to be removed.
8550 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8551 structures for PKCS7_sign(). They are now set up by the relevant public
8556 * Add provisional EC pkey method with support for ECDSA and ECDH.
8560 * Add support for key derivation (agreement) in the API, DH method and
8565 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8566 public and private key formats. As a side effect these add additional
8567 command line functionality not previously available: DSA signatures can be
8568 generated and verified using pkeyutl and DH key support and generation in
8575 *Oliver Tappe <zooey@hirschkaefer.de>*
8577 * New make target "install_html_docs" installs HTML renditions of the
8580 *Oliver Tappe <zooey@hirschkaefer.de>*
8582 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8583 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8584 support key and parameter generation and add initial key generation
8585 functionality for RSA.
8589 * Add functions for main EVP_PKEY_method operations. The undocumented
8590 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8591 `EVP_PKEY_{encrypt,decrypt}_old`.
8595 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8596 key API, doesn't do much yet.
8600 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8601 public key algorithms. New option to openssl utility:
8602 "list-public-key-algorithms" to print out info.
8606 * Implement the Supported Elliptic Curves Extension for
8607 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8611 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8612 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8616 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8617 utilities such as rsa, dsa, dsaparam etc except they process any key
8622 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8623 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8624 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8629 * Initial support for pluggable public key ASN1.
8630 De-spaghettify the public key ASN1 handling. Move public and private
8631 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8632 algorithm specific handling to a single module within the relevant
8633 algorithm directory. Add functions to allow (near) opaque processing
8634 of public and private key structures.
8638 * Implement the Supported Point Formats Extension for
8639 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8643 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8644 for the psk identity [hint] and the psk callback functions to the
8645 SSL_SESSION, SSL and SSL_CTX structure.
8648 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8652 SSL_CTX_use_psk_identity_hint
8653 SSL_get_psk_identity_hint
8654 SSL_get_psk_identity
8655 SSL_use_psk_identity_hint
8657 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8659 * Add RFC 3161 compliant time stamp request creation, response generation
8660 and response verification functionality.
8662 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8664 * Add initial support for TLS extensions, specifically for the server_name
8665 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8666 have new members for a host name. The SSL data structure has an
8667 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8668 stored in that context to allow for session resumption, even after the
8669 SSL has been switched to a new SSL_CTX in reaction to a client's
8670 server_name extension.
8672 New functions (subject to change):
8674 SSL_get_servername()
8675 SSL_get_servername_type()
8678 New CTRL codes and macros (subject to change):
8680 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8681 - SSL_CTX_set_tlsext_servername_callback()
8682 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8683 - SSL_CTX_set_tlsext_servername_arg()
8684 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8686 openssl s_client has a new '-servername ...' option.
8688 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8689 '-key2 ...', '-servername_fatal' (subject to change). This allows
8690 testing the HostName extension for a specific single host name ('-cert'
8691 and '-key' remain fallbacks for handshakes without HostName
8692 negotiation). If the unrecognized_name alert has to be sent, this by
8693 default is a warning; it becomes fatal with the '-servername_fatal'
8696 *Peter Sylvester, Remy Allais, Christophe Renou*
8698 * Whirlpool hash implementation is added.
8702 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8703 bn(64,32). Because of instruction set limitations it doesn't have
8704 any negative impact on performance. This was done mostly in order
8705 to make it possible to share assembler modules, such as bn_mul_mont
8706 implementations, between 32- and 64-bit builds without hassle.
8710 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8711 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8716 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8717 dedicated Montgomery multiplication procedure, is introduced.
8718 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8719 "64-bit" performance on certain 32-bit targets.
8723 * New option SSL_OP_NO_COMP to disable use of compression selectively
8724 in SSL structures. New SSL ctrl to set maximum send fragment size.
8725 Save memory by setting the I/O buffer sizes dynamically instead of
8726 using the maximum available value.
8730 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8731 in addition to the text details.
8735 * Very, very preliminary EXPERIMENTAL support for printing of general
8736 ASN1 structures. This currently produces rather ugly output and doesn't
8737 handle several customised structures at all.
8741 * Integrated support for PVK file format and some related formats such
8742 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8743 these in the 'rsa' and 'dsa' utilities.
8747 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8751 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8752 place for the (very old) "NETSCAPE" format certificates which are now
8753 handled using new ASN1 code equivalents.
8757 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8758 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8759 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8763 * Modify CRL distribution points extension code to print out previously
8764 unsupported fields. Enhance extension setting code to allow setting of
8769 * Add print and set support for Issuing Distribution Point CRL extension.
8773 * Change 'Configure' script to enable Camellia by default.
8780 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8782 * When rejecting SSL/TLS records due to an incorrect version number, never
8783 update s->server with a new major version number. As of
8784 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8785 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8786 the previous behavior could result in a read attempt at NULL when
8787 receiving specific incorrect SSL/TLS records once record payload
8788 protection is active. ([CVE-2010-0740])
8790 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8792 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8793 could be crashed if the relevant tables were not present (e.g. chrooted).
8795 *Tomas Hoger <thoger@redhat.com>*
8797 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8799 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8801 *Martin Olsson, Neel Mehta*
8803 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8804 accommodate for stack sorting, always a write lock!).
8808 * On some versions of WIN32 Heap32Next is very slow. This can cause
8809 excessive delays in the RAND_poll(): over a minute. As a workaround
8810 include a time check in the inner Heap32Next loop too.
8814 * The code that handled flushing of data in SSL/TLS originally used the
8815 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8816 the problem outlined in PR#1949. The fix suggested there however can
8817 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8818 of Apache). So instead simplify the code to flush unconditionally.
8819 This should be fine since flushing with no data to flush is a no op.
8823 * Handle TLS versions 2.0 and later properly and correctly use the
8824 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8825 off ancient servers have a habit of sticking around for a while...
8829 * Modify compression code so it frees up structures without using the
8830 ex_data callbacks. This works around a problem where some applications
8831 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8832 restarting) then use compression (e.g. SSL with compression) later.
8833 This results in significant per-connection memory leaks and
8834 has caused some security issues including CVE-2008-1678 and
8839 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8840 change when encrypting or decrypting.
8844 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8845 connect and renegotiate with servers which do not support RI.
8846 Until RI is more widely deployed this option is enabled by default.
8850 * Add "missing" ssl ctrls to clear options and mode.
8854 * If client attempts to renegotiate and doesn't support RI respond with
8855 a no_renegotiation alert as required by RFC5746. Some renegotiating
8856 TLS clients will continue a connection gracefully when they receive
8857 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8858 waiting for a server hello which it will never receive. Now we treat a
8859 received no_renegotiation alert as a fatal error. This is because
8860 applications requesting a renegotiation might well expect it to succeed
8861 and would have no code in place to handle the server denying it so the
8862 only safe thing to do is to terminate the connection.
8866 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8867 peer supports secure renegotiation and 0 otherwise. Print out peer
8868 renegotiation support in s_client/s_server.
8872 * Replace the highly broken and deprecated SPKAC certification method with
8873 the updated NID creation version. This should correctly handle UTF8.
8877 * Implement RFC5746. Re-enable renegotiation but require the extension
8878 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8879 turns out to be a bad idea. It has been replaced by
8880 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8881 SSL_CTX_set_options(). This is really not recommended unless you
8882 know what you are doing.
8884 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8886 * Fixes to stateless session resumption handling. Use initial_ctx when
8887 issuing and attempting to decrypt tickets in case it has changed during
8888 servername handling. Use a non-zero length session ID when attempting
8889 stateless session resumption: this makes it possible to determine if
8890 a resumption has occurred immediately after receiving server hello
8891 (several places in OpenSSL subtly assume this) instead of later in
8896 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8897 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8898 fixes for a few places where the return code is not checked
8901 *Julia Lawall <julia@diku.dk>*
8903 * Add --strict-warnings option to Configure script to include devteam
8904 warnings in other configurations.
8908 * Add support for --libdir option and LIBDIR variable in makefiles. This
8909 makes it possible to install openssl libraries in locations which
8910 have names other than "lib", for example "/usr/lib64" which some
8913 *Steve Henson, based on patch from Jeremy Utley*
8915 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8916 X690 8.9.12 and can produce some misleading textual output of OIDs.
8918 *Steve Henson, reported by Dan Kaminsky*
8920 * Delete MD2 from algorithm tables. This follows the recommendation in
8921 several standards that it is not used in new applications due to
8922 several cryptographic weaknesses. For binary compatibility reasons
8923 the MD2 API is still compiled in by default.
8927 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8932 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8933 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8936 *Guenter <lists@gknw.net>*
8938 * Fix the server certificate chain building code to use X509_verify_cert(),
8939 it used to have an ad-hoc builder which was unable to cope with anything
8940 other than a simple chain.
8942 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8944 * Don't check self signed certificate signatures in X509_verify_cert()
8945 by default (a flag can override this): it just wastes time without
8946 adding any security. As a useful side effect self signed root CAs
8947 with non-FIPS digests are now usable in FIPS mode.
8951 * In dtls1_process_out_of_seq_message() the check if the current message
8952 is already buffered was missing. For every new message was memory
8953 allocated, allowing an attacker to perform an denial of service attack
8954 with sending out of seq handshake messages until there is no memory
8955 left. Additionally every future message was buffered, even if the
8956 sequence number made no sense and would be part of another handshake.
8957 So only messages with sequence numbers less than 10 in advance will be
8958 buffered. ([CVE-2009-1378])
8960 *Robin Seggelmann, discovered by Daniel Mentz*
8962 * Records are buffered if they arrive with a future epoch to be
8963 processed after finishing the corresponding handshake. There is
8964 currently no limitation to this buffer allowing an attacker to perform
8965 a DOS attack with sending records with future epochs until there is no
8966 memory left. This patch adds the pqueue_size() function to determine
8967 the size of a buffer and limits the record buffer to 100 entries.
8970 *Robin Seggelmann, discovered by Daniel Mentz*
8972 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8973 parent structure is freed. ([CVE-2009-1379])
8977 * Handle non-blocking I/O properly in SSL_shutdown() call.
8979 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8981 * Add `2.5.4.*` OIDs
8983 *Ilya O. <vrghost@gmail.com>*
8985 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8987 * Disable renegotiation completely - this fixes a severe security
8988 problem ([CVE-2009-3555]) at the cost of breaking all
8989 renegotiation. Renegotiation can be re-enabled by setting
8990 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8991 run-time. This is really not recommended unless you know what
8996 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8998 * Don't set val to NULL when freeing up structures, it is freed up by
8999 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9000 zeroing past the valid field. ([CVE-2009-0789])
9002 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9004 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9005 checked correctly. This would allow some invalid signed attributes to
9006 appear to verify correctly. ([CVE-2009-0591])
9008 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9010 * Reject UniversalString and BMPString types with invalid lengths. This
9011 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9012 a legal length. ([CVE-2009-0590])
9016 * Set S/MIME signing as the default purpose rather than setting it
9017 unconditionally. This allows applications to override it at the store
9022 * Permit restricted recursion of ASN1 strings. This is needed in practice
9023 to handle some structures.
9027 * Improve efficiency of mem_gets: don't search whole buffer each time
9030 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9032 * New -hex option for openssl rand.
9036 * Print out UTF8String and NumericString when parsing ASN1.
9040 * Support NumericString type for name components.
9044 * Allow CC in the environment to override the automatically chosen
9045 compiler. Note that nothing is done to ensure flags work with the
9050 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9052 * Properly check EVP_VerifyFinal() and similar return values
9055 *Ben Laurie, Bodo Moeller, Google Security Team*
9057 * Enable TLS extensions by default.
9061 * Allow the CHIL engine to be loaded, whether the application is
9062 multithreaded or not. (This does not release the developer from the
9063 obligation to set up the dynamic locking callbacks.)
9065 *Sander Temme <sander@temme.net>*
9067 * Use correct exit code if there is an error in dgst command.
9069 *Steve Henson; problem pointed out by Roland Dirlewanger*
9071 * Tweak Configure so that you need to say "experimental-jpake" to enable
9072 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9076 * Add experimental JPAKE support, including demo authentication in
9077 s_client and s_server.
9081 * Set the comparison function in v3_addr_canonize().
9083 *Rob Austein <sra@hactrn.net>*
9085 * Add support for XMPP STARTTLS in s_client.
9087 *Philip Paeps <philip@freebsd.org>*
9089 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9090 to ensure that even with this option, only ciphersuites in the
9091 server's preference list will be accepted. (Note that the option
9092 applies only when resuming a session, so the earlier behavior was
9093 just about the algorithm choice for symmetric cryptography.)
9097 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9099 * Fix NULL pointer dereference if a DTLS server received
9100 ChangeCipherSpec as first record ([CVE-2009-1386]).
9104 * Fix a state transition in s3_srvr.c and d1_srvr.c
9105 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9109 * The fix in 0.9.8c that supposedly got rid of unsafe
9110 double-checked locking was incomplete for RSA blinding,
9111 addressing just one layer of what turns out to have been
9112 doubly unsafe triple-checked locking.
9114 So now fix this for real by retiring the MONT_HELPER macro
9115 in crypto/rsa/rsa_eay.c.
9117 *Bodo Moeller; problem pointed out by Marius Schilder*
9119 * Various precautionary measures:
9121 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9123 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9124 (NB: This would require knowledge of the secret session ticket key
9125 to exploit, in which case you'd be SOL either way.)
9127 - Change bn_nist.c so that it will properly handle input BIGNUMs
9128 outside the expected range.
9130 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9133 *Neel Mehta, Bodo Moeller*
9135 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9136 the load fails. Useful for distros.
9138 *Ben Laurie and the FreeBSD team*
9140 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9144 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9148 * Expand ENGINE to support engine supplied SSL client certificate functions.
9150 This work was sponsored by Logica.
9154 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9155 keystores. Support for SSL/TLS client authentication too.
9156 Not compiled unless enable-capieng specified to Configure.
9158 This work was sponsored by Logica.
9162 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9163 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9164 attribute creation routines such as certificate requests and PKCS#12
9169 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9171 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9172 handshake which could lead to a client crash as found using the
9173 Codenomicon TLS test suite ([CVE-2008-1672])
9175 *Steve Henson, Mark Cox*
9177 * Fix double free in TLS server name extensions which could lead to
9178 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9182 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9184 Clear the error queue to ensure that error entries left from
9185 older function calls do not interfere with the correct operation.
9187 *Lutz Jaenicke, Erik de Castro Lopo*
9189 * Remove root CA certificates of commercial CAs:
9191 The OpenSSL project does not recommend any specific CA and does not
9192 have any policy with respect to including or excluding any CA.
9193 Therefore it does not make any sense to ship an arbitrary selection
9194 of root CA certificates with the OpenSSL software.
9198 * RSA OAEP patches to fix two separate invalid memory reads.
9199 The first one involves inputs when 'lzero' is greater than
9200 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9201 before the beginning of from). The second one involves inputs where
9202 the 'db' section contains nothing but zeroes (there is a one-byte
9203 invalid read after the end of 'db').
9205 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9207 * Partial backport from 0.9.9-dev:
9209 Introduce bn_mul_mont (dedicated Montgomery multiplication
9210 procedure) as a candidate for BIGNUM assembler implementation.
9211 While 0.9.9-dev uses assembler for various architectures, only
9212 x86_64 is available by default here in the 0.9.8 branch, and
9213 32-bit x86 is available through a compile-time setting.
9215 To try the 32-bit x86 assembler implementation, use Configure
9216 option "enable-montasm" (which exists only for this backport).
9218 As "enable-montasm" for 32-bit x86 disclaims code stability
9219 anyway, in this constellation we activate additional code
9220 backported from 0.9.9-dev for further performance improvements,
9221 namely BN_from_montgomery_word. (To enable this otherwise,
9222 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9224 *Andy Polyakov (backport partially by Bodo Moeller)*
9226 * Add TLS session ticket callback. This allows an application to set
9227 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9228 values. This is useful for key rollover for example where several key
9229 sets may exist with different names.
9233 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9234 This was broken until now in 0.9.8 releases, such that the only way
9235 a registered ENGINE could be used (assuming it initialises
9236 successfully on the host) was to explicitly set it as the default
9237 for the relevant algorithms. This is in contradiction with 0.9.7
9238 behaviour and the documentation. With this fix, when an ENGINE is
9239 registered into a given algorithm's table of implementations, the
9240 'uptodate' flag is reset so that auto-discovery will be used next
9241 time a new context for that algorithm attempts to select an
9244 *Ian Lister (tweaked by Geoff Thorpe)*
9246 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9247 implementation in the following ways:
9249 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9252 Lack of BER streaming support means one pass streaming processing is
9253 only supported if data is detached: setting the streaming flag is
9254 ignored for embedded content.
9256 CMS support is disabled by default and must be explicitly enabled
9257 with the enable-cms configuration option.
9261 * Update the GMP engine glue to do direct copies between BIGNUM and
9262 mpz_t when openssl and GMP use the same limb size. Otherwise the
9263 existing "conversion via a text string export" trick is still used.
9265 *Paul Sheer <paulsheer@gmail.com>*
9267 * Zlib compression BIO. This is a filter BIO which compressed and
9268 uncompresses any data passed through it.
9272 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9273 RFC3394 compatible AES key wrapping.
9277 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9278 sets string data without copying. X509_ALGOR_set0() and
9279 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9280 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9281 from an X509_ATTRIBUTE structure optionally checking it occurs only
9282 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9287 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9288 to get the expected BN_FLG_CONSTTIME behavior.
9290 *Bodo Moeller (Google)*
9294 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9295 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9296 - added some more tests to do_tests.pl
9297 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9298 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9299 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9300 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9301 - various changes to netware.pl to enable gcc-cross builds on Win32
9303 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9304 - various changes to fix missing prototype warnings
9305 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9306 - added AES, WHIRLPOOL and CPUID assembler code to build files
9307 - added missing AES assembler make rules to mk1mf.pl
9308 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9310 *Guenter Knauf <eflash@gmx.net>*
9312 * Implement certificate status request TLS extension defined in RFC3546.
9313 A client can set the appropriate parameters and receive the encoded
9314 OCSP response via a callback. A server can query the supplied parameters
9315 and set the encoded OCSP response in the callback. Add simplified examples
9316 to s_client and s_server.
9320 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9323 + Binary incompatibility of ssl_ctx_st structure
9324 + DTLS interoperation with non-compliant servers
9325 + Don't call get_session_cb() without proposed session
9326 + Fix ia64 assembler code
9328 *Andy Polyakov, Steve Henson*
9330 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9332 * DTLS Handshake overhaul. There were longstanding issues with
9333 OpenSSL DTLS implementation, which were making it impossible for
9334 RFC 4347 compliant client to communicate with OpenSSL server.
9335 Unfortunately just fixing these incompatibilities would "cut off"
9336 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9337 server keeps tolerating non RFC compliant syntax. The opposite is
9338 not true, 0.9.8f client can not communicate with earlier server.
9339 This update even addresses CVE-2007-4995.
9343 * Changes to avoid need for function casts in OpenSSL: some compilers
9344 (gcc 4.2 and later) reject their use.
9345 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9348 * Add RFC4507 support to OpenSSL. This includes the corrections in
9349 RFC4507bis. The encrypted ticket format is an encrypted encoded
9350 SSL_SESSION structure, that way new session features are automatically
9353 If a client application caches session in an SSL_SESSION structure
9354 support is transparent because tickets are now stored in the encoded
9357 The SSL_CTX structure automatically generates keys for ticket
9358 protection in servers so again support should be possible
9359 with no application modification.
9361 If a client or server wishes to disable RFC4507 support then the option
9362 SSL_OP_NO_TICKET can be set.
9364 Add a TLS extension debugging callback to allow the contents of any client
9365 or server extensions to be examined.
9367 This work was sponsored by Google.
9371 * Add initial support for TLS extensions, specifically for the server_name
9372 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9373 have new members for a host name. The SSL data structure has an
9374 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9375 stored in that context to allow for session resumption, even after the
9376 SSL has been switched to a new SSL_CTX in reaction to a client's
9377 server_name extension.
9379 New functions (subject to change):
9381 SSL_get_servername()
9382 SSL_get_servername_type()
9385 New CTRL codes and macros (subject to change):
9387 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9388 - SSL_CTX_set_tlsext_servername_callback()
9389 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9390 - SSL_CTX_set_tlsext_servername_arg()
9391 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9393 openssl s_client has a new '-servername ...' option.
9395 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9396 '-key2 ...', '-servername_fatal' (subject to change). This allows
9397 testing the HostName extension for a specific single host name ('-cert'
9398 and '-key' remain fallbacks for handshakes without HostName
9399 negotiation). If the unrecognized_name alert has to be sent, this by
9400 default is a warning; it becomes fatal with the '-servername_fatal'
9403 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9405 * Add AES and SSE2 assembly language support to VC++ build.
9409 * Mitigate attack on final subtraction in Montgomery reduction.
9413 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9414 (which previously caused an internal error).
9418 * Squeeze another 10% out of IGE mode when in != out.
9422 * AES IGE mode speedup.
9424 *Dean Gaudet (Google)*
9426 * Add the Korean symmetric 128-bit cipher SEED (see
9427 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9428 add SEED ciphersuites from RFC 4162:
9430 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9431 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9432 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9433 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9435 To minimize changes between patchlevels in the OpenSSL 0.9.8
9436 series, SEED remains excluded from compilation unless OpenSSL
9437 is configured with 'enable-seed'.
9439 *KISA, Bodo Moeller*
9441 * Mitigate branch prediction attacks, which can be practical if a
9442 single processor is shared, allowing a spy process to extract
9443 information. For detailed background information, see
9444 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9445 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9446 and Necessary Software Countermeasures"). The core of the change
9447 are new versions BN_div_no_branch() and
9448 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9449 respectively, which are slower, but avoid the security-relevant
9450 conditional branches. These are automatically called by BN_div()
9451 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9452 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9453 remove a conditional branch.
9455 BN_FLG_CONSTTIME is the new name for the previous
9456 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9457 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9458 in the exponent causes BN_mod_exp_mont() to use the alternative
9459 implementation in BN_mod_exp_mont_consttime().) The old name
9460 remains as a deprecated alias.
9462 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9463 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9464 constant-time implementations for more than just exponentiation.
9465 Here too the old name is kept as a deprecated alias.
9467 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9468 the BN_BLINDING structure gets an independent copy of the
9469 modulus. This means that the previous `BIGNUM *m` argument to
9470 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9471 essentially becomes `const BIGNUM *m`, although we can't actually
9472 change this in the header file before 0.9.9. It allows
9473 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9474 enable BN_FLG_CONSTTIME.
9476 *Matthew D Wood (Intel Corp)*
9478 * In the SSL/TLS server implementation, be strict about session ID
9479 context matching (which matters if an application uses a single
9480 external cache for different purposes). Previously,
9481 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9482 set. This did ensure strict client verification, but meant that,
9483 with applications using a single external cache for quite
9484 different requirements, clients could circumvent ciphersuite
9485 restrictions for a given session ID context by starting a session
9486 in a different context.
9490 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9491 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9492 authentication-only ciphersuites.
9496 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9497 not complete and could lead to a possible single byte overflow
9498 ([CVE-2007-5135]) [Ben Laurie]
9500 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9502 * Since AES128 and AES256 (and similarly Camellia128 and
9503 Camellia256) share a single mask bit in the logic of
9504 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9505 kludge to work properly if AES128 is available and AES256 isn't
9506 (or if Camellia128 is available and Camellia256 isn't).
9510 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9511 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9512 When a point or a seed is encoded in a BIT STRING, we need to
9513 prevent the removal of trailing zero bits to get the proper DER
9514 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9515 of a NamedBitList, for which trailing 0 bits need to be removed.)
9519 * Have SSL/TLS server implementation tolerate "mismatched" record
9520 protocol version while receiving ClientHello even if the
9521 ClientHello is fragmented. (The server can't insist on the
9522 particular protocol version it has chosen before the ServerHello
9523 message has informed the client about his choice.)
9527 * Add RFC 3779 support.
9529 *Rob Austein for ARIN, Ben Laurie*
9531 * Load error codes if they are not already present instead of using a
9532 static variable. This allows them to be cleanly unloaded and reloaded.
9533 Improve header file function name parsing.
9537 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9538 or CAPABILITY handshake as required by RFCs.
9542 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9544 * Introduce limits to prevent malicious keys being able to
9545 cause a denial of service. ([CVE-2006-2940])
9547 *Steve Henson, Bodo Moeller*
9549 * Fix ASN.1 parsing of certain invalid structures that can result
9550 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9552 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9553 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9555 * Fix SSL client code which could crash if connecting to a
9556 malicious SSLv2 server. ([CVE-2006-4343])
9558 *Tavis Ormandy and Will Drewry, Google Security Team*
9560 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9561 match only those. Before that, "AES256-SHA" would be interpreted
9562 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9563 the same strength classification in 0.9.7h) as we currently only
9564 have a single AES bit in the ciphersuite description bitmap.
9565 That change, however, also applied to ciphersuite strings such as
9566 "RC4-MD5" that intentionally matched multiple ciphersuites --
9567 namely, SSL 2.0 ciphersuites in addition to the more common ones
9568 from SSL 3.0/TLS 1.0.
9570 So we change the selection algorithm again: Naming an explicit
9571 ciphersuite selects this one ciphersuite, and any other similar
9572 ciphersuite (same bitmap) from *other* protocol versions.
9573 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9574 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9576 Since SSL 2.0 does not have any ciphersuites for which the
9577 128/256 bit distinction would be relevant, this works for now.
9578 The proper fix will be to use different bits for AES128 and
9579 AES256, which would have avoided the problems from the beginning;
9580 however, bits are scarce, so we can only do this in a new release
9581 (not just a patchlevel) when we can change the SSL_CIPHER
9582 definition to split the single 'unsigned long mask' bitmap into
9583 multiple values to extend the available space.
9587 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9589 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9590 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9592 * Add AES IGE and biIGE modes.
9596 * Change the Unix randomness entropy gathering to use poll() when
9597 possible instead of select(), since the latter has some
9598 undesirable limitations.
9600 *Darryl Miles via Richard Levitte and Bodo Moeller*
9602 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9603 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9604 cannot be implicitly activated as part of, e.g., the "AES" alias.
9605 However, please upgrade to OpenSSL 0.9.9[-dev] for
9606 non-experimental use of the ECC ciphersuites to get TLS extension
9607 support, which is required for curve and point format negotiation
9608 to avoid potential handshake problems.
9612 * Disable rogue ciphersuites:
9614 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9615 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9616 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9618 The latter two were purportedly from
9619 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9622 Also deactivate the remaining ciphersuites from
9623 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9624 unofficial, and the ID has long expired.
9628 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9629 dual-core machines) and other potential thread-safety issues.
9633 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9634 versions), which is now available for royalty-free use
9635 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9636 Also, add Camellia TLS ciphersuites from RFC 4132.
9638 To minimize changes between patchlevels in the OpenSSL 0.9.8
9639 series, Camellia remains excluded from compilation unless OpenSSL
9640 is configured with 'enable-camellia'.
9644 * Disable the padding bug check when compression is in use. The padding
9645 bug check assumes the first packet is of even length, this is not
9646 necessarily true if compression is enabled and can result in false
9647 positives causing handshake failure. The actual bug test is ancient
9648 code so it is hoped that implementations will either have fixed it by
9649 now or any which still have the bug do not support compression.
9653 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9655 * When applying a cipher rule check to see if string match is an explicit
9656 cipher suite and only match that one cipher suite if it is.
9660 * Link in manifests for VC++ if needed.
9662 *Austin Ziegler <halostatue@gmail.com>*
9664 * Update support for ECC-based TLS ciphersuites according to
9665 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9666 TLS extensions, which are supported starting with the 0.9.9
9667 branch, not in the OpenSSL 0.9.8 branch).
9671 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9672 opaque EVP_CIPHER_CTX handling.
9676 * Fixes and enhancements to zlib compression code. We now only use
9677 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9678 to conform with the standards mentioned here:
9679 <http://www.zlib.net/DLL_FAQ.txt>
9680 Static zlib linking now works on Windows and the new --with-zlib-include
9681 --with-zlib-lib options to Configure can be used to supply the location
9682 of the headers and library. Gracefully handle case where zlib library
9687 * Several fixes and enhancements to the OID generation code. The old code
9688 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9689 handle numbers larger than ULONG_MAX, truncated printing and had a
9690 non standard OBJ_obj2txt() behaviour.
9694 * Add support for building of engines under engine/ as shared libraries
9695 under VC++ build system.
9699 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9700 Hopefully, we will not see any false combination of paths any more.
9704 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9706 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9707 (part of SSL_OP_ALL). This option used to disable the
9708 countermeasure against man-in-the-middle protocol-version
9709 rollback in the SSL 2.0 server implementation, which is a bad
9710 idea. ([CVE-2005-2969])
9712 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9713 for Information Security, National Institute of Advanced Industrial
9714 Science and Technology [AIST], Japan)*
9716 * Add two function to clear and return the verify parameter flags.
9720 * Keep cipherlists sorted in the source instead of sorting them at
9721 runtime, thus removing the need for a lock.
9725 * Avoid some small subgroup attacks in Diffie-Hellman.
9727 *Nick Mathewson and Ben Laurie*
9729 * Add functions for well-known primes.
9733 * Extended Windows CE support.
9735 *Satoshi Nakamura and Andy Polyakov*
9737 * Initialize SSL_METHOD structures at compile time instead of during
9738 runtime, thus removing the need for a lock.
9742 * Make PKCS7_decrypt() work even if no certificate is supplied by
9743 attempting to decrypt each encrypted key in turn. Add support to
9748 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9750 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9753 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9757 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9758 key into the same file any more.
9762 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9766 * Add -utf8 command line and config file option to 'ca'.
9768 *Stefan <stf@udoma.org*
9770 * Removed the macro des_crypt(), as it seems to conflict with some
9771 libraries. Use DES_crypt().
9775 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9776 involves renaming the source and generated shared-libs for
9777 both. The engines will accept the corrected or legacy ids
9778 ('ncipher' and '4758_cca' respectively) when binding. NB,
9779 this only applies when building 'shared'.
9781 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9783 * Add attribute functions to EVP_PKEY structure. Modify
9784 PKCS12_create() to recognize a CSP name attribute and
9785 use it. Make -CSP option work again in pkcs12 utility.
9789 * Add new functionality to the bn blinding code:
9790 - automatic re-creation of the BN_BLINDING parameters after
9791 a fixed number of uses (currently 32)
9792 - add new function for parameter creation
9793 - introduce flags to control the update behaviour of the
9794 BN_BLINDING parameters
9795 - hide BN_BLINDING structure
9796 Add a second BN_BLINDING slot to the RSA structure to improve
9797 performance when a single RSA object is shared among several
9802 * Add support for DTLS.
9804 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9806 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9807 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9811 * Remove buggy and incomplete DH cert support from
9812 ssl/ssl_rsa.c and ssl/s3_both.c
9816 * Use SHA-1 instead of MD5 as the default digest algorithm for
9817 the apps/openssl applications.
9821 * Compile clean with "-Wall -Wmissing-prototypes
9822 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9823 DEBUG_SAFESTACK must also be set.
9827 * Change ./Configure so that certain algorithms can be disabled by default.
9828 The new counterpiece to "no-xxx" is "enable-xxx".
9830 The patented RC5 and MDC2 algorithms will now be disabled unless
9831 "enable-rc5" and "enable-mdc2", respectively, are specified.
9833 (IDEA remains enabled despite being patented. This is because IDEA
9834 is frequently required for interoperability, and there is no license
9835 fee for non-commercial use. As before, "no-idea" can be used to
9836 avoid this algorithm.)
9840 * Add processing of proxy certificates (see RFC 3820). This work was
9841 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9842 EGEE (Enabling Grids for E-science in Europe).
9846 * RC4 performance overhaul on modern architectures/implementations, such
9847 as Intel P4, IA-64 and AMD64.
9851 * New utility extract-section.pl. This can be used specify an alternative
9852 section number in a pod file instead of having to treat each file as
9853 a separate case in Makefile. This can be done by adding two lines to the
9856 =for comment openssl_section:XXX
9858 The blank line is mandatory.
9862 * New arguments -certform, -keyform and -pass for s_client and s_server
9863 to allow alternative format key and certificate files and passphrase
9868 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9869 update associated structures and add various utility functions.
9871 Add new policy related verify parameters, include policy checking in
9872 standard verify code. Enhance 'smime' application with extra parameters
9873 to support policy checking and print out.
9877 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9878 Nehemiah processors. These extensions support AES encryption in hardware
9879 as well as RNG (though RNG support is currently disabled).
9881 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9883 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9887 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9889 *Andy Polyakov and a number of other people*
9891 * Improved PowerPC platform support. Most notably BIGNUM assembler
9892 implementation contributed by IBM.
9894 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9896 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9897 exponent rather than 'unsigned long'. There is a corresponding change to
9898 the new 'rsa_keygen' element of the RSA_METHOD structure.
9900 *Jelte Jansen, Geoff Thorpe*
9902 * Functionality for creating the initial serial number file is now
9903 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9905 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9906 number file to 1, which is bound to cause problems. To avoid
9907 the problems while respecting compatibility between different 0.9.7
9908 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9909 CA.pl for serial number initialization. With the new release 0.9.8,
9910 we can fix the problem directly in the 'ca' utility.)
9914 * Reduced header interdependencies by declaring more opaque objects in
9915 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9916 give fewer recursive includes, which could break lazy source code - so
9917 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9918 developers should define this symbol when building and using openssl to
9919 ensure they track the recommended behaviour, interfaces, [etc], but
9920 backwards-compatible behaviour prevails when this isn't defined.
9924 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9928 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9929 This will generate a random key of the appropriate length based on the
9930 cipher context. The EVP_CIPHER can provide its own random key generation
9931 routine to support keys of a specific form. This is used in the des and
9932 3des routines to generate a key of the correct parity. Update S/MIME
9933 code to use new functions and hence generate correct parity DES keys.
9934 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9935 valid (weak or incorrect parity).
9939 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9940 as looking them up. This is useful when the verified structure may contain
9941 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9942 present unless the new PKCS7_NO_CRL flag is asserted.
9946 * Extend ASN1 oid configuration module. It now additionally accepts the
9949 shortName = some long name, 1.2.3.4
9953 * Reimplemented the BN_CTX implementation. There is now no more static
9954 limitation on the number of variables it can handle nor the depth of the
9955 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9956 information can now expand as required, and rather than having a single
9957 static array of bignums, BN_CTX now uses a linked-list of such arrays
9958 allowing it to expand on demand whilst maintaining the usefulness of
9959 BN_CTX's "bundling".
9963 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9964 to allow all RSA operations to function using a single BN_CTX.
9968 * Preliminary support for certificate policy evaluation and checking. This
9969 is initially intended to pass the tests outlined in "Conformance Testing
9970 of Relying Party Client Certificate Path Processing Logic" v1.07.
9974 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9975 remained unused and not that useful. A variety of other little bignum
9976 tweaks and fixes have also been made continuing on from the audit (see
9981 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9982 associated ASN1, EVP and SSL functions and old ASN1 macros.
9986 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9987 and this should never fail. So the return value from the use of
9988 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9989 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9993 * BN_CTX_get() should return zero-valued bignums, providing the same
9994 initialised value as BN_new().
9996 *Geoff Thorpe, suggested by Ulf Möller*
9998 * Support for inhibitAnyPolicy certificate extension.
10002 * An audit of the BIGNUM code is underway, for which debugging code is
10003 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10004 is considered valid when processing BIGNUMs, and causes execution to
10005 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10006 further steps are taken to deliberately pollute unused data in BIGNUM
10007 structures to try and expose faulty code further on. For now, openssl will
10008 (in its default mode of operation) continue to tolerate the inconsistent
10009 forms that it has tolerated in the past, but authors and packagers should
10010 consider trying openssl and their own applications when compiled with
10011 these debugging symbols defined. It will help highlight potential bugs in
10012 their own code, and will improve the test coverage for OpenSSL itself. At
10013 some point, these tighter rules will become openssl's default to improve
10014 maintainability, though the assert()s and other overheads will remain only
10015 in debugging configurations. See bn.h for more details.
10017 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10019 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10020 that can only be obtained through BN_CTX_new() (which implicitly
10021 initialises it). The presence of this function only made it possible
10022 to overwrite an existing structure (and cause memory leaks).
10026 * Because of the callback-based approach for implementing LHASH as a
10027 template type, lh_insert() adds opaque objects to hash-tables and
10028 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10029 to clean up those corresponding objects before destroying the hash table
10030 (and losing the object pointers). So some over-zealous constifications in
10031 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10032 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10033 prototyped to have "const" restrictions on the object pointers they are
10034 given (and so aren't required to cast them away any more).
10038 * The tmdiff.h API was so ugly and minimal that our own timing utility
10039 (speed) prefers to use its own implementation. The two implementations
10040 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10041 its object type properly exposed (MS_TM) instead of casting to/from
10042 `char *`. This may still change yet if someone realises MS_TM and
10044 aren't necessarily the greatest nomenclatures - but this is what was used
10045 internally to the implementation so I've used that for now.
10049 * Ensure that deprecated functions do not get compiled when
10050 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10051 the self-tests were still using deprecated key-generation functions so
10052 these have been updated also.
10056 * Reorganise PKCS#7 code to separate the digest location functionality
10057 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10058 New function PKCS7_set_digest() to set the digest type for PKCS#7
10059 digestedData type. Add additional code to correctly generate the
10060 digestedData type and add support for this type in PKCS7 initialization
10065 * New function PKCS7_set0_type_other() this initializes a PKCS7
10066 structure of type "other".
10070 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10071 sure the loop does correctly stop and breaking ("division by zero")
10072 modulus operations are not performed. The (pre-generated) prime
10073 table crypto/bn/bn_prime.h was already correct, but it could not be
10074 re-generated on some platforms because of the "division by zero"
10075 situation in the script.
10077 *Ralf S. Engelschall*
10079 * Update support for ECC-based TLS ciphersuites according to
10080 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10081 SHA-1 now is only used for "small" curves (where the
10082 representation of a field element takes up to 24 bytes); for
10083 larger curves, the field element resulting from ECDH is directly
10084 used as premaster secret.
10086 *Douglas Stebila (Sun Microsystems Laboratories)*
10088 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10089 curve secp160r1 to the tests.
10091 *Douglas Stebila (Sun Microsystems Laboratories)*
10093 * Add the possibility to load symbols globally with DSO.
10095 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10097 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10098 control of the error stack.
10102 * Add support for STORE in ENGINE.
10106 * Add the STORE type. The intention is to provide a common interface
10107 to certificate and key stores, be they simple file-based stores, or
10108 HSM-type store, or LDAP stores, or...
10109 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10113 * Add a generic structure called OPENSSL_ITEM. This can be used to
10114 pass a list of arguments to any function as well as provide a way
10115 for a function to pass data back to the caller.
10119 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10120 works like BUF_strdup() but can be used to duplicate a portion of
10121 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10126 * Add the function sk_find_ex() which works like sk_find(), but will
10127 return an index to an element even if an exact match couldn't be
10128 found. The index is guaranteed to point at the element where the
10129 searched-for key would be inserted to preserve sorting order.
10133 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10134 takes an extra flags argument for optional functionality. Currently,
10135 the following flags are defined:
10137 OBJ_BSEARCH_VALUE_ON_NOMATCH
10138 This one gets OBJ_bsearch_ex() to return a pointer to the first
10139 element where the comparing function returns a negative or zero
10142 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10143 This one gets OBJ_bsearch_ex() to return a pointer to the first
10144 element where the comparing function returns zero. This is useful
10145 if there are more than one element where the comparing function
10150 * Make it possible to create self-signed certificates with 'openssl ca'
10151 in such a way that the self-signed certificate becomes part of the
10152 CA database and uses the same mechanisms for serial number generation
10153 as all other certificate signing. The new flag '-selfsign' enables
10154 this functionality. Adapt CA.sh and CA.pl.in.
10158 * Add functionality to check the public key of a certificate request
10159 against a given private. This is useful to check that a certificate
10160 request can be signed by that key (self-signing).
10164 * Make it possible to have multiple active certificates with the same
10165 subject in the CA index file. This is done only if the keyword
10166 'unique_subject' is set to 'no' in the main CA section (default
10167 if 'CA_default') of the configuration file. The value is saved
10168 with the database itself in a separate index attribute file,
10169 named like the index file with '.attr' appended to the name.
10173 * Generate multi-valued AVAs using '+' notation in config files for
10178 * Support for nameConstraints certificate extension.
10182 * Support for policyConstraints certificate extension.
10186 * Support for policyMappings certificate extension.
10190 * Make sure the default DSA_METHOD implementation only uses its
10191 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10192 and change its own handlers to be NULL so as to remove unnecessary
10193 indirection. This lets alternative implementations fallback to the
10194 default implementation more easily.
10198 * Support for directoryName in GeneralName related extensions
10203 * Make it possible to link applications using Makefile.shared.
10204 Make that possible even when linking against static libraries!
10208 * Support for single pass processing for S/MIME signing. This now
10209 means that S/MIME signing can be done from a pipe, in addition
10210 cleartext signing (multipart/signed type) is effectively streaming
10211 and the signed data does not need to be all held in memory.
10213 This is done with a new flag PKCS7_STREAM. When this flag is set
10214 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10215 is done after the data is output (and digests calculated) in
10216 SMIME_write_PKCS7().
10220 * Add full support for -rpath/-R, both in shared libraries and
10221 applications, at least on the platforms where it's known how
10226 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10227 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10228 will now compute a table of multiples of the generator that
10229 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10230 faster (notably in the case of a single point multiplication,
10231 scalar * generator).
10233 *Nils Larsch, Bodo Moeller*
10235 * IPv6 support for certificate extensions. The various extensions
10236 which use the IP:a.b.c.d can now take IPv6 addresses using the
10237 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10242 * Added an ENGINE that implements RSA by performing private key
10243 exponentiations with the GMP library. The conversions to and from
10244 GMP's mpz_t format aren't optimised nor are any montgomery forms
10245 cached, and on x86 it appears OpenSSL's own performance has caught up.
10246 However there are likely to be other architectures where GMP could
10247 provide a boost. This ENGINE is not built in by default, but it can be
10248 specified at Configure time and should be accompanied by the necessary
10249 linker additions, eg;
10250 ./config -DOPENSSL_USE_GMP -lgmp
10254 * "openssl engine" will not display ENGINE/DSO load failure errors when
10255 testing availability of engines with "-t" - the old behaviour is
10256 produced by increasing the feature's verbosity with "-tt".
10260 * ECDSA routines: under certain error conditions uninitialized BN objects
10261 could be freed. Solution: make sure initialization is performed early
10262 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10267 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10268 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10269 software implementations. For DSA and DH, parameter generation can
10270 also be overridden by providing the appropriate method callbacks.
10274 * Change the "progress" mechanism used in key-generation and
10275 primality testing to functions that take a new BN_GENCB pointer in
10276 place of callback/argument pairs. The new API functions have `_ex`
10277 postfixes and the older functions are reimplemented as wrappers for
10278 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10279 declarations of the old functions to help (graceful) attempts to
10280 migrate to the new functions. Also, the new key-generation API
10281 functions operate on a caller-supplied key-structure and return
10282 success/failure rather than returning a key or NULL - this is to
10283 help make "keygen" another member function of RSA_METHOD etc.
10285 Example for using the new callback interface:
10287 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10288 void *my_arg = ...;
10291 BN_GENCB_set(&my_cb, my_callback, my_arg);
10293 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10294 /* For the meaning of a, b in calls to my_callback(), see the
10295 * documentation of the function that calls the callback.
10296 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10297 * my_callback should return 1 if it wants BN_is_prime_ex()
10298 * to continue, or 0 to stop.
10303 * Change the ZLIB compression method to be stateful, and make it
10304 available to TLS with the number defined in
10305 draft-ietf-tls-compression-04.txt.
10309 * Add the ASN.1 structures and functions for CertificatePair, which
10310 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10312 CertificatePair ::= SEQUENCE {
10313 forward [0] Certificate OPTIONAL,
10314 reverse [1] Certificate OPTIONAL,
10315 -- at least one of the pair shall be present -- }
10317 Also implement the PEM functions to read and write certificate
10318 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10320 This needed to be defined, mostly for the sake of the LDAP
10321 attribute crossCertificatePair, but may prove useful elsewhere as
10326 * Make it possible to inhibit symlinking of shared libraries in
10327 Makefile.shared, for Cygwin's sake.
10331 * Extend the BIGNUM API by creating a function
10332 void BN_set_negative(BIGNUM *a, int neg);
10333 and a macro that behave like
10334 int BN_is_negative(const BIGNUM *a);
10336 to avoid the need to access 'a->neg' directly in applications.
10340 * Implement fast modular reduction for pseudo-Mersenne primes
10341 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10342 EC_GROUP_new_curve_GFp() will now automatically use this
10345 *Nils Larsch <nla@trustcenter.de>*
10347 * Add new lock type (CRYPTO_LOCK_BN).
10351 * Change the ENGINE framework to automatically load engines
10352 dynamically from specific directories unless they could be
10353 found to already be built in or loaded. Move all the
10354 current engines except for the cryptodev one to a new
10355 directory engines/.
10356 The engines in engines/ are built as shared libraries if
10357 the "shared" options was given to ./Configure or ./config.
10358 Otherwise, they are inserted in libcrypto.a.
10359 /usr/local/ssl/engines is the default directory for dynamic
10360 engines, but that can be overridden at configure time through
10361 the usual use of --prefix and/or --openssldir, and at run
10362 time with the environment variable OPENSSL_ENGINES.
10364 *Geoff Thorpe and Richard Levitte*
10366 * Add Makefile.shared, a helper makefile to build shared
10367 libraries. Adapt Makefile.org.
10371 * Add version info to Win32 DLLs.
10373 *Peter 'Luna' Runestig" <peter@runestig.com>*
10375 * Add new 'medium level' PKCS#12 API. Certificates and keys
10376 can be added using this API to created arbitrary PKCS#12
10377 files while avoiding the low level API.
10379 New options to PKCS12_create(), key or cert can be NULL and
10380 will then be omitted from the output file. The encryption
10381 algorithm NIDs can be set to -1 for no encryption, the mac
10382 iteration count can be set to 0 to omit the mac.
10384 Enhance pkcs12 utility by making the -nokeys and -nocerts
10385 options work when creating a PKCS#12 file. New option -nomac
10386 to omit the mac, NONE can be set for an encryption algorithm.
10387 New code is modified to use the enhanced PKCS12_create()
10388 instead of the low level API.
10392 * Extend ASN1 encoder to support indefinite length constructed
10393 encoding. This can output sequences tags and octet strings in
10394 this form. Modify pk7_asn1.c to support indefinite length
10395 encoding. This is experimental and needs additional code to
10396 be useful, such as an ASN1 bio and some enhanced streaming
10399 Extend template encode functionality so that tagging is passed
10400 down to the template encoder.
10404 * Let 'openssl req' fail if an argument to '-newkey' is not
10405 recognized instead of using RSA as a default.
10409 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10410 As these are not official, they are not included in "ALL";
10411 the "ECCdraft" ciphersuite group alias can be used to select them.
10413 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10415 * Add ECDH engine support.
10417 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10419 * Add ECDH in new directory crypto/ecdh/.
10421 *Douglas Stebila (Sun Microsystems Laboratories)*
10423 * Let BN_rand_range() abort with an error after 100 iterations
10424 without success (which indicates a broken PRNG).
10428 * Change BN_mod_sqrt() so that it verifies that the input value
10429 is really the square of the return value. (Previously,
10430 BN_mod_sqrt would show GIGO behaviour.)
10434 * Add named elliptic curves over binary fields from X9.62, SECG,
10435 and WAP/WTLS; add OIDs that were still missing.
10437 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10439 * Extend the EC library for elliptic curves over binary fields
10440 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10443 EC_GF2m_simple_method
10447 EC_GROUP_new_curve_GF2m
10448 EC_GROUP_set_curve_GF2m
10449 EC_GROUP_get_curve_GF2m
10450 EC_POINT_set_affine_coordinates_GF2m
10451 EC_POINT_get_affine_coordinates_GF2m
10452 EC_POINT_set_compressed_coordinates_GF2m
10454 Point compression for binary fields is disabled by default for
10455 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10458 As binary polynomials are represented as BIGNUMs, various members
10459 of the EC_GROUP and EC_POINT data structures can be shared
10460 between the implementations for prime fields and binary fields;
10461 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10462 are essentially identical to their `..._GFp` counterparts.
10463 (For simplicity, the `..._GFp` prefix has been dropped from
10464 various internal method names.)
10466 An internal 'field_div' method (similar to 'field_mul' and
10467 'field_sqr') has been added; this is used only for binary fields.
10469 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10471 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10472 through methods ('mul', 'precompute_mult').
10474 The generic implementations (now internally called 'ec_wNAF_mul'
10475 and 'ec_wNAF_precomputed_mult') remain the default if these
10476 methods are undefined.
10478 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10480 * New function EC_GROUP_get_degree, which is defined through
10481 EC_METHOD. For curves over prime fields, this returns the bit
10482 length of the modulus.
10484 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10486 * New functions EC_GROUP_dup, EC_POINT_dup.
10487 (These simply call ..._new and ..._copy).
10489 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10491 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10492 Polynomials are represented as BIGNUMs (where the sign bit is not
10493 used) in the following functions [macros]:
10496 BN_GF2m_sub [= BN_GF2m_add]
10497 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10498 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10499 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10501 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10502 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10503 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10504 BN_GF2m_cmp [= BN_ucmp]
10506 (Note that only the 'mod' functions are actually for fields GF(2^m).
10507 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10509 For some functions, an the irreducible polynomial defining a
10510 field can be given as an 'unsigned int[]' with strictly
10511 decreasing elements giving the indices of those bits that are set;
10512 i.e., p[] represents the polynomial
10513 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10515 p[0] > p[1] > ... > p[k] = 0.
10516 This applies to the following functions:
10519 BN_GF2m_mod_mul_arr
10520 BN_GF2m_mod_sqr_arr
10521 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10522 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10523 BN_GF2m_mod_exp_arr
10524 BN_GF2m_mod_sqrt_arr
10525 BN_GF2m_mod_solve_quad_arr
10529 Conversion can be performed by the following functions:
10534 bntest.c has additional tests for binary polynomial arithmetic.
10536 Two implementations for BN_GF2m_mod_div() are available.
10537 The default algorithm simply uses BN_GF2m_mod_inv() and
10538 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10539 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10540 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10542 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10544 * Add new error code 'ERR_R_DISABLED' that can be used when some
10545 functionality is disabled at compile-time.
10547 *Douglas Stebila <douglas.stebila@sun.com>*
10549 * Change default behaviour of 'openssl asn1parse' so that more
10550 information is visible when viewing, e.g., a certificate:
10552 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10553 mode the content of non-printable OCTET STRINGs is output in a
10554 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10555 avoid the appearance of a printable string.
10557 *Nils Larsch <nla@trustcenter.de>*
10559 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10561 EC_GROUP_set_asn1_flag()
10562 EC_GROUP_get_asn1_flag()
10563 EC_GROUP_set_point_conversion_form()
10564 EC_GROUP_get_point_conversion_form()
10565 These control ASN1 encoding details:
10566 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10567 has been set to OPENSSL_EC_NAMED_CURVE.
10568 - Points are encoded in uncompressed form by default; options for
10569 asn1_for are as for point2oct, namely
10570 POINT_CONVERSION_COMPRESSED
10571 POINT_CONVERSION_UNCOMPRESSED
10572 POINT_CONVERSION_HYBRID
10574 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10576 EC_GROUP_set_seed()
10577 EC_GROUP_get0_seed()
10578 EC_GROUP_get_seed_len()
10579 This is used only for ASN1 purposes (so far).
10581 *Nils Larsch <nla@trustcenter.de>*
10583 * Add 'field_type' member to EC_METHOD, which holds the NID
10584 of the appropriate field type OID. The new function
10585 EC_METHOD_get_field_type() returns this value.
10587 *Nils Larsch <nla@trustcenter.de>*
10590 EC_POINT_point2bn()
10591 EC_POINT_bn2point()
10592 EC_POINT_point2hex()
10593 EC_POINT_hex2point()
10594 providing useful interfaces to EC_POINT_point2oct() and
10595 EC_POINT_oct2point().
10597 *Nils Larsch <nla@trustcenter.de>*
10599 * Change internals of the EC library so that the functions
10600 EC_GROUP_set_generator()
10601 EC_GROUP_get_generator()
10602 EC_GROUP_get_order()
10603 EC_GROUP_get_cofactor()
10604 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10605 to methods, which would lead to unnecessary code duplication when
10606 adding different types of curves.
10608 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10610 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10611 arithmetic, and such that modified wNAFs are generated
10612 (which avoid length expansion in many cases).
10616 * Add a function EC_GROUP_check_discriminant() (defined via
10617 EC_METHOD) that verifies that the curve discriminant is non-zero.
10619 Add a function EC_GROUP_check() that makes some sanity tests
10620 on a EC_GROUP, its generator and order. This includes
10621 EC_GROUP_check_discriminant().
10623 *Nils Larsch <nla@trustcenter.de>*
10625 * Add ECDSA in new directory crypto/ecdsa/.
10627 Add applications 'openssl ecparam' and 'openssl ecdsa'
10628 (these are based on 'openssl dsaparam' and 'openssl dsa').
10630 ECDSA support is also included in various other files across the
10631 library. Most notably,
10632 - 'openssl req' now has a '-newkey ecdsa:file' option;
10633 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10634 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10635 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10636 them suitable for ECDSA where domain parameters must be
10637 extracted before the specific public key;
10638 - ECDSA engine support has been added.
10640 *Nils Larsch <nla@trustcenter.de>*
10642 * Include some named elliptic curves, and add OIDs from X9.62,
10643 SECG, and WAP/WTLS. Each curve can be obtained from the new
10645 EC_GROUP_new_by_curve_name(),
10646 and the list of available named curves can be obtained with
10647 EC_get_builtin_curves().
10648 Also add a 'curve_name' member to EC_GROUP objects, which can be
10650 EC_GROUP_set_curve_name()
10651 EC_GROUP_get_curve_name()
10653 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10655 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10656 was actually never needed) and in BN_mul(). The removal in BN_mul()
10657 required a small change in bn_mul_part_recursive() and the addition
10658 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10659 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10660 bn_sub_words() and bn_add_words() except they take arrays with
10665 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10667 * Cleanse PEM buffers before freeing them since they may contain
10670 *Benjamin Bennett <ben@psc.edu>*
10672 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10673 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10674 authentication-only ciphersuites.
10678 * Since AES128 and AES256 share a single mask bit in the logic of
10679 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10680 kludge to work properly if AES128 is available and AES256 isn't.
10684 * Expand security boundary to match 1.1.1 module.
10688 * Remove redundant features: hash file source, editing of test vectors
10689 modify fipsld to use external fips_premain.c signature.
10693 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10694 run algorithm test programs.
10698 * Make algorithm test programs more tolerant of whitespace.
10702 * Have SSL/TLS server implementation tolerate "mismatched" record
10703 protocol version while receiving ClientHello even if the
10704 ClientHello is fragmented. (The server can't insist on the
10705 particular protocol version it has chosen before the ServerHello
10706 message has informed the client about his choice.)
10710 * Load error codes if they are not already present instead of using a
10711 static variable. This allows them to be cleanly unloaded and reloaded.
10715 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10717 * Introduce limits to prevent malicious keys being able to
10718 cause a denial of service. ([CVE-2006-2940])
10720 *Steve Henson, Bodo Moeller*
10722 * Fix ASN.1 parsing of certain invalid structures that can result
10723 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10725 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10726 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10728 * Fix SSL client code which could crash if connecting to a
10729 malicious SSLv2 server. ([CVE-2006-4343])
10731 *Tavis Ormandy and Will Drewry, Google Security Team*
10733 * Change ciphersuite string processing so that an explicit
10734 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10735 will no longer include "AES128-SHA"), and any other similar
10736 ciphersuite (same bitmap) from *other* protocol versions (so that
10737 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10738 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10739 changes from 0.9.8b and 0.9.8d.
10743 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10745 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10746 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10748 * Change the Unix randomness entropy gathering to use poll() when
10749 possible instead of select(), since the latter has some
10750 undesirable limitations.
10752 *Darryl Miles via Richard Levitte and Bodo Moeller*
10754 * Disable rogue ciphersuites:
10756 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10757 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10758 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10760 The latter two were purportedly from
10761 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10764 Also deactivate the remaining ciphersuites from
10765 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10766 unofficial, and the ID has long expired.
10770 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10771 dual-core machines) and other potential thread-safety issues.
10775 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10777 * Adapt fipsld and the build system to link against the validated FIPS
10778 module in FIPS mode.
10782 * Fixes for VC++ 2005 build under Windows.
10786 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10787 from a Windows bash shell such as MSYS. It is autodetected from the
10788 "config" script when run from a VC++ environment. Modify standard VC++
10789 build to use fipscanister.o from the GNU make build.
10793 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10795 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10796 The value now differs depending on if you build for FIPS or not.
10797 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10798 safely run with a non-FIPSed libcrypto, as it may crash because of
10799 the difference induced by this change.
10803 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10805 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10806 (part of SSL_OP_ALL). This option used to disable the
10807 countermeasure against man-in-the-middle protocol-version
10808 rollback in the SSL 2.0 server implementation, which is a bad
10809 idea. ([CVE-2005-2969])
10811 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10812 for Information Security, National Institute of Advanced Industrial
10813 Science and Technology [AIST, Japan)]*
10815 * Minimal support for X9.31 signatures and PSS padding modes. This is
10816 mainly for FIPS compliance and not fully integrated at this stage.
10820 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10821 the exponentiation using a fixed-length exponent. (Otherwise,
10822 the information leaked through timing could expose the secret key
10823 after many signatures; cf. Bleichenbacher's attack on DSA with
10828 * Make a new fixed-window mod_exp implementation the default for
10829 RSA, DSA, and DH private-key operations so that the sequence of
10830 squares and multiplies and the memory access pattern are
10831 independent of the particular secret key. This will mitigate
10832 cache-timing and potential related attacks.
10834 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10835 and this is automatically used by BN_mod_exp_mont() if the new flag
10836 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10837 will use this BN flag for private exponents unless the flag
10838 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10839 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10841 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10843 * Change the client implementation for SSLv23_method() and
10844 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10845 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10846 (Previously, the SSL 2.0 backwards compatible Client Hello
10847 message format would be used even with SSL_OP_NO_SSLv2.)
10851 * Add support for smime-type MIME parameter in S/MIME messages which some
10856 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10857 a threadsafe manner. Modify rsa code to use new function and add calls
10858 to dsa and dh code (which had race conditions before).
10862 * Include the fixed error library code in the C error file definitions
10863 instead of fixing them up at runtime. This keeps the error code
10864 structures constant.
10868 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10870 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10873 * Fixes for newer kerberos headers. NB: the casts are needed because
10874 the 'length' field is signed on one version and unsigned on another
10875 with no (?) obvious way to tell the difference, without these VC++
10876 complains. Also the "definition" of FAR (blank) is no longer included
10877 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10878 some needed definitions.
10882 * Undo Cygwin change.
10886 * Added support for proxy certificates according to RFC 3820.
10887 Because they may be a security thread to unaware applications,
10888 they must be explicitly allowed in run-time. See
10889 docs/HOWTO/proxy_certificates.txt for further information.
10893 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10895 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10896 server and client random values. Previously
10897 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10898 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10900 This change has negligible security impact because:
10902 1. Server and client random values still have 24 bytes of pseudo random
10905 2. Server and client random values are sent in the clear in the initial
10908 3. The master secret is derived using the premaster secret (48 bytes in
10909 size for static RSA ciphersuites) as well as client server and random
10912 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10915 *Stephen Henson, reported by UK NISCC*
10917 * Use Windows randomness collection on Cygwin.
10921 * Fix hang in EGD/PRNGD query when communication socket is closed
10922 prematurely by EGD/PRNGD.
10924 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10926 * Prompt for pass phrases when appropriate for PKCS12 input format.
10930 * Back-port of selected performance improvements from development
10931 branch, as well as improved support for PowerPC platforms.
10935 * Add lots of checks for memory allocation failure, error codes to indicate
10936 failure and freeing up memory if a failure occurs.
10938 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10940 * Add new -passin argument to dgst.
10944 * Perform some character comparisons of different types in X509_NAME_cmp:
10945 this is needed for some certificates that re-encode DNs into UTF8Strings
10946 (in violation of RFC3280) and can't or won't issue name rollover
10951 * Make an explicit check during certificate validation to see that
10952 the CA setting in each certificate on the chain is correct. As a
10953 side effect always do the following basic checks on extensions,
10954 not just when there's an associated purpose to the check:
10956 - if there is an unhandled critical extension (unless the user
10957 has chosen to ignore this fault)
10958 - if the path length has been exceeded (if one is set at all)
10959 - that certain extensions fit the associated purpose (if one has
10964 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10966 * Avoid a race condition when CRLs are checked in a multi threaded
10967 environment. This would happen due to the reordering of the revoked
10968 entries during signature checking and serial number lookup. Now the
10969 encoding is cached and the serial number sort performed under a lock.
10970 Add new STACK function sk_is_sorted().
10974 * Add Delta CRL to the extension code.
10978 * Various fixes to s3_pkt.c so alerts are sent properly.
10980 *David Holmes <d.holmes@f5.com>*
10982 * Reduce the chances of duplicate issuer name and serial numbers (in
10983 violation of RFC3280) using the OpenSSL certificate creation utilities.
10984 This is done by creating a random 64 bit value for the initial serial
10985 number when a serial number file is created or when a self signed
10986 certificate is created using 'openssl req -x509'. The initial serial
10987 number file is created using 'openssl x509 -next_serial' in CA.pl
10988 rather than being initialized to 1.
10992 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10994 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10995 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
10997 *Joe Orton, Steve Henson*
10999 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11002 *Joe Orton, Steve Henson*
11004 * Make it possible to have multiple active certificates with the same
11005 subject in the CA index file. This is done only if the keyword
11006 'unique_subject' is set to 'no' in the main CA section (default
11007 if 'CA_default') of the configuration file. The value is saved
11008 with the database itself in a separate index attribute file,
11009 named like the index file with '.attr' appended to the name.
11013 * X509 verify fixes. Disable broken certificate workarounds when
11014 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11015 keyUsage extension present. Don't accept CRLs with unhandled critical
11016 extensions: since verify currently doesn't process CRL extensions this
11017 rejects a CRL with *any* critical extensions. Add new verify error codes
11022 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11023 A clarification of RFC2560 will require the use of OCTET STRINGs and
11024 some implementations cannot handle the current raw format. Since OpenSSL
11025 copies and compares OCSP nonces as opaque blobs without any attempt at
11026 parsing them this should not create any compatibility issues.
11030 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11031 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11032 this HMAC (and other) operations are several times slower than OpenSSL
11037 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11039 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11041 * Use the correct content when signing type "other".
11045 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11047 * Fix various bugs revealed by running the NISCC test suite:
11049 Stop out of bounds reads in the ASN1 code when presented with
11050 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11052 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11054 If verify callback ignores invalid public key errors don't try to check
11055 certificate signature with the NULL public key.
11059 * New -ignore_err option in ocsp application to stop the server
11060 exiting on the first error in a request.
11064 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11065 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11070 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11071 extra data after the compression methods not only for TLS 1.0
11072 but also for SSL 3.0 (as required by the specification).
11074 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11076 * Change X509_certificate_type() to mark the key as exported/exportable
11077 when it's 512 *bits* long, not 512 bytes.
11081 * Change AES_cbc_encrypt() so it outputs exact multiple of
11082 blocks during encryption.
11086 * Various fixes to base64 BIO and non blocking I/O. On write
11087 flushes were not handled properly if the BIO retried. On read
11088 data was not being buffered properly and had various logic bugs.
11089 This also affects blocking I/O when the data being decoded is a
11094 * Various S/MIME bugfixes and compatibility changes:
11095 output correct application/pkcs7 MIME type if
11096 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11097 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11098 of files as .eml work). Correctly handle very long lines in MIME
11103 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11105 * Countermeasure against the Klima-Pokorny-Rosa extension of
11106 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11107 a protocol version number mismatch like a decryption error
11108 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11112 * Turn on RSA blinding by default in the default implementation
11113 to avoid a timing attack. Applications that don't want it can call
11114 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11115 They would be ill-advised to do so in most cases.
11117 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11119 * Change RSA blinding code so that it works when the PRNG is not
11120 seeded (in this case, the secret RSA exponent is abused as
11121 an unpredictable seed -- if it is not unpredictable, there
11122 is no point in blinding anyway). Make RSA blinding thread-safe
11123 by remembering the creator's thread ID in rsa->blinding and
11124 having all other threads use local one-time blinding factors
11125 (this requires more computation than sharing rsa->blinding, but
11126 avoids excessive locking; and if an RSA object is not shared
11127 between threads, blinding will still be very fast).
11131 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11132 ENGINE as defaults for all supported algorithms irrespective of
11133 the 'flags' parameter. 'flags' is now honoured, so applications
11134 should make sure they are passing it correctly.
11138 * Target "mingw" now allows native Windows code to be generated in
11139 the Cygwin environment as well as with the MinGW compiler.
11143 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11145 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11146 via timing by performing a MAC computation even if incorrect
11147 block cipher padding has been found. This is a countermeasure
11148 against active attacks where the attacker has to distinguish
11149 between bad padding and a MAC verification error. ([CVE-2003-0078])
11151 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11152 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11153 Martin Vuagnoux (EPFL, Ilion)*
11155 * Make the no-err option work as intended. The intention with no-err
11156 is not to have the whole error stack handling routines removed from
11157 libcrypto, it's only intended to remove all the function name and
11158 reason texts, thereby removing some of the footprint that may not
11159 be interesting if those errors aren't displayed anyway.
11161 NOTE: it's still possible for any application or module to have its
11162 own set of error texts inserted. The routines are there, just not
11163 used by default when no-err is given.
11167 * Add support for FreeBSD on IA64.
11169 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11171 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11172 Kerberos function mit_des_cbc_cksum(). Before this change,
11173 the value returned by DES_cbc_cksum() was like the one from
11174 mit_des_cbc_cksum(), except the bytes were swapped.
11176 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11178 * Allow an application to disable the automatic SSL chain building.
11179 Before this a rather primitive chain build was always performed in
11180 ssl3_output_cert_chain(): an application had no way to send the
11181 correct chain if the automatic operation produced an incorrect result.
11183 Now the chain builder is disabled if either:
11185 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11187 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11189 The reasoning behind this is that an application would not want the
11190 auto chain building to take place if extra chain certificates are
11191 present and it might also want a means of sending no additional
11192 certificates (for example the chain has two certificates and the
11197 * Add the possibility to build without the ENGINE framework.
11199 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11201 * Under Win32 gmtime() can return NULL: check return value in
11202 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11206 * DSA routines: under certain error conditions uninitialized BN objects
11207 could be freed. Solution: make sure initialization is performed early
11208 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11209 Nils Larsch <nla@trustcenter.de> via PR#459)
11213 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11214 checked on reconnect on the client side, therefore session resumption
11215 could still fail with a "ssl session id is different" error. This
11216 behaviour is masked when SSL_OP_ALL is used due to
11217 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11218 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11219 followup to PR #377.
11223 * IA-32 assembler support enhancements: unified ELF targets, support
11224 for SCO/Caldera platforms, fix for Cygwin shared build.
11228 * Add support for FreeBSD on sparc64. As a consequence, support for
11229 FreeBSD on non-x86 processors is separate from x86 processors on
11230 the config script, much like the NetBSD support.
11232 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11234 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11236 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11239 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11240 code (06) was taken as the first octet of the session ID and the last
11241 octet was ignored consequently. As a result SSLv2 client side session
11242 caching could not have worked due to the session ID mismatch between
11244 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11249 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11250 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11255 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11256 seems that in spite of existing for more than a year, many application
11257 author have done nothing to provide the necessary callbacks, which
11258 means that this particular engine will not work properly anywhere.
11259 This is a very unfortunate situation which forces us, in the name
11260 of usability, to give the hw_ncipher.c a static lock, which is part
11262 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11263 appear in 0.9.8 or later. We EXPECT application authors to have
11264 dealt properly with this when 0.9.8 is released (unless we actually
11265 make such changes in the libcrypto locking code that changes will
11266 have to be made anyway).
11270 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11271 octets have been read, EOF or an error occurs. Without this change
11272 some truncated ASN1 structures will not produce an error.
11276 * Disable Heimdal support, since it hasn't been fully implemented.
11277 Still give the possibility to force the use of Heimdal, but with
11278 warnings and a request that patches get sent to openssl-dev.
11282 * Add the VC-CE target, introduce the WINCE sysname, and add
11283 INSTALL.WCE and appropriate conditionals to make it build.
11285 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11287 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11288 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11289 edit numbers of the version.
11291 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11293 * Introduce safe string copy and catenation functions
11294 (BUF_strlcpy() and BUF_strlcat()).
11296 *Ben Laurie (CHATS) and Richard Levitte*
11298 * Avoid using fixed-size buffers for one-line DNs.
11300 *Ben Laurie (CHATS)*
11302 * Add BUF_MEM_grow_clean() to avoid information leakage when
11303 resizing buffers containing secrets, and use where appropriate.
11305 *Ben Laurie (CHATS)*
11307 * Avoid using fixed size buffers for configuration file location.
11309 *Ben Laurie (CHATS)*
11311 * Avoid filename truncation for various CA files.
11313 *Ben Laurie (CHATS)*
11315 * Use sizeof in preference to magic numbers.
11317 *Ben Laurie (CHATS)*
11319 * Avoid filename truncation in cert requests.
11321 *Ben Laurie (CHATS)*
11323 * Add assertions to check for (supposedly impossible) buffer
11326 *Ben Laurie (CHATS)*
11328 * Don't cache truncated DNS entries in the local cache (this could
11329 potentially lead to a spoofing attack).
11331 *Ben Laurie (CHATS)*
11333 * Fix various buffers to be large enough for hex/decimal
11334 representations in a platform independent manner.
11336 *Ben Laurie (CHATS)*
11338 * Add CRYPTO_realloc_clean() to avoid information leakage when
11339 resizing buffers containing secrets, and use where appropriate.
11341 *Ben Laurie (CHATS)*
11343 * Add BIO_indent() to avoid much slightly worrying code to do
11346 *Ben Laurie (CHATS)*
11348 * Convert sprintf()/BIO_puts() to BIO_printf().
11350 *Ben Laurie (CHATS)*
11352 * buffer_gets() could terminate with the buffer only half
11355 *Ben Laurie (CHATS)*
11357 * Add assertions to prevent user-supplied crypto functions from
11358 overflowing internal buffers by having large block sizes, etc.
11360 *Ben Laurie (CHATS)*
11362 * New OPENSSL_assert() macro (similar to assert(), but enabled
11365 *Ben Laurie (CHATS)*
11367 * Eliminate unused copy of key in RC4.
11369 *Ben Laurie (CHATS)*
11371 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11373 *Ben Laurie (CHATS)*
11375 * Fix off-by-one error in EGD path.
11377 *Ben Laurie (CHATS)*
11379 * If RANDFILE path is too long, ignore instead of truncating.
11381 *Ben Laurie (CHATS)*
11383 * Eliminate unused and incorrectly sized X.509 structure
11386 *Ben Laurie (CHATS)*
11388 * Eliminate unused and dangerous function knumber().
11390 *Ben Laurie (CHATS)*
11392 * Eliminate unused and dangerous structure, KSSL_ERR.
11394 *Ben Laurie (CHATS)*
11396 * Protect against overlong session ID context length in an encoded
11397 session object. Since these are local, this does not appear to be
11400 *Ben Laurie (CHATS)*
11402 * Change from security patch (see 0.9.6e below) that did not affect
11403 the 0.9.6 release series:
11405 Remote buffer overflow in SSL3 protocol - an attacker could
11406 supply an oversized master key in Kerberos-enabled versions.
11409 *Ben Laurie (CHATS)*
11411 * Change the SSL kerb5 codes to match RFC 2712.
11415 * Make -nameopt work fully for req and add -reqopt switch.
11417 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11419 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11421 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11423 * Make sure tests can be performed even if the corresponding algorithms
11424 have been removed entirely. This was also the last step to make
11425 OpenSSL compilable with DJGPP under all reasonable conditions.
11427 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11429 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11430 to allow version independent disabling of normally unselected ciphers,
11431 which may be activated as a side-effect of selecting a single cipher.
11433 (E.g., cipher list string "RSA" enables ciphersuites that are left
11434 out of "ALL" because they do not provide symmetric encryption.
11435 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11437 *Lutz Jaenicke, Bodo Moeller*
11439 * Add appropriate support for separate platform-dependent build
11440 directories. The recommended way to make a platform-dependent
11441 build directory is the following (tested on Linux), maybe with
11444 # Place yourself outside of the OpenSSL source tree. In
11445 # this example, the environment variable OPENSSL_SOURCE
11446 # is assumed to contain the absolute OpenSSL source directory.
11447 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11448 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11449 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11450 mkdir -p `dirname $F`
11451 ln -s $OPENSSL_SOURCE/$F $F
11454 To be absolutely sure not to disturb the source tree, a "make clean"
11455 is a good thing. If it isn't successful, don't worry about it,
11456 it probably means the source directory is very clean.
11460 * Make sure any ENGINE control commands make local copies of string
11461 pointers passed to them whenever necessary. Otherwise it is possible
11462 the caller may have overwritten (or deallocated) the original string
11463 data when a later ENGINE operation tries to use the stored values.
11465 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11467 * Improve diagnostics in file reading and command-line digests.
11469 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11471 * Add AES modes CFB and OFB to the object database. Correct an
11472 error in AES-CFB decryption.
11476 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11477 allows existing EVP_CIPHER_CTX structures to be reused after
11478 calling `EVP_*Final()`. This behaviour is used by encryption
11479 BIOs and some applications. This has the side effect that
11480 applications must explicitly clean up cipher contexts with
11481 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11485 * Check the values of dna and dnb in bn_mul_recursive before calling
11486 bn_mul_comba (a non zero value means the a or b arrays do not contain
11487 n2 elements) and fallback to bn_mul_normal if either is not zero.
11491 * Fix escaping of non-ASCII characters when using the -subj option
11492 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11496 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11497 form for "surname", serialNumber has no short form.
11498 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11499 therefore remove "mail" short name for "internet 7".
11500 The OID for unique identifiers in X509 certificates is
11501 x500UniqueIdentifier, not uniqueIdentifier.
11502 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11506 * Add an "init" command to the ENGINE config module and auto initialize
11507 ENGINEs. Without any "init" command the ENGINE will be initialized
11508 after all ctrl commands have been executed on it. If init=1 the
11509 ENGINE is initialized at that point (ctrls before that point are run
11510 on the uninitialized ENGINE and after on the initialized one). If
11511 init=0 then the ENGINE will not be initialized at all.
11515 * Fix the 'app_verify_callback' interface so that the user-defined
11516 argument is actually passed to the callback: In the
11517 SSL_CTX_set_cert_verify_callback() prototype, the callback
11518 declaration has been changed from
11521 int (*cb)(X509_STORE_CTX *,void *);
11522 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11523 i=s->ctx->app_verify_callback(&ctx)
11524 has been changed into
11525 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11527 To update applications using SSL_CTX_set_cert_verify_callback(),
11528 a dummy argument can be added to their callback functions.
11530 *D. K. Smetters <smetters@parc.xerox.com>*
11532 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11534 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11536 * Add and OPENSSL_LOAD_CONF define which will cause
11537 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11538 This allows older applications to transparently support certain
11539 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11540 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11541 load the config file and OPENSSL_add_all_algorithms_conf() which will
11542 always load it have also been added.
11546 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11547 Adjust NIDs and EVP layer.
11549 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11551 * Config modules support in openssl utility.
11553 Most commands now load modules from the config file,
11554 though in a few (such as version) this isn't done
11555 because it couldn't be used for anything.
11557 In the case of ca and req the config file used is
11558 the same as the utility itself: that is the -config
11559 command line option can be used to specify an
11564 * Move default behaviour from OPENSSL_config(). If appname is NULL
11565 use "openssl_conf" if filename is NULL use default openssl config file.
11569 * Add an argument to OPENSSL_config() to allow the use of an alternative
11570 config section name. Add a new flag to tolerate a missing config file
11571 and move code to CONF_modules_load_file().
11575 * Support for crypto accelerator cards from Accelerated Encryption
11576 Processing, www.aep.ie. (Use engine 'aep')
11577 The support was copied from 0.9.6c [engine] and adapted/corrected
11578 to work with the new engine framework.
11580 *AEP Inc. and Richard Levitte*
11582 * Support for SureWare crypto accelerator cards from Baltimore
11583 Technologies. (Use engine 'sureware')
11584 The support was copied from 0.9.6c [engine] and adapted
11585 to work with the new engine framework.
11589 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11590 make the newer ENGINE framework commands for the CHIL engine work.
11592 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11594 * Make it possible to produce shared libraries on ReliantUNIX.
11596 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11598 * Add the configuration target debug-linux-ppro.
11599 Make 'openssl rsa' use the general key loading routines
11600 implemented in apps.c, and make those routines able to
11601 handle the key format FORMAT_NETSCAPE and the variant
11604 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11606 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11608 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11610 * Add -keyform to rsautl, and document -engine.
11612 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11614 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11615 BIO_R_NO_SUCH_FILE error code rather than the generic
11616 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11620 * Add new functions
11621 ERR_peek_last_error
11622 ERR_peek_last_error_line
11623 ERR_peek_last_error_line_data.
11624 These are similar to
11626 ERR_peek_error_line
11627 ERR_peek_error_line_data,
11628 but report on the latest error recorded rather than the first one
11629 still in the error queue.
11631 *Ben Laurie, Bodo Moeller*
11633 * default_algorithms option in ENGINE config module. This allows things
11635 default_algorithms = ALL
11636 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11640 * Preliminary ENGINE config module.
11644 * New experimental application configuration code.
11648 * Change the AES code to follow the same name structure as all other
11649 symmetric ciphers, and behave the same way. Move everything to
11650 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11652 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11654 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11656 *Ben Laurie and Theo de Raadt*
11658 * Add option to output public keys in req command.
11660 *Massimiliano Pala madwolf@openca.org*
11662 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11663 (up to about 10% better than before for P-192 and P-224).
11667 * New functions/macros
11669 SSL_CTX_set_msg_callback(ctx, cb)
11670 SSL_CTX_set_msg_callback_arg(ctx, arg)
11671 SSL_set_msg_callback(ssl, cb)
11672 SSL_set_msg_callback_arg(ssl, arg)
11674 to request calling a callback function
11676 void cb(int write_p, int version, int content_type,
11677 const void *buf, size_t len, SSL *ssl, void *arg)
11679 whenever a protocol message has been completely received
11680 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11681 protocol version according to which the SSL library interprets
11682 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11683 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11684 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11685 specification (change_cipher_spec(20), alert(21), handshake(22)).
11686 'buf' and 'len' point to the actual message, 'ssl' to the
11687 SSL object, and 'arg' is the application-defined value set by
11688 SSL[_CTX]_set_msg_callback_arg().
11690 'openssl s_client' and 'openssl s_server' have new '-msg' options
11691 to enable a callback that displays all protocol messages.
11695 * Change the shared library support so shared libraries are built as
11696 soon as the corresponding static library is finished, and thereby get
11697 openssl and the test programs linked against the shared library.
11698 This still only happens when the keyword "shard" has been given to
11699 the configuration scripts.
11701 NOTE: shared library support is still an experimental thing, and
11702 backward binary compatibility is still not guaranteed.
11704 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11706 * Add support for Subject Information Access extension.
11708 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11710 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11711 additional bytes when new memory had to be allocated, not just
11712 when reusing an existing buffer.
11716 * New command line and configuration option 'utf8' for the req command.
11717 This allows field values to be specified as UTF8 strings.
11721 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11722 runs for the former and machine-readable output for the latter.
11726 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11727 of the e-mail address in the DN (i.e., it will go into a certificate
11728 extension only). The new configuration file option 'email_in_dn = no'
11729 has the same effect.
11731 *Massimiliano Pala madwolf@openca.org*
11733 * Change all functions with names starting with `des_` to be starting
11734 with `DES_` instead. Add wrappers that are compatible with libdes,
11735 but are named `_ossl_old_des_*`. Finally, add macros that map the
11736 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11737 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11738 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11741 Since we provide two compatibility mappings, the user needs to
11742 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11743 compatibility is desired. The default (i.e., when that macro
11744 isn't defined) is OpenSSL 0.9.6c compatibility.
11746 There are also macros that enable and disable the support of old
11747 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11748 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11749 are defined, the default will apply: to support the old des routines.
11751 In either case, one must include openssl/des.h to get the correct
11752 definitions. Do not try to just include openssl/des_old.h, that
11755 NOTE: This is a major break of an old API into a new one. Software
11756 authors are encouraged to switch to the `DES_` style functions. Some
11757 time in the future, des_old.h and the libdes compatibility functions
11758 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11759 default), and then completely removed.
11763 * Test for certificates which contain unsupported critical extensions.
11764 If such a certificate is found during a verify operation it is
11765 rejected by default: this behaviour can be overridden by either
11766 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11767 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11768 X509_supported_extension() has also been added which returns 1 if a
11769 particular extension is supported.
11773 * Modify the behaviour of EVP cipher functions in similar way to digests
11774 to retain compatibility with existing code.
11778 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11779 compatibility with existing code. In particular the 'ctx' parameter does
11780 not have to be to be initialized before the call to EVP_DigestInit() and
11781 it is tidied up after a call to EVP_DigestFinal(). New function
11782 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11783 EVP_MD_CTX_copy() changed to not require the destination to be
11784 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11785 requires the destination to be valid.
11787 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11788 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11792 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11793 so that complete 'Handshake' protocol structures are kept in memory
11794 instead of overwriting 'msg_type' and 'length' with 'body' data.
11798 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11800 *Massimo Santin via Richard Levitte*
11802 * Major restructuring to the underlying ENGINE code. This includes
11803 reduction of linker bloat, separation of pure "ENGINE" manipulation
11804 (initialisation, etc) from functionality dealing with implementations
11805 of specific crypto interfaces. This change also introduces integrated
11806 support for symmetric ciphers and digest implementations - so ENGINEs
11807 can now accelerate these by providing EVP_CIPHER and EVP_MD
11808 implementations of their own. This is detailed in
11809 [crypto/engine/README.md](crypto/engine/README.md)
11810 as it couldn't be adequately described here. However, there are a few
11811 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11812 were changed in the original introduction of ENGINE code have now
11813 reverted back - the hooking from this code to ENGINE is now a good
11814 deal more passive and at run-time, operations deal directly with
11815 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11816 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11817 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11818 they were not being used by the framework as there is no concept of a
11819 BIGNUM_METHOD and they could not be generalised to the new
11820 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11821 ENGINE_cpy() has been removed as it cannot be consistently defined in
11826 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11830 * Change mkdef.pl to sort symbols that get the same entry number,
11831 and make sure the automatically generated functions `ERR_load_*`
11832 become part of libeay.num as well.
11836 * New function SSL_renegotiate_pending(). This returns true once
11837 renegotiation has been requested (either SSL_renegotiate() call
11838 or HelloRequest/ClientHello received from the peer) and becomes
11839 false once a handshake has been completed.
11840 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11841 sends a HelloRequest, but does not ensure that a handshake takes
11842 place. SSL_renegotiate_pending() is useful for checking if the
11843 client has followed the request.)
11847 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11848 By default, clients may request session resumption even during
11849 renegotiation (if session ID contexts permit); with this option,
11850 session resumption is possible only in the first handshake.
11852 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11853 more bits available for options that should not be part of
11854 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11858 * Add some demos for certificate and certificate request creation.
11862 * Make maximum certificate chain size accepted from the peer application
11863 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11864 "Douglas E. Engert" <deengert@anl.gov>.
11868 * Add support for shared libraries for Unixware-7
11869 (Boyd Lynn Gerber <gerberb@zenez.com>).
11873 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11874 be done prior to destruction. Use this to unload error strings from
11875 ENGINEs that load their own error strings. NB: This adds two new API
11876 functions to "get" and "set" this destroy handler in an ENGINE.
11880 * Alter all existing ENGINE implementations (except "openssl" and
11881 "openbsd") to dynamically instantiate their own error strings. This
11882 makes them more flexible to be built both as statically-linked ENGINEs
11883 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11884 Also, add stub code to each that makes building them as self-contained
11885 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11889 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11890 implementations into applications that are completely implemented in
11891 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11892 commands that can be used to configure what shared-library to load and
11893 to control aspects of the way it is handled. Also, made an update to
11894 the [README-Engine.md](README-Engine.md) file
11895 that brings its information up-to-date and
11896 provides some information and instructions on the "dynamic" ENGINE
11897 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11901 * Make it possible to unload ranges of ERR strings with a new
11902 "ERR_unload_strings" function.
11906 * Add a copy() function to EVP_MD.
11910 * Make EVP_MD routines take a context pointer instead of just the
11911 md_data void pointer.
11915 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11916 that the digest can only process a single chunk of data
11917 (typically because it is provided by a piece of
11918 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11919 is only going to provide a single chunk of data, and hence the
11920 framework needn't accumulate the data for oneshot drivers.
11924 * As with "ERR", make it possible to replace the underlying "ex_data"
11925 functions. This change also alters the storage and management of global
11926 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11927 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11928 index counters. The API functions that use this state have been changed
11929 to take a "class_index" rather than pointers to the class's local STACK
11930 and counter, and there is now an API function to dynamically create new
11931 classes. This centralisation allows us to (a) plug a lot of the
11932 thread-safety problems that existed, and (b) makes it possible to clean
11933 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11934 such data would previously have always leaked in application code and
11935 workarounds were in place to make the memory debugging turn a blind eye
11936 to it. Application code that doesn't use this new function will still
11937 leak as before, but their memory debugging output will announce it now
11938 rather than letting it slide.
11940 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11941 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11942 has a return value to indicate success or failure.
11946 * Make it possible to replace the underlying "ERR" functions such that the
11947 global state (2 LHASH tables and 2 locks) is only used by the "default"
11948 implementation. This change also adds two functions to "get" and "set"
11949 the implementation prior to it being automatically set the first time
11950 any other ERR function takes place. Ie. an application can call "get",
11951 pass the return value to a module it has just loaded, and that module
11952 can call its own "set" function using that value. This means the
11953 module's "ERR" operations will use (and modify) the error state in the
11954 application and not in its own statically linked copy of OpenSSL code.
11958 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11959 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11960 the operation, and provides a more encapsulated way for external code
11961 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11962 to use these functions rather than manually incrementing the counts.
11964 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11968 * Add EVP test program.
11972 * Add symmetric cipher support to ENGINE. Expect the API to change!
11976 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11977 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11978 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11979 These allow a CRL to be built without having to access X509_CRL fields
11980 directly. Modify 'ca' application to use new functions.
11984 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11985 bug workarounds. Rollback attack detection is a security feature.
11986 The problem will only arise on OpenSSL servers when TLSv1 is not
11987 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11988 Software authors not wanting to support TLSv1 will have special reasons
11989 for their choice and can explicitly enable this option.
11991 *Bodo Moeller, Lutz Jaenicke*
11993 * Rationalise EVP so it can be extended: don't include a union of
11994 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11995 (similar to those existing for EVP_CIPHER_CTX).
12000 EVP_MD_CTX_init(&md); /* new function call */
12001 EVP_DigestInit(&md, EVP_sha1());
12002 EVP_DigestUpdate(&md, in, len);
12003 EVP_DigestFinal(&md, out, NULL);
12004 EVP_MD_CTX_cleanup(&md); /* new function call */
12008 * Make DES key schedule conform to the usual scheme, as well as
12009 correcting its structure. This means that calls to DES functions
12010 now have to pass a pointer to a des_key_schedule instead of a
12011 plain des_key_schedule (which was actually always a pointer
12014 des_key_schedule ks;
12016 des_set_key_checked(..., &ks);
12017 des_ncbc_encrypt(..., &ks, ...);
12019 (Note that a later change renames 'des_...' into 'DES_...'.)
12023 * Initial reduction of linker bloat: the use of some functions, such as
12024 PEM causes large amounts of unused functions to be linked in due to
12025 poor organisation. For example pem_all.c contains every PEM function
12026 which has a knock on effect of linking in large amounts of (unused)
12027 ASN1 code. Grouping together similar functions and splitting unrelated
12028 functions prevents this.
12032 * Cleanup of EVP macros.
12036 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12037 correct `_ecb suffix`.
12041 * Add initial OCSP responder support to ocsp application. The
12042 revocation information is handled using the text based index
12043 use by the ca application. The responder can either handle
12044 requests generated internally, supplied in files (for example
12045 via a CGI script) or using an internal minimal server.
12049 * Add configuration choices to get zlib compression for TLS.
12053 * Changes to Kerberos SSL for RFC 2712 compliance:
12054 1. Implemented real KerberosWrapper, instead of just using
12055 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12056 2. Implemented optional authenticator field of KerberosWrapper.
12058 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12059 and authenticator structs; see crypto/krb5/.
12061 Generalized Kerberos calls to support multiple Kerberos libraries.
12062 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12063 via Richard Levitte*
12065 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12066 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12067 values for each of the key sizes rather than having just
12068 parameters (and 'speed' generating keys each time).
12072 * Speed up EVP routines.
12075 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12076 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12077 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12078 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12080 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12081 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12082 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12085 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12087 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12091 * Added the OS2-EMX target.
12093 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12095 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12096 to support NCONF routines in extension code. New function CONF_set_nconf()
12097 to allow functions which take an NCONF to also handle the old LHASH
12098 structure: this means that the old CONF compatible routines can be
12099 retained (in particular wrt extensions) without having to duplicate the
12100 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12104 * Enhance the general user interface with mechanisms for inner control
12105 and with possibilities to have yes/no kind of prompts.
12109 * Change all calls to low level digest routines in the library and
12110 applications to use EVP. Add missing calls to HMAC_cleanup() and
12111 don't assume HMAC_CTX can be copied using memcpy().
12113 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12115 * Add the possibility to control engines through control names but with
12116 arbitrary arguments instead of just a string.
12117 Change the key loaders to take a UI_METHOD instead of a callback
12118 function pointer. NOTE: this breaks binary compatibility with earlier
12119 versions of OpenSSL [engine].
12120 Adapt the nCipher code for these new conditions and add a card insertion
12125 * Enhance the general user interface with mechanisms to better support
12126 dialog box interfaces, application-defined prompts, the possibility
12127 to use defaults (for example default passwords from somewhere else)
12128 and interrupts/cancellations.
12132 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12133 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12137 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12138 tidy up some unnecessarily weird code in 'sk_new()').
12140 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12142 * Change the key loading routines for ENGINEs to use the same kind
12143 callback (pem_password_cb) as all other routines that need this
12148 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12149 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12150 than this minimum value is recommended.
12154 * New random seeder for OpenVMS, using the system process statistics
12155 that are easily reachable.
12159 * Windows apparently can't transparently handle global
12160 variables defined in DLLs. Initialisations such as:
12162 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12164 won't compile. This is used by the any applications that need to
12165 declare their own ASN1 modules. This was fixed by adding the option
12166 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12167 needed for static libraries under Win32.
12171 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12172 setting of purpose and trust fields. New X509_STORE trust and
12173 purpose functions and tidy up setting in other SSL functions.
12177 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12178 structure. These are inherited by X509_STORE_CTX when it is
12179 initialised. This allows various defaults to be set in the
12180 X509_STORE structure (such as flags for CRL checking and custom
12181 purpose or trust settings) for functions which only use X509_STORE_CTX
12182 internally such as S/MIME.
12184 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12185 trust settings if they are not set in X509_STORE. This allows X509_STORE
12186 purposes and trust (in S/MIME for example) to override any set by default.
12188 Add command line options for CRL checking to smime, s_client and s_server
12193 * Initial CRL based revocation checking. If the CRL checking flag(s)
12194 are set then the CRL is looked up in the X509_STORE structure and
12195 its validity and signature checked, then if the certificate is found
12196 in the CRL the verify fails with a revoked error.
12198 Various new CRL related callbacks added to X509_STORE_CTX structure.
12200 Command line options added to 'verify' application to support this.
12202 This needs some additional work, such as being able to handle multiple
12203 CRLs with different times, extension based lookup (rather than just
12204 by subject name) and ultimately more complete V2 CRL extension
12209 * Add a general user interface API (crypto/ui/). This is designed
12210 to replace things like des_read_password and friends (backward
12211 compatibility functions using this new API are provided).
12212 The purpose is to remove prompting functions from the DES code
12213 section as well as provide for prompting through dialog boxes in
12214 a window system and the like.
12218 * Add "ex_data" support to ENGINE so implementations can add state at a
12219 per-structure level rather than having to store it globally.
12223 * Make it possible for ENGINE structures to be copied when retrieved by
12224 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12225 This causes the "original" ENGINE structure to act like a template,
12226 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12227 operational state can be localised to each ENGINE structure, despite the
12228 fact they all share the same "methods". New ENGINE structures returned in
12229 this case have no functional references and the return value is the single
12230 structural reference. This matches the single structural reference returned
12231 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12236 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12237 needs to match any other type at all we need to manually clear the
12242 * Changes to the "openssl engine" utility to include;
12243 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12244 about an ENGINE's available control commands.
12245 - executing control commands from command line arguments using the
12246 '-pre' and '-post' switches. '-post' is only used if '-t' is
12247 specified and the ENGINE is successfully initialised. The syntax for
12248 the individual commands are colon-separated, for example;
12249 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12253 * New dynamic control command support for ENGINEs. ENGINEs can now
12254 declare their own commands (numbers), names (strings), descriptions,
12255 and input types for run-time discovery by calling applications. A
12256 subset of these commands are implicitly classed as "executable"
12257 depending on their input type, and only these can be invoked through
12258 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12259 can be based on user input, config files, etc). The distinction is
12260 that "executable" commands cannot return anything other than a boolean
12261 result and can only support numeric or string input, whereas some
12262 discoverable commands may only be for direct use through
12263 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12264 pointers, or other custom uses. The "executable" commands are to
12265 support parameterisations of ENGINE behaviour that can be
12266 unambiguously defined by ENGINEs and used consistently across any
12267 OpenSSL-based application. Commands have been added to all the
12268 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12269 control over shared-library paths without source code alterations.
12273 * Changed all ENGINE implementations to dynamically allocate their
12274 ENGINEs rather than declaring them statically. Apart from this being
12275 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12276 this also allows the implementations to compile without using the
12277 internal engine_int.h header.
12281 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12282 'const' value. Any code that should be able to modify a RAND_METHOD
12283 should already have non-const pointers to it (ie. they should only
12284 modify their own ones).
12288 * Made a variety of little tweaks to the ENGINE code.
12289 - "atalla" and "ubsec" string definitions were moved from header files
12290 to C code. "nuron" string definitions were placed in variables
12291 rather than hard-coded - allowing parameterisation of these values
12292 later on via ctrl() commands.
12293 - Removed unused "#if 0"'d code.
12294 - Fixed engine list iteration code so it uses ENGINE_free() to release
12295 structural references.
12296 - Constified the RAND_METHOD element of ENGINE structures.
12297 - Constified various get/set functions as appropriate and added
12298 missing functions (including a catch-all ENGINE_cpy that duplicates
12299 all ENGINE values onto a new ENGINE except reference counts/state).
12300 - Removed NULL parameter checks in get/set functions. Setting a method
12301 or function to NULL is a way of cancelling out a previously set
12302 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12303 and doesn't justify the extra error symbols and code.
12304 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12305 flags from engine_int.h to engine.h.
12306 - Changed prototypes for ENGINE handler functions (init(), finish(),
12307 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12311 * Implement binary inversion algorithm for BN_mod_inverse in addition
12312 to the algorithm using long division. The binary algorithm can be
12313 used only if the modulus is odd. On 32-bit systems, it is faster
12314 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12315 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12316 up to 450 bits. In 64-bit environments, the binary algorithm
12317 appears to be advantageous for much longer moduli; here we use it
12318 for moduli up to 2048 bits.
12322 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12323 could not support the combine flag in choice fields.
12327 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12328 extensions from a certificate request to the certificate.
12332 * Allow multiple 'certopt' and 'nameopt' options to be separated
12333 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12334 file: this allows the display of the certificate about to be
12335 signed to be customised, to allow certain fields to be included
12336 or excluded and extension details. The old system didn't display
12337 multicharacter strings properly, omitted fields not in the policy
12338 and couldn't display additional details such as extensions.
12342 * Function EC_POINTs_mul for multiple scalar multiplication
12343 of an arbitrary number of elliptic curve points
12344 \sum scalars[i]*points[i],
12345 optionally including the generator defined for the EC_GROUP:
12346 scalar*generator + \sum scalars[i]*points[i].
12348 EC_POINT_mul is a simple wrapper function for the typical case
12349 that the point list has just one item (besides the optional
12354 * First EC_METHODs for curves over GF(p):
12356 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12357 operations and provides various method functions that can also
12358 operate with faster implementations of modular arithmetic.
12360 EC_GFp_mont_method() reuses most functions that are part of
12361 EC_GFp_simple_method, but uses Montgomery arithmetic.
12363 *Bodo Moeller; point addition and point doubling
12364 implementation directly derived from source code provided by
12365 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12367 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12368 crypto/ec/ec_lib.c):
12370 Curves are EC_GROUP objects (with an optional group generator)
12371 based on EC_METHODs that are built into the library.
12373 Points are EC_POINT objects based on EC_GROUP objects.
12375 Most of the framework would be able to handle curves over arbitrary
12376 finite fields, but as there are no obvious types for fields other
12377 than GF(p), some functions are limited to that for now.
12381 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12382 that the file contains a complete HTTP response.
12386 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12387 change the def and num file printf format specifier from "%-40sXXX"
12388 to "%-39s XXX". The latter will always guarantee a space after the
12389 field while the former will cause them to run together if the field
12390 is 40 of more characters long.
12394 * Constify the cipher and digest 'method' functions and structures
12395 and modify related functions to take constant EVP_MD and EVP_CIPHER
12400 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12401 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12405 * Modify `EVP_Digest*()` routines so they now return values. Although the
12406 internal software routines can never fail additional hardware versions
12411 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12413 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12414 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12417 ERR_R_NESTED_ASN1_ERROR
12419 ERR_R_MISSING_ASN1_EOS
12420 were 4 .. 9, conflicting with
12421 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12423 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12424 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12426 Add new error code 'ERR_R_INTERNAL_ERROR'.
12430 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12435 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12436 sets the subject name for a new request or supersedes the
12437 subject name in a given request. Formats that can be parsed are
12438 'CN=Some Name, OU=myOU, C=IT'
12440 'CN=Some Name/OU=myOU/C=IT'.
12442 Add options '-batch' and '-verbose' to 'openssl req'.
12444 *Massimiliano Pala <madwolf@hackmasters.net>*
12446 * Introduce the possibility to access global variables through
12447 functions on platform were that's the best way to handle exporting
12448 global variables in shared libraries. To enable this functionality,
12449 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12450 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12451 is normally done by Configure or something similar).
12453 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12454 in the source file (foo.c) like this:
12456 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12457 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12459 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12460 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12462 OPENSSL_DECLARE_GLOBAL(int,foo);
12463 #define foo OPENSSL_GLOBAL_REF(foo)
12464 OPENSSL_DECLARE_GLOBAL(double,bar);
12465 #define bar OPENSSL_GLOBAL_REF(bar)
12467 The #defines are very important, and therefore so is including the
12468 header file everywhere where the defined globals are used.
12470 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12471 of ASN.1 items, but that structure is a bit different.
12473 The largest change is in util/mkdef.pl which has been enhanced with
12474 better and easier to understand logic to choose which symbols should
12475 go into the Windows .def files as well as a number of fixes and code
12476 cleanup (among others, algorithm keywords are now sorted
12477 lexicographically to avoid constant rewrites).
12481 * In BN_div() keep a copy of the sign of 'num' before writing the
12482 result to 'rm' because if rm==num the value will be overwritten
12483 and produce the wrong result if 'num' is negative: this caused
12484 problems with BN_mod() and BN_nnmod().
12488 * Function OCSP_request_verify(). This checks the signature on an
12489 OCSP request and verifies the signer certificate. The signer
12490 certificate is just checked for a generic purpose and OCSP request
12495 * Add OCSP_check_validity() function to check the validity of OCSP
12496 responses. OCSP responses are prepared in real time and may only
12497 be a few seconds old. Simply checking that the current time lies
12498 between thisUpdate and nextUpdate max reject otherwise valid responses
12499 caused by either OCSP responder or client clock inaccuracy. Instead
12500 we allow thisUpdate and nextUpdate to fall within a certain period of
12501 the current time. The age of the response can also optionally be
12502 checked. Two new options -validity_period and -status_age added to
12507 * If signature or public key algorithm is unrecognized print out its
12508 OID rather that just UNKNOWN.
12512 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12513 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12514 ID to be generated from the issuer certificate alone which can then be
12515 passed to OCSP_id_issuer_cmp().
12519 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12520 ASN1 modules to export functions returning ASN1_ITEM pointers
12521 instead of the ASN1_ITEM structures themselves. This adds several
12522 new macros which allow the underlying ASN1 function/structure to
12523 be accessed transparently. As a result code should not use ASN1_ITEM
12524 references directly (such as &X509_it) but instead use the relevant
12525 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12526 use of the new ASN1 code on platforms where exporting structures
12527 is problematical (for example in shared libraries) but exporting
12528 functions returning pointers to structures is not.
12532 * Add support for overriding the generation of SSL/TLS session IDs.
12533 These callbacks can be registered either in an SSL_CTX or per SSL.
12534 The purpose of this is to allow applications to control, if they wish,
12535 the arbitrary values chosen for use as session IDs, particularly as it
12536 can be useful for session caching in multiple-server environments. A
12537 command-line switch for testing this (and any client code that wishes
12538 to use such a feature) has been added to "s_server".
12540 *Geoff Thorpe, Lutz Jaenicke*
12542 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12543 of the form `#if defined(...) || defined(...) || ...` and
12544 `#if !defined(...) && !defined(...) && ...`. This also avoids
12545 the growing number of special cases it was previously handling.
12549 * Make all configuration macros available for application by making
12550 sure they are available in opensslconf.h, by giving them names starting
12551 with `OPENSSL_` to avoid conflicts with other packages and by making
12552 sure e_os2.h will cover all platform-specific cases together with
12554 Additionally, it is now possible to define configuration/platform-
12555 specific names (called "system identities"). In the C code, these
12556 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12557 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12558 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12563 * New option -set_serial to 'req' and 'x509' this allows the serial
12564 number to use to be specified on the command line. Previously self
12565 signed certificates were hard coded with serial number 0 and the
12566 CA options of 'x509' had to use a serial number in a file which was
12571 * New options to 'ca' utility to support V2 CRL entry extensions.
12572 Currently CRL reason, invalidity date and hold instruction are
12573 supported. Add new CRL extensions to V3 code and some new objects.
12577 * New function EVP_CIPHER_CTX_set_padding() this is used to
12578 disable standard block padding (aka PKCS#5 padding) in the EVP
12579 API, which was previously mandatory. This means that the data is
12580 not padded in any way and so the total length much be a multiple
12581 of the block size, otherwise an error occurs.
12585 * Initial (incomplete) OCSP SSL support.
12589 * New function OCSP_parse_url(). This splits up a URL into its host,
12590 port and path components: primarily to parse OCSP URLs. New -url
12591 option to ocsp utility.
12595 * New nonce behavior. The return value of OCSP_check_nonce() now
12596 reflects the various checks performed. Applications can decide
12597 whether to tolerate certain situations such as an absent nonce
12598 in a response when one was present in a request: the ocsp application
12599 just prints out a warning. New function OCSP_add1_basic_nonce()
12600 this is to allow responders to include a nonce in a response even if
12601 the request is nonce-less.
12605 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12606 skipped when using openssl x509 multiple times on a single input file,
12607 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12611 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12612 set string type: to handle setting ASN1_TIME structures. Fix ca
12613 utility to correctly initialize revocation date of CRLs.
12617 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12618 the clients preferred ciphersuites and rather use its own preferences.
12619 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12620 Internet Explorer by ensuring unchanged hash method during stepup.
12621 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12625 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12626 to aes and add a new 'exist' option to print out symbols that don't
12631 * Additional options to ocsp utility to allow flags to be set and
12632 additional certificates supplied.
12636 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12637 OCSP client a number of certificate to only verify the response
12642 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12643 handle the new API. Currently only ECB, CBC modes supported. Add new
12646 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12647 Encryption Standard (AES) Ciphersuites for Transport Layer
12648 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12649 not enabled by default and were not part of the "ALL" ciphersuite
12650 alias because they were not yet official; they could be
12651 explicitly requested by specifying the "AESdraft" ciphersuite
12652 group alias. In the final release of OpenSSL 0.9.7, the group
12653 alias is called "AES" and is part of "ALL".)
12655 *Ben Laurie, Steve Henson, Bodo Moeller*
12657 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12658 request to response.
12662 * Functions for OCSP responders. OCSP_request_onereq_count(),
12663 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12664 extract information from a certificate request. OCSP_response_create()
12665 creates a response and optionally adds a basic response structure.
12666 OCSP_basic_add1_status() adds a complete single response to a basic
12667 response and returns the OCSP_SINGLERESP structure just added (to allow
12668 extensions to be included for example). OCSP_basic_add1_cert() adds a
12669 certificate to a basic response and OCSP_basic_sign() signs a basic
12670 response with various flags. New helper functions ASN1_TIME_check()
12671 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12672 (converts ASN1_TIME to GeneralizedTime).
12676 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12677 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12678 structure from a certificate. X509_pubkey_digest() digests the public_key
12679 contents: this is used in various key identifiers.
12683 * Make sk_sort() tolerate a NULL argument.
12685 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12687 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12688 passed by the function are trusted implicitly. If any of them signed the
12689 response then it is assumed to be valid and is not verified.
12693 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12694 to data. This was previously part of the PKCS7 ASN1 code. This
12695 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12696 *Steve Henson, reported by Kenneth R. Robinette
12697 <support@securenetterm.com>*
12699 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12700 routines: without these tracing memory leaks is very painful.
12701 Fix leaks in PKCS12 and PKCS7 routines.
12705 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12706 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12707 effectively meant GeneralizedTime would never be used. Now it
12708 is initialised to -1 but X509_time_adj() now has to check the value
12709 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12710 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12711 *Steve Henson, reported by Kenneth R. Robinette
12712 <support@securenetterm.com>*
12714 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12715 result in a zero length in the ASN1_INTEGER structure which was
12716 not consistent with the structure when d2i_ASN1_INTEGER() was used
12717 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12718 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12719 where it did not print out a minus for negative ASN1_INTEGER.
12723 * Add summary printout to ocsp utility. The various functions which
12724 convert status values to strings have been renamed to:
12725 OCSP_response_status_str(), OCSP_cert_status_str() and
12726 OCSP_crl_reason_str() and are no longer static. New options
12727 to verify nonce values and to disable verification. OCSP response
12728 printout format cleaned up.
12732 * Add additional OCSP certificate checks. These are those specified
12733 in RFC2560. This consists of two separate checks: the CA of the
12734 certificate being checked must either be the OCSP signer certificate
12735 or the issuer of the OCSP signer certificate. In the latter case the
12736 OCSP signer certificate must contain the OCSP signing extended key
12737 usage. This check is performed by attempting to match the OCSP
12738 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12739 in the OCSP_CERTID structures of the response.
12743 * Initial OCSP certificate verification added to OCSP_basic_verify()
12744 and related routines. This uses the standard OpenSSL certificate
12745 verify routines to perform initial checks (just CA validity) and
12746 to obtain the certificate chain. Then additional checks will be
12747 performed on the chain. Currently the root CA is checked to see
12748 if it is explicitly trusted for OCSP signing. This is used to set
12749 a root CA as a global signing root: that is any certificate that
12750 chains to that CA is an acceptable OCSP signing certificate.
12754 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12755 extensions from a separate configuration file.
12756 As when reading extensions from the main configuration file,
12757 the '-extensions ...' option may be used for specifying the
12760 *Massimiliano Pala <madwolf@comune.modena.it>*
12762 * New OCSP utility. Allows OCSP requests to be generated or
12763 read. The request can be sent to a responder and the output
12764 parsed, outputted or printed in text form. Not complete yet:
12765 still needs to check the OCSP response validity.
12769 * New subcommands for 'openssl ca':
12770 `openssl ca -status <serial>` prints the status of the cert with
12771 the given serial number (according to the index file).
12772 `openssl ca -updatedb` updates the expiry status of certificates
12775 *Massimiliano Pala <madwolf@comune.modena.it>*
12777 * New '-newreq-nodes' command option to CA.pl. This is like
12778 '-newreq', but calls 'openssl req' with the '-nodes' option
12779 so that the resulting key is not encrypted.
12781 *Damien Miller <djm@mindrot.org>*
12783 * New configuration for the GNU Hurd.
12785 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12787 * Initial code to implement OCSP basic response verify. This
12788 is currently incomplete. Currently just finds the signer's
12789 certificate and verifies the signature on the response.
12793 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12794 value of OPENSSLDIR. This is available via the new '-d' option
12795 to 'openssl version', and is also included in 'openssl version -a'.
12799 * Allowing defining memory allocation callbacks that will be given
12800 file name and line number information in additional arguments
12801 (a `const char*` and an int). The basic functionality remains, as
12802 well as the original possibility to just replace malloc(),
12803 realloc() and free() by functions that do not know about these
12804 additional arguments. To register and find out the current
12805 settings for extended allocation functions, the following
12806 functions are provided:
12808 CRYPTO_set_mem_ex_functions
12809 CRYPTO_set_locked_mem_ex_functions
12810 CRYPTO_get_mem_ex_functions
12811 CRYPTO_get_locked_mem_ex_functions
12813 These work the same way as CRYPTO_set_mem_functions and friends.
12814 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12815 extended allocation function is enabled.
12816 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12817 a conventional allocation function is enabled.
12819 *Richard Levitte, Bodo Moeller*
12821 * Finish off removing the remaining LHASH function pointer casts.
12822 There should no longer be any prototype-casting required when using
12823 the LHASH abstraction, and any casts that remain are "bugs". See
12824 the callback types and macros at the head of lhash.h for details
12825 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12829 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12830 If /dev/[u]random devices are not available or do not return enough
12831 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12833 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12834 /etc/entropy will be queried once each in this sequence, querying stops
12835 when enough entropy was collected without querying more sockets.
12839 * Change the Unix RAND_poll() variant to be able to poll several
12840 random devices, as specified by DEVRANDOM, until a sufficient amount
12841 of data has been collected. We spend at most 10 ms on each file
12842 (select timeout) and read in non-blocking mode. DEVRANDOM now
12843 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12844 (previously it was just the string "/dev/urandom"), so on typical
12845 platforms the 10 ms delay will never occur.
12846 Also separate out the Unix variant to its own file, rand_unix.c.
12847 For VMS, there's a currently-empty rand_vms.c.
12851 * Move OCSP client related routines to ocsp_cl.c. These
12852 provide utility functions which an application needing
12853 to issue a request to an OCSP responder and analyse the
12854 response will typically need: as opposed to those which an
12855 OCSP responder itself would need which will be added later.
12857 OCSP_request_sign() signs an OCSP request with an API similar
12858 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12859 response. OCSP_response_get1_basic() extracts basic response
12860 from response. OCSP_resp_find_status(): finds and extracts status
12861 information from an OCSP_CERTID structure (which will be created
12862 when the request structure is built). These are built from lower
12863 level functions which work on OCSP_SINGLERESP structures but
12864 won't normally be used unless the application wishes to examine
12865 extensions in the OCSP response for example.
12867 Replace nonce routines with a pair of functions.
12868 OCSP_request_add1_nonce() adds a nonce value and optionally
12869 generates a random value. OCSP_check_nonce() checks the
12870 validity of the nonce in an OCSP response.
12874 * Change function OCSP_request_add() to OCSP_request_add0_id().
12875 This doesn't copy the supplied OCSP_CERTID and avoids the
12876 need to free up the newly created id. Change return type
12877 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12878 This can then be used to add extensions to the request.
12879 Deleted OCSP_request_new(), since most of its functionality
12880 is now in OCSP_REQUEST_new() (and the case insensitive name
12881 clash) apart from the ability to set the request name which
12882 will be added elsewhere.
12886 * Update OCSP API. Remove obsolete extensions argument from
12887 various functions. Extensions are now handled using the new
12888 OCSP extension code. New simple OCSP HTTP function which
12889 can be used to send requests and parse the response.
12893 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12894 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12895 uses the special reorder version of SET OF to sort the attributes
12896 and reorder them to match the encoded order. This resolves a long
12897 standing problem: a verify on a PKCS7 structure just after signing
12898 it used to fail because the attribute order did not match the
12899 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12900 it uses the received order. This is necessary to tolerate some broken
12901 software that does not order SET OF. This is handled by encoding
12902 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12903 to produce the required SET OF.
12907 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12908 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12909 files to get correct declarations of the ASN.1 item variables.
12913 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12914 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12915 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12916 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12917 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12918 ASN1_ITEM and no wrapper functions.
12922 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12923 replace the old function pointer based I/O routines. Change most of
12924 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12928 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12929 lines, recognize more "algorithms" that can be deselected, and make
12930 it complain about algorithm deselection that isn't recognised.
12934 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12935 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12936 to use new functions. Add NO_ASN1_OLD which can be set to remove
12937 some old style ASN1 functions: this can be used to determine if old
12938 code will still work when these eventually go away.
12942 * New extension functions for OCSP structures, these follow the
12943 same conventions as certificates and CRLs.
12947 * New function X509V3_add1_i2d(). This automatically encodes and
12948 adds an extension. Its behaviour can be customised with various
12949 flags to append, replace or delete. Various wrappers added for
12950 certificates and CRLs.
12954 * Fix to avoid calling the underlying ASN1 print routine when
12955 an extension cannot be parsed. Correct a typo in the
12956 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12960 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12961 entries for variables.
12965 * Add functionality to apps/openssl.c for detecting locking
12966 problems: As the program is single-threaded, all we have
12967 to do is register a locking callback using an array for
12968 storing which locks are currently held by the program.
12972 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12973 SSL_get_ex_data_X509_STORE_idx(), which is used in
12974 ssl_verify_cert_chain() and thus can be called at any time
12975 during TLS/SSL handshakes so that thread-safety is essential.
12976 Unfortunately, the ex_data design is not at all suited
12977 for multi-threaded use, so it probably should be abolished.
12981 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12983 *Broadcom, tweaked and integrated by Geoff Thorpe*
12985 * Move common extension printing code to new function
12986 X509V3_print_extensions(). Reorganise OCSP print routines and
12987 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12991 * New function X509_signature_print() to remove duplication in some
12996 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12997 set (this was treated exactly the same as SET OF previously). This
12998 is used to reorder the STACK representing the structure to match the
12999 encoding. This will be used to get round a problem where a PKCS7
13000 structure which was signed could not be verified because the STACK
13001 order did not reflect the encoded order.
13005 * Reimplement the OCSP ASN1 module using the new code.
13009 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13010 for its ASN1 operations. The old style function pointers still exist
13011 for now but they will eventually go away.
13015 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13016 completely replaces the old ASN1 functionality with a table driven
13017 encoder and decoder which interprets an ASN1_ITEM structure describing
13018 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13019 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13020 has also been converted to the new form.
13024 * Change BN_mod_exp_recp so that negative moduli are tolerated
13025 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13026 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13027 for negative moduli.
13031 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13032 of not touching the result's sign bit.
13036 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13041 * Changed the LHASH code to use prototypes for callbacks, and created
13042 macros to declare and implement thin (optionally static) functions
13043 that provide type-safety and avoid function pointer casting for the
13044 type-specific callbacks.
13048 * Added Kerberos Cipher Suites to be used with TLS, as written in
13050 *Veers Staats <staatsvr@asc.hpc.mil>,
13051 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13053 * Reformat the FAQ so the different questions and answers can be divided
13054 in sections depending on the subject.
13058 * Have the zlib compression code load ZLIB.DLL dynamically under
13063 * New function BN_mod_sqrt for computing square roots modulo a prime
13064 (using the probabilistic Tonelli-Shanks algorithm unless
13065 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13066 be handled deterministically).
13068 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13070 * Make BN_mod_inverse faster by explicitly handling small quotients
13071 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13072 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13076 * New function BN_kronecker.
13080 * Fix BN_gcd so that it works on negative inputs; the result is
13081 positive unless both parameters are zero.
13082 Previously something reasonably close to an infinite loop was
13083 possible because numbers could be growing instead of shrinking
13084 in the implementation of Euclid's algorithm.
13088 * Fix BN_is_word() and BN_is_one() macros to take into account the
13089 sign of the number in question.
13091 Fix BN_is_word(a,w) to work correctly for w == 0.
13093 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13094 because its test if the absolute value of 'a' equals 'w'.
13095 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13096 it exists mostly for use in the implementations of BN_is_zero(),
13097 BN_is_one(), and BN_is_word().
13101 * New function BN_swap.
13105 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13106 the exponentiation functions are more likely to produce reasonable
13107 results on negative inputs.
13111 * Change BN_mod_mul so that the result is always non-negative.
13112 Previously, it could be negative if one of the factors was negative;
13113 I don't think anyone really wanted that behaviour.
13117 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13118 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13119 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13120 and add new functions:
13129 BN_mod_lshift1_quick
13131 BN_mod_lshift_quick
13133 These functions always generate non-negative results.
13135 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13136 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13138 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13139 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13140 be reduced modulo `m`.
13142 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13145 The following entry accidentally appeared in the CHANGES file
13146 distributed with OpenSSL 0.9.7. The modifications described in
13147 it do *not* apply to OpenSSL 0.9.7.
13149 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13150 was actually never needed) and in BN_mul(). The removal in BN_mul()
13151 required a small change in bn_mul_part_recursive() and the addition
13152 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13153 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13154 bn_sub_words() and bn_add_words() except they take arrays with
13160 * In 'openssl passwd', verify passwords read from the terminal
13161 unless the '-salt' option is used (which usually means that
13162 verification would just waste user's time since the resulting
13163 hash is going to be compared with some given password hash)
13164 or the new '-noverify' option is used.
13166 This is an incompatible change, but it does not affect
13167 non-interactive use of 'openssl passwd' (passwords on the command
13168 line, '-stdin' option, '-in ...' option) and thus should not
13169 cause any problems.
13173 * Remove all references to RSAref, since there's no more need for it.
13177 * Make DSO load along a path given through an environment variable
13178 (SHLIB_PATH) with shl_load().
13182 * Constify the ENGINE code as a result of BIGNUM constification.
13183 Also constify the RSA code and most things related to it. In a
13184 few places, most notable in the depth of the ASN.1 code, ugly
13185 casts back to non-const were required (to be solved at a later
13190 * Make it so the openssl application has all engines loaded by default.
13194 * Constify the BIGNUM routines a little more.
13198 * Add the following functions:
13200 ENGINE_load_cswift()
13202 ENGINE_load_atalla()
13203 ENGINE_load_nuron()
13204 ENGINE_load_builtin_engines()
13206 That way, an application can itself choose if external engines that
13207 are built-in in OpenSSL shall ever be used or not. The benefit is
13208 that applications won't have to be linked with libdl or other dso
13209 libraries unless it's really needed.
13211 Changed 'openssl engine' to load all engines on demand.
13212 Changed the engine header files to avoid the duplication of some
13213 declarations (they differed!).
13217 * 'openssl engine' can now list capabilities.
13221 * Better error reporting in 'openssl engine'.
13225 * Never call load_dh_param(NULL) in s_server.
13229 * Add engine application. It can currently list engines by name and
13230 identity, and test if they are actually available.
13234 * Improve RPM specification file by forcing symbolic linking and making
13235 sure the installed documentation is also owned by root.root.
13237 *Damien Miller <djm@mindrot.org>*
13239 * Give the OpenSSL applications more possibilities to make use of
13240 keys (public as well as private) handled by engines.
13244 * Add OCSP code that comes from CertCo.
13248 * Add VMS support for the Rijndael code.
13252 * Added untested support for Nuron crypto accelerator.
13256 * Add support for external cryptographic devices. This code was
13257 previously distributed separately as the "engine" branch.
13259 *Geoff Thorpe, Richard Levitte*
13261 * Rework the filename-translation in the DSO code. It is now possible to
13262 have far greater control over how a "name" is turned into a filename
13263 depending on the operating environment and any oddities about the
13264 different shared library filenames on each system.
13268 * Support threads on FreeBSD-elf in Configure.
13272 * Fix for SHA1 assembly problem with MASM: it produces
13273 warnings about corrupt line number information when assembling
13274 with debugging information. This is caused by the overlapping
13277 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13280 NCONF_get_number() has no error checking at all. As a replacement,
13281 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13282 promoted strongly. The old NCONF_get_number is kept around for
13283 binary backward compatibility.
13284 Make it possible for methods to load from something other than a BIO,
13285 by providing a function pointer that is given a name instead of a BIO.
13286 For example, this could be used to load configuration data from an
13291 * Fix for non blocking accept BIOs. Added new I/O special reason
13292 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13293 with non blocking I/O was not possible because no retry code was
13294 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13299 * Added the beginnings of Rijndael support.
13303 * Fix for bug in DirectoryString mask setting. Add support for
13304 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13305 to allow certificate printing to more controllable, additional
13306 'certopt' option to 'x509' to allow new printing options to be
13311 * Clean old EAY MD5 hack from e_os.h.
13315 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13317 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13318 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13320 *Joe Orton, Steve Henson*
13322 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13324 * Fix additional bug revealed by the NISCC test suite:
13326 Stop bug triggering large recursion when presented with
13327 certain ASN.1 tags ([CVE-2003-0851])
13331 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13333 * Fix various bugs revealed by running the NISCC test suite:
13335 Stop out of bounds reads in the ASN1 code when presented with
13336 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13338 If verify callback ignores invalid public key errors don't try to check
13339 certificate signature with the NULL public key.
13343 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13344 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13349 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13350 extra data after the compression methods not only for TLS 1.0
13351 but also for SSL 3.0 (as required by the specification).
13353 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13355 * Change X509_certificate_type() to mark the key as exported/exportable
13356 when it's 512 *bits* long, not 512 bytes.
13360 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13362 * Countermeasure against the Klima-Pokorny-Rosa extension of
13363 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13364 a protocol version number mismatch like a decryption error
13365 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13369 * Turn on RSA blinding by default in the default implementation
13370 to avoid a timing attack. Applications that don't want it can call
13371 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13372 They would be ill-advised to do so in most cases.
13374 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13376 * Change RSA blinding code so that it works when the PRNG is not
13377 seeded (in this case, the secret RSA exponent is abused as
13378 an unpredictable seed -- if it is not unpredictable, there
13379 is no point in blinding anyway). Make RSA blinding thread-safe
13380 by remembering the creator's thread ID in rsa->blinding and
13381 having all other threads use local one-time blinding factors
13382 (this requires more computation than sharing rsa->blinding, but
13383 avoids excessive locking; and if an RSA object is not shared
13384 between threads, blinding will still be very fast).
13388 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13390 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13391 via timing by performing a MAC computation even if incorrect
13392 block cipher padding has been found. This is a countermeasure
13393 against active attacks where the attacker has to distinguish
13394 between bad padding and a MAC verification error. ([CVE-2003-0078])
13396 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13397 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13398 Martin Vuagnoux (EPFL, Ilion)*
13400 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13402 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13403 memory from its contents. This is done with a counter that will
13404 place alternating values in each byte. This can be used to solve
13405 two issues: 1) the removal of calls to memset() by highly optimizing
13406 compilers, and 2) cleansing with other values than 0, since those can
13407 be read through on certain media, for example a swap space on disk.
13411 * Bugfix: client side session caching did not work with external caching,
13412 because the session->cipher setting was not restored when reloading
13413 from the external cache. This problem was masked, when
13414 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13415 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13419 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13420 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13422 *Zeev Lieber <zeev-l@yahoo.com>*
13424 * Undo an undocumented change introduced in 0.9.6e which caused
13425 repeated calls to OpenSSL_add_all_ciphers() and
13426 OpenSSL_add_all_digests() to be ignored, even after calling
13431 * Change the default configuration reader to deal with last line not
13432 being properly terminated.
13436 * Change X509_NAME_cmp() so it applies the special rules on handling
13437 DN values that are of type PrintableString, as well as RDNs of type
13438 emailAddress where the value has the type ia5String.
13440 *stefank@valicert.com via Richard Levitte*
13442 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13443 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13444 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13445 the bitwise-OR of the two for use by the majority of applications
13446 wanting this behaviour, and update the docs. The documented
13447 behaviour and actual behaviour were inconsistent and had been
13448 changing anyway, so this is more a bug-fix than a behavioural
13451 *Geoff Thorpe, diagnosed by Nadav Har'El*
13453 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13454 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13458 * Fix initialization code race conditions in
13459 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13460 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13461 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13462 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13463 ssl2_get_cipher_by_char(),
13464 ssl3_get_cipher_by_char().
13466 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13468 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13469 the cached sessions are flushed, as the remove_cb() might use ex_data
13470 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13471 (see [openssl.org #212]).
13473 *Geoff Thorpe, Lutz Jaenicke*
13475 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13476 length, instead of the encoding length to d2i_ASN1_OBJECT.
13480 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13482 * [In 0.9.6g-engine release:]
13483 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13485 *Lynn Gazis <lgazis@rainbow.com>*
13487 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13489 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13490 and get fix the header length calculation.
13491 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13492 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13494 * Use proper error handling instead of 'assertions' in buffer
13495 overflow checks added in 0.9.6e. This prevents DoS (the
13496 assertions could call abort()).
13498 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13500 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13502 * Add various sanity checks to asn1_get_length() to reject
13503 the ASN1 length bytes if they exceed sizeof(long), will appear
13504 negative or the content length exceeds the length of the
13507 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13509 * Fix cipher selection routines: ciphers without encryption had no flags
13510 for the cipher strength set and where therefore not handled correctly
13511 by the selection routines (PR #130).
13515 * Fix EVP_dsa_sha macro.
13520 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13521 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13522 that was added in OpenSSL 0.9.6d.
13524 As the countermeasure turned out to be incompatible with some
13525 broken SSL implementations, the new option is part of SSL_OP_ALL.
13526 SSL_OP_ALL is usually employed when compatibility with weird SSL
13527 implementations is desired (e.g. '-bugs' option to 's_client' and
13528 's_server'), so the new option is automatically set in many
13533 * Changes in security patch:
13535 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13536 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13537 Air Force Materiel Command, USAF, under agreement number
13540 * Add various sanity checks to asn1_get_length() to reject
13541 the ASN1 length bytes if they exceed sizeof(long), will appear
13542 negative or the content length exceeds the length of the
13543 supplied buffer. ([CVE-2002-0659])
13545 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13547 * Assertions for various potential buffer overflows, not known to
13548 happen in practice.
13550 *Ben Laurie (CHATS)*
13552 * Various temporary buffers to hold ASCII versions of integers were
13553 too small for 64 bit platforms. ([CVE-2002-0655])
13554 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13556 * Remote buffer overflow in SSL3 protocol - an attacker could
13557 supply an oversized session ID to a client. ([CVE-2002-0656])
13559 *Ben Laurie (CHATS)*
13561 * Remote buffer overflow in SSL2 protocol - an attacker could
13562 supply an oversized client master key. ([CVE-2002-0656])
13564 *Ben Laurie (CHATS)*
13566 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13568 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13569 encoded as NULL) with id-dsa-with-sha1.
13571 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13573 * Check various `X509_...()` return values in apps/req.c.
13575 *Nils Larsch <nla@trustcenter.de>*
13577 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13578 an end-of-file condition would erroneously be flagged, when the CRLF
13579 was just at the end of a processed block. The bug was discovered when
13580 processing data through a buffering memory BIO handing the data to a
13581 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13582 <ptsekov@syntrex.com> and Nedelcho Stanev.
13586 * Implement a countermeasure against a vulnerability recently found
13587 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13588 before application data chunks to avoid the use of known IVs
13589 with data potentially chosen by the attacker.
13593 * Fix length checks in ssl3_get_client_hello().
13597 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13598 to prevent ssl3_read_internal() from incorrectly assuming that
13599 ssl3_read_bytes() found application data while handshake
13600 processing was enabled when in fact s->s3->in_read_app_data was
13601 merely automatically cleared during the initial handshake.
13603 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13605 * Fix object definitions for Private and Enterprise: they were not
13606 recognized in their shortname (=lowercase) representation. Extend
13607 obj_dat.pl to issue an error when using undefined keywords instead
13608 of silently ignoring the problem (Svenning Sorensen
13609 <sss@sss.dnsalias.net>).
13613 * Fix DH_generate_parameters() so that it works for 'non-standard'
13614 generators, i.e. generators other than 2 and 5. (Previously, the
13615 code did not properly initialise the 'add' and 'rem' values to
13616 BN_generate_prime().)
13618 In the new general case, we do not insist that 'generator' is
13619 actually a primitive root: This requirement is rather pointless;
13620 a generator of the order-q subgroup is just as good, if not
13625 * Map new X509 verification errors to alerts. Discovered and submitted by
13626 Tom Wu <tom@arcot.com>.
13630 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13631 returning non-zero before the data has been completely received
13632 when using non-blocking I/O.
13634 *Bodo Moeller; problem pointed out by John Hughes*
13636 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13638 *Ben Laurie, Lutz Jaenicke*
13640 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13641 Yoram Zahavi <YoramZ@gilian.com>).
13645 * Add information about CygWin 1.3 and on, and preserve proper
13646 configuration for the versions before that.
13648 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13650 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13651 check whether we deal with a copy of a session and do not delete from
13652 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13653 <izhar@checkpoint.com>.
13657 * Do not store session data into the internal session cache, if it
13658 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13659 flag is set). Proposed by Aslam <aslam@funk.com>.
13663 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13668 * [In 0.9.6d-engine release:]
13669 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13671 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13673 * Add the configuration target linux-s390x.
13675 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13677 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13678 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13679 variable as an indication that a ClientHello message has been
13680 received. As the flag value will be lost between multiple
13681 invocations of ssl3_accept when using non-blocking I/O, the
13682 function may not be aware that a handshake has actually taken
13683 place, thus preventing a new session from being added to the
13686 To avoid this problem, we now set s->new_session to 2 instead of
13687 using a local variable.
13689 *Lutz Jaenicke, Bodo Moeller*
13691 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13692 if the SSL_R_LENGTH_MISMATCH error is detected.
13694 *Geoff Thorpe, Bodo Moeller*
13696 * New 'shared_ldflag' column in Configure platform table.
13700 * Fix EVP_CIPHER_mode macro.
13702 *"Dan S. Camper" <dan@bti.net>*
13704 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13705 type, we must throw them away by setting rr->length to 0.
13707 *D P Chang <dpc@qualys.com>*
13709 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13711 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13712 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13713 worked incorrectly for those cases where range = `10..._2` and
13714 `3*range` is two bits longer than range.)
13718 * Only add signing time to PKCS7 structures if it is not already
13723 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13724 OBJ_ld_ce should be OBJ_id_ce.
13725 Also some ip-pda OIDs in crypto/objects/objects.txt were
13726 incorrect (cf. RFC 3039).
13728 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13730 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13731 returns early because it has nothing to do.
13733 *Andy Schneider <andy.schneider@bjss.co.uk>*
13735 * [In 0.9.6c-engine release:]
13736 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13738 *Andy Schneider <andy.schneider@bjss.co.uk>*
13740 * [In 0.9.6c-engine release:]
13741 Add support for Cryptographic Appliance's keyserver technology.
13742 (Use engine 'keyclient')
13744 *Cryptographic Appliances and Geoff Thorpe*
13746 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13747 is called via tools/c89.sh because arguments have to be
13748 rearranged (all '-L' options must appear before the first object
13751 *Richard Shapiro <rshapiro@abinitio.com>*
13753 * [In 0.9.6c-engine release:]
13754 Add support for Broadcom crypto accelerator cards, backported
13757 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13759 * [In 0.9.6c-engine release:]
13760 Add support for SureWare crypto accelerator cards from
13761 Baltimore Technologies. (Use engine 'sureware')
13763 *Baltimore Technologies and Mark Cox*
13765 * [In 0.9.6c-engine release:]
13766 Add support for crypto accelerator cards from Accelerated
13767 Encryption Processing, www.aep.ie. (Use engine 'aep')
13769 *AEP Inc. and Mark Cox*
13771 * Add a configuration entry for gcc on UnixWare.
13773 *Gary Benson <gbenson@redhat.com>*
13775 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13776 messages are stored in a single piece (fixed-length part and
13777 variable-length part combined) and fix various bugs found on the way.
13781 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13782 instead. BIO_gethostbyname() does not know what timeouts are
13783 appropriate, so entries would stay in cache even when they have
13785 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13787 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13788 faced with a pathologically small ClientHello fragment that does
13789 not contain client_version: Instead of aborting with an error,
13790 simply choose the highest available protocol version (i.e.,
13791 TLS 1.0 unless it is disabled). In practice, ClientHello
13792 messages are never sent like this, but this change gives us
13793 strictly correct behaviour at least for TLS.
13797 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13798 never resets s->method to s->ctx->method when called from within
13799 one of the SSL handshake functions.
13801 *Bodo Moeller; problem pointed out by Niko Baric*
13803 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13804 (sent using the client's version number) if client_version is
13805 smaller than the protocol version in use. Also change
13806 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13807 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13808 the client will at least see that alert.
13812 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13817 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13818 client receives HelloRequest while in a handshake.
13820 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13822 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13823 should end in 'break', not 'goto end' which circumvents various
13824 cleanups done in state SSL_ST_OK. But session related stuff
13825 must be disabled for SSL_ST_OK in the case that we just sent a
13828 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13829 before just sending a HelloRequest.
13831 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13833 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13834 reveal whether illegal block cipher padding was found or a MAC
13835 verification error occurred. (Neither SSLerr() codes nor alerts
13836 are directly visible to potential attackers, but the information
13837 may leak via logfiles.)
13839 Similar changes are not required for the SSL 2.0 implementation
13840 because the number of padding bytes is sent in clear for SSL 2.0,
13841 and the extra bytes are just ignored. However ssl/s2_pkt.c
13842 failed to verify that the purported number of padding bytes is in
13847 * Add OpenUNIX-8 support including shared libraries
13848 (Boyd Lynn Gerber <gerberb@zenez.com>).
13852 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13853 'wristwatch attack' using huge encoding parameters (cf.
13854 James H. Manger's CRYPTO 2001 paper). Note that the
13855 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13856 encoding parameters and hence was not vulnerable.
13860 * BN_sqr() bug fix.
13862 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13864 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13865 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13866 followed by modular reduction.
13868 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13870 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13871 equivalent based on BN_pseudo_rand() instead of BN_rand().
13875 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13876 This function was broken, as the check for a new client hello message
13877 to handle SGC did not allow these large messages.
13878 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13882 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13886 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13887 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13891 * Rework the configuration and shared library support for Tru64 Unix.
13892 The configuration part makes use of modern compiler features and
13893 still retains old compiler behavior for those that run older versions
13894 of the OS. The shared library support part includes a variant that
13895 uses the RPATH feature, and is available through the special
13896 configuration target "alpha-cc-rpath", which will never be selected
13899 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13901 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13902 with the same message size as in ssl3_get_certificate_request().
13903 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13904 messages might inadvertently be reject as too long.
13906 *Petr Lampa <lampa@fee.vutbr.cz>*
13908 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13912 * Modified SSL library such that the verify_callback that has been set
13913 specifically for an SSL object with SSL_set_verify() is actually being
13914 used. Before the change, a verify_callback set with this function was
13915 ignored and the verify_callback() set in the SSL_CTX at the time of
13916 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13917 to allow the necessary settings.
13921 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13922 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13923 done automatically (in contradiction to the requirements of the C
13924 standard). This made problems when used from OpenSSH.
13928 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13929 dh->length and always used
13931 BN_rand_range(priv_key, dh->p).
13933 BN_rand_range() is not necessary for Diffie-Hellman, and this
13934 specific range makes Diffie-Hellman unnecessarily inefficient if
13935 dh->length (recommended exponent length) is much smaller than the
13936 length of dh->p. We could use BN_rand_range() if the order of
13937 the subgroup was stored in the DH structure, but we only have
13942 BN_rand(priv_key, l, ...)
13944 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13951 RSA_eay_public_encrypt
13952 RSA_eay_private_decrypt
13953 RSA_eay_private_encrypt (signing)
13954 RSA_eay_public_decrypt (signature verification)
13956 (default implementations for RSA_public_encrypt,
13957 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13958 always reject numbers >= n.
13962 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13963 to synchronize access to 'locking_thread'. This is necessary on
13964 systems where access to 'locking_thread' (an 'unsigned long'
13965 variable) is not atomic.
13969 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13970 *before* setting the 'crypto_lock_rand' flag. The previous code had
13971 a race condition if 0 is a valid thread ID.
13973 *Travis Vitek <vitek@roguewave.com>*
13975 * Add support for shared libraries under Irix.
13977 *Albert Chin-A-Young <china@thewrittenword.com>*
13979 * Add configuration option to build on Linux on both big-endian and
13980 little-endian MIPS.
13982 *Ralf Baechle <ralf@uni-koblenz.de>*
13984 * Add the possibility to create shared libraries on HP-UX.
13988 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13990 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13991 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13992 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13993 PRNG state recovery was possible based on the output of
13994 one PRNG request appropriately sized to gain knowledge on
13995 'md' followed by enough consecutive 1-byte PRNG requests
13996 to traverse all of 'state'.
13998 1. When updating 'md_local' (the current thread's copy of 'md')
13999 during PRNG output generation, hash all of the previous
14000 'md_local' value, not just the half used for PRNG output.
14002 2. Make the number of bytes from 'state' included into the hash
14003 independent from the number of PRNG bytes requested.
14005 The first measure alone would be sufficient to avoid
14006 Markku-Juhani's attack. (Actually it had never occurred
14007 to me that the half of 'md_local' used for chaining was the
14008 half from which PRNG output bytes were taken -- I had always
14009 assumed that the secret half would be used.) The second
14010 measure makes sure that additional data from 'state' is never
14011 mixed into 'md_local' in small portions; this heuristically
14012 further strengthens the PRNG.
14016 * Fix crypto/bn/asm/mips3.s.
14020 * When only the key is given to "enc", the IV is undefined. Print out
14021 an error message in this case.
14025 * Handle special case when X509_NAME is empty in X509 printing routines.
14029 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14030 positive and less than q.
14034 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14035 used: it isn't thread safe and the add_lock_callback should handle
14038 *Paul Rose <Paul.Rose@bridge.com>*
14040 * Verify that incoming data obeys the block size in
14041 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14047 *Ulf Möller, Bodo Möller*
14049 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14050 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14051 when fixing the server behaviour for backwards-compatible 'client
14052 hello' messages. (Note that the attack is impractical against
14053 SSL 3.0 and TLS 1.0 anyway because length and version checking
14054 means that the probability of guessing a valid ciphertext is
14055 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14058 Before 0.9.5, the countermeasure (hide the error by generating a
14059 random 'decryption result') did not work properly because
14060 ERR_clear_error() was missing, meaning that SSL_get_error() would
14061 detect the supposedly ignored error.
14063 Both problems are now fixed.
14067 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14068 (previously it was 1024).
14072 * Fix for compatibility mode trust settings: ignore trust settings
14073 unless some valid trust or reject settings are present.
14077 * Fix for blowfish EVP: its a variable length cipher.
14081 * Fix various bugs related to DSA S/MIME verification. Handle missing
14082 parameters in DSA public key structures and return an error in the
14083 DSA routines if parameters are absent.
14087 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14088 in the current directory if neither $RANDFILE nor $HOME was set.
14089 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14090 caused some confusion to Windows users who haven't defined $HOME.
14091 Thus RAND_file_name() is changed again: e_os.h can define a
14092 DEFAULT_HOME, which will be used if $HOME is not set.
14093 For Windows, we use "C:"; on other platforms, we still require
14094 environment variables.
14096 * Move 'if (!initialized) RAND_poll()' into regions protected by
14097 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14098 having multiple threads call RAND_poll() concurrently.
14102 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14103 combination of a flag and a thread ID variable.
14104 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14105 flag), *other* threads can enter ssleay_add_bytes without obeying
14106 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14107 that they do not hold after the first thread unsets add_do_not_lock).
14111 * Change bctest again: '-x' expressions are not available in all
14112 versions of 'test'.
14116 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14118 * Fix a couple of memory leaks in PKCS7_dataDecode()
14120 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14122 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14123 the default extension for executables, if any. Also, make the perl
14124 scripts that use symlink() to test if it really exists and use "cp"
14125 if it doesn't. All this made OpenSSL compilable and installable in
14130 * Fix for asn1_GetSequence() for indefinite length constructed data.
14131 If SEQUENCE is length is indefinite just set c->slen to the total
14132 amount of data available.
14134 *Steve Henson, reported by shige@FreeBSD.org*
14136 *This change does not apply to 0.9.7.*
14138 * Change bctest to avoid here-documents inside command substitution
14139 (workaround for FreeBSD /bin/sh bug).
14140 For compatibility with Ultrix, avoid shell functions (introduced
14141 in the bctest version that searches along $PATH).
14145 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14146 with des_encrypt() defined on some operating systems, like Solaris
14151 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14152 On the Importance of Eliminating Errors in Cryptographic
14153 Computations, J. Cryptology 14 (2001) 2, 101-119,
14154 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14158 * MIPS assembler BIGNUM division bug fix.
14162 * Disabled incorrect Alpha assembler code.
14166 * Fix PKCS#7 decode routines so they correctly update the length
14167 after reading an EOC for the EXPLICIT tag.
14171 *This change does not apply to 0.9.7.*
14173 * Fix bug in PKCS#12 key generation routines. This was triggered
14174 if a 3DES key was generated with a 0 initial byte. Include
14175 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14176 (but broken) behaviour.
14180 * Enhance bctest to search for a working bc along $PATH and print
14183 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14185 * Fix memory leaks in err.c: free err_data string if necessary;
14186 don't write to the wrong index in ERR_set_error_data.
14190 * Implement ssl23_peek (analogous to ssl23_read), which previously
14195 * Replace rdtsc with `_emit` statements for VC++ version 5.
14197 *Jeremy Cooper <jeremy@baymoo.org>*
14199 * Make it possible to reuse SSLv2 sessions.
14203 * In copy_email() check for >= 0 as a return value for
14204 X509_NAME_get_index_by_NID() since 0 is a valid index.
14206 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14208 * Avoid coredump with unsupported or invalid public keys by checking if
14209 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14210 PKCS7_verify() fails with non detached data.
14214 * Don't use getenv in library functions when run as setuid/setgid.
14215 New function OPENSSL_issetugid().
14219 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14220 due to incorrect handling of multi-threading:
14222 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14224 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14226 3. Count how many times MemCheck_off() has been called so that
14227 nested use can be treated correctly. This also avoids
14228 inband-signalling in the previous code (which relied on the
14229 assumption that thread ID 0 is impossible).
14233 * Add "-rand" option also to s_client and s_server.
14237 * Fix CPU detection on Irix 6.x.
14238 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14239 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14241 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14246 *This change does not apply to 0.9.7.*
14248 * Use the cached encoding of an X509_NAME structure rather than
14249 copying it. This is apparently the reason for the libsafe "errors"
14250 but the code is actually correct.
14254 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14255 Bleichenbacher's DSA attack.
14256 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14257 to be set and top=0 forces the highest bit to be set; top=-1 is new
14258 and leaves the highest bit random.
14260 *Ulf Moeller, Bodo Moeller*
14262 * In the `NCONF_...`-based implementations for `CONF_...` queries
14263 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14264 a temporary CONF structure with the data component set to NULL
14265 (which gives segmentation faults in lh_retrieve).
14266 Instead, use NULL for the CONF pointer in CONF_get_string and
14267 CONF_get_number (which may use environment variables) and directly
14268 return NULL from CONF_get_section.
14272 * Fix potential buffer overrun for EBCDIC.
14276 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14277 keyUsage if basicConstraints absent for a CA.
14281 * Make SMIME_write_PKCS7() write mail header values with a format that
14282 is more generally accepted (no spaces before the semicolon), since
14283 some programs can't parse those values properly otherwise. Also make
14284 sure BIO's that break lines after each write do not create invalid
14289 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14290 macros previously used would not encode an empty SEQUENCE OF
14291 and break the signature.
14295 *This change does not apply to 0.9.7.*
14297 * Zero the premaster secret after deriving the master secret in
14302 * Add some EVP_add_digest_alias registrations (as found in
14303 OpenSSL_add_all_digests()) to SSL_library_init()
14304 aka OpenSSL_add_ssl_algorithms(). This provides improved
14305 compatibility with peers using X.509 certificates
14306 with unconventional AlgorithmIdentifier OIDs.
14310 * Fix for Irix with NO_ASM.
14312 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14314 * ./config script fixes.
14316 *Ulf Moeller, Richard Levitte*
14318 * Fix 'openssl passwd -1'.
14322 * Change PKCS12_key_gen_asc() so it can cope with non null
14323 terminated strings whose length is passed in the passlen
14324 parameter, for example from PEM callbacks. This was done
14325 by adding an extra length parameter to asc2uni().
14327 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14329 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14330 call failed, free the DSA structure.
14334 * Fix to uni2asc() to cope with zero length Unicode strings.
14335 These are present in some PKCS#12 files.
14339 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14340 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14341 when writing a 32767 byte record.
14343 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14345 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14346 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14348 (RSA objects have a reference count access to which is protected
14349 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14350 so they are meant to be shared between threads.)
14351 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14352 "Reddie, Steven" <Steven.Reddie@ca.com>*
14354 * Fix a deadlock in CRYPTO_mem_leaks().
14358 * Use better test patterns in bntest.
14362 * rand_win.c fix for Borland C.
14366 * BN_rshift bugfix for n == 0.
14370 * Add a 'bctest' script that checks for some known 'bc' bugs
14371 so that 'make test' does not abort just because 'bc' is broken.
14375 * Store verify_result within SSL_SESSION also for client side to
14376 avoid potential security hole. (Re-used sessions on the client side
14377 always resulted in verify_result==X509_V_OK, not using the original
14378 result of the server certificate verification.)
14382 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14383 SSL3_RT_APPLICATION_DATA, return 0.
14384 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14389 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14390 releases, have been re-implemented by renaming the previous
14391 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14392 and ssl3_read_internal, respectively, and adding 'peek' parameters
14393 to them. The new ssl[23]_{read,peek} functions are calls to
14394 ssl[23]_read_internal with the 'peek' flag set appropriately.
14395 A 'peek' parameter has also been added to ssl3_read_bytes, which
14396 does the actual work for ssl3_read_internal.
14400 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14401 the method-specific "init()" handler. Also clean up ex_data after
14402 calling the method-specific "finish()" handler. Previously, this was
14403 happening the other way round.
14407 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14408 The previous value, 12, was not always sufficient for BN_mod_exp().
14412 * Make sure that shared libraries get the internal name engine with
14413 the full version number and not just 0. This should mark the
14414 shared libraries as not backward compatible. Of course, this should
14415 be changed again when we can guarantee backward binary compatibility.
14419 * Fix typo in get_cert_by_subject() in by_dir.c
14421 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14423 * Rework the system to generate shared libraries:
14425 - Make note of the expected extension for the shared libraries and
14426 if there is a need for symbolic links from for example libcrypto.so.0
14427 to libcrypto.so.0.9.7. There is extended info in Configure for
14430 - Make as few rebuilds of the shared libraries as possible.
14432 - Still avoid linking the OpenSSL programs with the shared libraries.
14434 - When installing, install the shared libraries separately from the
14439 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14441 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14442 and not in SSL_clear because the latter is also used by the
14443 accept/connect functions; previously, the settings made by
14444 SSL_set_read_ahead would be lost during the handshake.
14446 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14448 * Correct util/mkdef.pl to be selective about disabled algorithms.
14449 Previously, it would create entries for disabled algorithms no
14454 * Added several new manual pages for SSL_* function.
14458 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14460 * In ssl23_get_client_hello, generate an error message when faced
14461 with an initial SSL 3.0/TLS record that is too small to contain the
14462 first two bytes of the ClientHello message, i.e. client_version.
14463 (Note that this is a pathologic case that probably has never happened
14464 in real life.) The previous approach was to use the version number
14465 from the record header as a substitute; but our protocol choice
14466 should not depend on that one because it is not authenticated
14467 by the Finished messages.
14471 * More robust randomness gathering functions for Windows.
14473 *Jeffrey Altman <jaltman@columbia.edu>*
14475 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14476 not set then we don't setup the error code for issuer check errors
14477 to avoid possibly overwriting other errors which the callback does
14478 handle. If an application does set the flag then we assume it knows
14479 what it is doing and can handle the new informational codes
14484 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14485 a general "ANY" type, as such it should be able to decode anything
14486 including tagged types. However it didn't check the class so it would
14487 wrongly interpret tagged types in the same way as their universal
14488 counterpart and unknown types were just rejected. Changed so that the
14489 tagged and unknown types are handled in the same way as a SEQUENCE:
14490 that is the encoding is stored intact. There is also a new type
14491 "V_ASN1_OTHER" which is used when the class is not universal, in this
14492 case we have no idea what the actual type is so we just lump them all
14497 * On VMS, stdout may very well lead to a file that is written to
14498 in a record-oriented fashion. That means that every write() will
14499 write a separate record, which will be read separately by the
14500 programs trying to read from it. This can be very confusing.
14502 The solution is to put a BIO filter in the way that will buffer
14503 text until a linefeed is reached, and then write everything a
14504 line at a time, so every record written will be an actual line,
14505 not chunks of lines and not (usually doesn't happen, but I've
14506 seen it once) several lines in one record. BIO_f_linebuffer() is
14509 Currently, it's a VMS-only method, because that's where it has
14510 been tested well enough.
14514 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14515 it can return incorrect results.
14516 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14517 but it was in 0.9.6-beta[12].)
14521 * Disable the check for content being present when verifying detached
14522 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14523 include zero length content when signing messages.
14527 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14528 BIO_ctrl (for BIO pairs).
14532 * Add DSO method for VMS.
14536 * Bug fix: Montgomery multiplication could produce results with the
14541 * Add RPM specification openssl.spec and modify it to build three
14542 packages. The default package contains applications, application
14543 documentation and run-time libraries. The devel package contains
14544 include files, static libraries and function documentation. The
14545 doc package contains the contents of the doc directory. The original
14546 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14550 * Add a large number of documentation files for many SSL routines.
14552 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14554 * Add a configuration entry for Sony News 4.
14556 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14558 * Don't set the two most significant bits to one when generating a
14559 random number < q in the DSA library.
14563 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14564 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14565 the underlying transport is blocking) if a handshake took place.
14566 (The default behaviour is needed by applications such as s_client
14567 and s_server that use select() to determine when to use SSL_read;
14568 but for applications that know in advance when to expect data, it
14569 just makes things more complicated.)
14573 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14578 * Add a few more EBCDIC conditionals that make `req` and `x509`
14579 work better on such systems.
14581 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14583 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14584 Update PKCS12_parse() so it copies the friendlyName and the
14585 keyid to the certificates aux info.
14589 * Fix bug in PKCS7_verify() which caused an infinite loop
14590 if there was more than one signature.
14592 *Sven Uszpelkat <su@celocom.de>*
14594 * Major change in util/mkdef.pl to include extra information
14595 about each symbol, as well as presenting variables as well
14596 as functions. This change means that there's n more need
14597 to rebuild the .num files when some algorithms are excluded.
14601 * Allow the verify time to be set by an application,
14602 rather than always using the current time.
14606 * Phase 2 verify code reorganisation. The certificate
14607 verify code now looks up an issuer certificate by a
14608 number of criteria: subject name, authority key id
14609 and key usage. It also verifies self signed certificates
14610 by the same criteria. The main comparison function is
14611 X509_check_issued() which performs these checks.
14613 Lot of changes were necessary in order to support this
14614 without completely rewriting the lookup code.
14616 Authority and subject key identifier are now cached.
14618 The LHASH 'certs' is X509_STORE has now been replaced
14619 by a STACK_OF(X509_OBJECT). This is mainly because an
14620 LHASH can't store or retrieve multiple objects with
14621 the same hash value.
14623 As a result various functions (which were all internal
14624 use only) have changed to handle the new X509_STORE
14625 structure. This will break anything that messed round
14626 with X509_STORE internally.
14628 The functions X509_STORE_add_cert() now checks for an
14629 exact match, rather than just subject name.
14631 The X509_STORE API doesn't directly support the retrieval
14632 of multiple certificates matching a given criteria, however
14633 this can be worked round by performing a lookup first
14634 (which will fill the cache with candidate certificates)
14635 and then examining the cache for matches. This is probably
14636 the best we can do without throwing out X509_LOOKUP
14637 entirely (maybe later...).
14639 The X509_VERIFY_CTX structure has been enhanced considerably.
14641 All certificate lookup operations now go via a get_issuer()
14642 callback. Although this currently uses an X509_STORE it
14643 can be replaced by custom lookups. This is a simple way
14644 to bypass the X509_STORE hackery necessary to make this
14645 work and makes it possible to use more efficient techniques
14646 in future. A very simple version which uses a simple
14647 STACK for its trusted certificate store is also provided
14648 using X509_STORE_CTX_trusted_stack().
14650 The verify_cb() and verify() callbacks now have equivalents
14651 in the X509_STORE_CTX structure.
14653 X509_STORE_CTX also has a 'flags' field which can be used
14654 to customise the verify behaviour.
14658 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14659 excludes S/MIME capabilities.
14663 * When a certificate request is read in keep a copy of the
14664 original encoding of the signed data and use it when outputting
14665 again. Signatures then use the original encoding rather than
14666 a decoded, encoded version which may cause problems if the
14667 request is improperly encoded.
14671 * For consistency with other BIO_puts implementations, call
14672 buffer_write(b, ...) directly in buffer_puts instead of calling
14675 In BIO_puts, increment b->num_write as in BIO_write.
14677 *Peter.Sylvester@EdelWeb.fr*
14679 * Fix BN_mul_word for the case where the word is 0. (We have to use
14680 BN_zero, we may not return a BIGNUM with an array consisting of
14681 words set to zero.)
14685 * Avoid calling abort() from within the library when problems are
14686 detected, except if preprocessor symbols have been defined
14687 (such as REF_CHECK, BN_DEBUG etc.).
14691 * New openssl application 'rsautl'. This utility can be
14692 used for low level RSA operations. DER public key
14693 BIO/fp routines also added.
14697 * New Configure entry and patches for compiling on QNX 4.
14699 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14701 * A demo state-machine implementation was sponsored by
14702 Nuron (<http://www.nuron.com/>) and is now available in
14703 demos/state_machine.
14707 * New options added to the 'dgst' utility for signature
14708 generation and verification.
14712 * Unrecognized PKCS#7 content types are now handled via a
14713 catch all ASN1_TYPE structure. This allows unsupported
14714 types to be stored as a "blob" and an application can
14715 encode and decode it manually.
14719 * Fix various signed/unsigned issues to make a_strex.c
14720 compile under VC++.
14722 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14724 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14725 length if passed a buffer. ASN1_INTEGER_to_BN failed
14726 if passed a NULL BN and its argument was negative.
14728 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14730 * Modification to PKCS#7 encoding routines to output definite
14731 length encoding. Since currently the whole structures are in
14732 memory there's not real point in using indefinite length
14733 constructed encoding. However if OpenSSL is compiled with
14734 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14738 * Added BIO_vprintf() and BIO_vsnprintf().
14742 * Added more prefixes to parse for in the strings written
14743 through a logging bio, to cover all the levels that are available
14744 through syslog. The prefixes are now:
14746 PANIC, EMERG, EMR => LOG_EMERG
14747 ALERT, ALR => LOG_ALERT
14748 CRIT, CRI => LOG_CRIT
14749 ERROR, ERR => LOG_ERR
14750 WARNING, WARN, WAR => LOG_WARNING
14751 NOTICE, NOTE, NOT => LOG_NOTICE
14752 INFO, INF => LOG_INFO
14753 DEBUG, DBG => LOG_DEBUG
14755 and as before, if none of those prefixes are present at the
14756 beginning of the string, LOG_ERR is chosen.
14758 On Win32, the `LOG_*` levels are mapped according to this:
14760 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14761 LOG_WARNING => EVENTLOG_WARNING_TYPE
14762 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14766 * Made it possible to reconfigure with just the configuration
14767 argument "reconf" or "reconfigure". The command line arguments
14768 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14769 and are retrieved from there when reconfiguring.
14775 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14777 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14781 * The obj_dat.pl script was messing up the sorting of object
14782 names. The reason was that it compared the quoted version
14783 of strings as a result "OCSP" > "OCSP Signing" because
14784 " > SPACE. Changed script to store unquoted versions of
14785 names and add quotes on output. It was also omitting some
14786 names from the lookup table if they were given a default
14787 value (that is if SN is missing it is given the same
14788 value as LN and vice versa), these are now added on the
14789 grounds that if an object has a name we should be able to
14790 look it up. Finally added warning output when duplicate
14791 short or long names are found.
14795 * Changes needed for Tandem NSK.
14797 *Scott Uroff <scott@xypro.com>*
14799 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14800 RSA_padding_check_SSLv23(), special padding was never detected
14801 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14802 version rollback attacks was not effective.
14804 In s23_clnt.c, don't use special rollback-attack detection padding
14805 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14806 client; similarly, in s23_srvr.c, don't do the rollback check if
14807 SSL 2.0 is the only protocol enabled in the server.
14811 * Make it possible to get hexdumps of unprintable data with 'openssl
14812 asn1parse'. By implication, the functions ASN1_parse_dump() and
14813 BIO_dump_indent() are added.
14817 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14818 these print out strings and name structures based on various
14819 flags including RFC2253 support and proper handling of
14820 multibyte characters. Added options to the 'x509' utility
14821 to allow the various flags to be set.
14825 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14826 Also change the functions X509_cmp_current_time() and
14827 X509_gmtime_adj() work with an ASN1_TIME structure,
14828 this will enable certificates using GeneralizedTime in validity
14829 dates to be checked.
14833 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14834 negative public key encodings) on by default,
14835 NO_NEG_PUBKEY_BUG can be set to disable it.
14839 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14840 content octets. An i2c_ASN1_OBJECT is unnecessary because
14841 the encoding can be trivially obtained from the structure.
14845 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14846 not read locks (`CRYPTO_r_[un]lock`).
14850 * A first attempt at creating official support for shared
14851 libraries through configuration. I've kept it so the
14852 default is static libraries only, and the OpenSSL programs
14853 are always statically linked for now, but there are
14854 preparations for dynamic linking in place.
14855 This has been tested on Linux and Tru64.
14859 * Randomness polling function for Win9x, as described in:
14860 Peter Gutmann, Software Generation of Practically Strong
14865 * Fix so PRNG is seeded in req if using an already existing
14870 * New options to smime application. -inform and -outform
14871 allow alternative formats for the S/MIME message including
14872 PEM and DER. The -content option allows the content to be
14873 specified separately. This should allow things like Netscape
14874 form signing output easier to verify.
14878 * Fix the ASN1 encoding of tags using the 'long form'.
14882 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14883 STRING types. These convert content octets to and from the
14884 underlying type. The actual tag and length octets are
14885 already assumed to have been read in and checked. These
14886 are needed because all other string types have virtually
14887 identical handling apart from the tag. By having versions
14888 of the ASN1 functions that just operate on content octets
14889 IMPLICIT tagging can be handled properly. It also allows
14890 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14891 and ASN1_INTEGER are identical apart from the tag.
14895 * Change the handling of OID objects as follows:
14897 - New object identifiers are inserted in objects.txt, following
14898 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14899 - objects.pl is used to process obj_mac.num and create a new
14901 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14904 This is currently kind of a hack, and the perl code in objects.pl
14905 isn't very elegant, but it works as I intended. The simplest way
14906 to check that it worked correctly is to look in obj_dat.h and
14907 check the array nid_objs and make sure the objects haven't moved
14908 around (this is important!). Additions are OK, as well as
14909 consistent name changes.
14913 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14917 * Addition of the command line parameter '-rand file' to 'openssl req'.
14918 The given file adds to whatever has already been seeded into the
14919 random pool through the RANDFILE configuration file option or
14920 environment variable, or the default random state file.
14924 * mkstack.pl now sorts each macro group into lexical order.
14925 Previously the output order depended on the order the files
14926 appeared in the directory, resulting in needless rewriting
14931 * Patches to make OpenSSL compile under Win32 again. Mostly
14932 work arounds for the VC++ problem that it treats func() as
14933 func(void). Also stripped out the parts of mkdef.pl that
14934 added extra typesafe functions: these no longer exist.
14938 * Reorganisation of the stack code. The macros are now all
14939 collected in safestack.h . Each macro is defined in terms of
14940 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14941 DEBUG_SAFESTACK is now handled in terms of function casts,
14942 this has the advantage of retaining type safety without the
14943 use of additional functions. If DEBUG_SAFESTACK is not defined
14944 then the non typesafe macros are used instead. Also modified the
14945 mkstack.pl script to handle the new form. Needs testing to see
14946 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14947 the default if no major problems. Similar behaviour for ASN1_SET_OF
14948 and PKCS12_STACK_OF.
14952 * When some versions of IIS use the 'NET' form of private key the
14953 key derivation algorithm is different. Normally MD5(password) is
14954 used as a 128 bit RC4 key. In the modified case
14955 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14956 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14957 as the old Netscape_RSA functions except they have an additional
14958 'sgckey' parameter which uses the modified algorithm. Also added
14959 an -sgckey command line option to the rsa utility. Thanks to
14960 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14961 algorithm to openssl-dev.
14965 * The evp_local.h macros were using 'c.##kname' which resulted in
14966 invalid expansion on some systems (SCO 5.0.5 for example).
14967 Corrected to 'c.kname'.
14969 *Phillip Porch <root@theporch.com>*
14971 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14972 a STACK of email addresses from a certificate or request, these look
14973 in the subject name and the subject alternative name extensions and
14974 omit any duplicate addresses.
14978 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14979 This makes DSA verification about 2 % faster.
14983 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14984 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14985 plus overhead for 1024 bit moduli).
14986 This makes exponentiations about 0.5 % faster for 1024 bit
14987 exponents (as measured by "openssl speed rsa2048").
14991 * Rename memory handling macros to avoid conflicts with other
14993 Malloc => OPENSSL_malloc
14994 Malloc_locked => OPENSSL_malloc_locked
14995 Realloc => OPENSSL_realloc
14996 Free => OPENSSL_free
15000 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15001 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15005 * CygWin32 support.
15007 *John Jarvie <jjarvie@newsguy.com>*
15009 * The type-safe stack code has been rejigged. It is now only compiled
15010 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15011 by default all type-specific stack functions are "#define"d back to
15012 standard stack functions. This results in more streamlined output
15013 but retains the type-safety checking possibilities of the original
15018 * The STACK code has been cleaned up, and certain type declarations
15019 that didn't make a lot of sense have been brought in line. This has
15020 also involved a cleanup of sorts in safestack.h to more correctly
15021 map type-safe stack functions onto their plain stack counterparts.
15022 This work has also resulted in a variety of "const"ifications of
15023 lots of the code, especially `_cmp` operations which should normally
15024 be prototyped with "const" parameters anyway.
15028 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15029 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15030 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15031 where all of 'md' is used each time the PRNG is used, but 'state'
15032 is used only indexed by a cyclic counter. As entropy may not be
15033 well distributed from the beginning, 'md' is important as a
15034 chaining variable. However, the output function chains only half
15035 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15036 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15037 in all of 'state' being rewritten, with the new values depending
15038 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15042 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15043 the handshake is continued after ssl_verify_cert_chain();
15044 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15045 can lead to 'unexplainable' connection aborts later.
15047 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15049 * Major EVP API cipher revision.
15050 Add hooks for extra EVP features. This allows various cipher
15051 parameters to be set in the EVP interface. Support added for variable
15052 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15053 setting of RC2 and RC5 parameters.
15055 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15058 Remove lots of duplicated code from the EVP library. For example *every*
15059 cipher init() function handles the 'iv' in the same way according to the
15060 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15061 for CFB and OFB modes they zero ctx->num.
15063 New functionality allows removal of S/MIME code RC2 hack.
15065 Most of the routines have the same form and so can be declared in terms
15068 By shifting this to the top level EVP_CipherInit() it can be removed from
15069 all individual ciphers. If the cipher wants to handle IVs or keys
15070 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15073 Change lots of functions like EVP_EncryptUpdate() to now return a
15074 value: although software versions of the algorithms cannot fail
15075 any installed hardware versions can.
15079 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15080 this option is set, tolerate broken clients that send the negotiated
15081 protocol version number instead of the requested protocol version
15086 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15087 i.e. non-zero for export ciphersuites, zero otherwise.
15088 Previous versions had this flag inverted, inconsistent with
15089 rsa_tmp_cb (..._TMP_RSA_CB).
15091 *Bodo Moeller; problem reported by Amit Chopra*
15093 * Add missing DSA library text string. Work around for some IIS
15094 key files with invalid SEQUENCE encoding.
15098 * Add a document (doc/standards.txt) that list all kinds of standards
15099 and so on that are implemented in OpenSSL.
15103 * Enhance c_rehash script. Old version would mishandle certificates
15104 with the same subject name hash and wouldn't handle CRLs at all.
15105 Added -fingerprint option to crl utility, to support new c_rehash
15110 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15114 * Fix for SSL server purpose checking. Server checking was
15115 rejecting certificates which had extended key usage present
15116 but no ssl client purpose.
15118 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15120 * Make PKCS#12 code work with no password. The PKCS#12 spec
15121 is a little unclear about how a blank password is handled.
15122 Since the password in encoded as a BMPString with terminating
15123 double NULL a zero length password would end up as just the
15124 double NULL. However no password at all is different and is
15125 handled differently in the PKCS#12 key generation code. NS
15126 treats a blank password as zero length. MSIE treats it as no
15127 password on export: but it will try both on import. We now do
15128 the same: PKCS12_parse() tries zero length and no password if
15129 the password is set to "" or NULL (NULL is now a valid password:
15130 it wasn't before) as does the pkcs12 application.
15134 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15135 perror when PEM_read_bio_X509_REQ fails, the error message must
15136 be obtained from the error queue.
15140 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15141 it in ERR_remove_state if appropriate, and change ERR_get_state
15142 accordingly to avoid race conditions (this is necessary because
15143 thread_hash is no longer constant once set).
15147 * Bugfix for linux-elf makefile.one.
15151 * RSA_get_default_method() will now cause a default
15152 RSA_METHOD to be chosen if one doesn't exist already.
15153 Previously this was only set during a call to RSA_new()
15154 or RSA_new_method(NULL) meaning it was possible for
15155 RSA_get_default_method() to return NULL.
15159 * Added native name translation to the existing DSO code
15160 that will convert (if the flag to do so is set) filenames
15161 that are sufficiently small and have no path information
15162 into a canonical native form. Eg. "blah" converted to
15163 "libblah.so" or "blah.dll" etc.
15167 * New function ERR_error_string_n(e, buf, len) which is like
15168 ERR_error_string(e, buf), but writes at most 'len' bytes
15169 including the 0 terminator. For ERR_error_string_n, 'buf'
15172 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15174 * CONF library reworked to become more general. A new CONF
15175 configuration file reader "class" is implemented as well as a
15176 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15177 old `CONF_*` functions are still there, but are reimplemented to
15178 work in terms of the new functions. Also, a set of functions
15179 to handle the internal storage of the configuration data is
15180 provided to make it easier to write new configuration file
15181 reader "classes" (I can definitely see something reading a
15182 configuration file in XML format, for example), called `_CONF_*`,
15183 or "the configuration storage API"...
15185 The new configuration file reading functions are:
15187 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15188 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15190 NCONF_default, NCONF_WIN32
15192 NCONF_dump_fp, NCONF_dump_bio
15194 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15195 NCONF_new creates a new CONF object. This works in the same way
15196 as other interfaces in OpenSSL, like the BIO interface.
15197 `NCONF_dump_*` dump the internal storage of the configuration file,
15198 which is useful for debugging. All other functions take the same
15199 arguments as the old `CONF_*` functions with the exception of the
15200 first that must be a `CONF *` instead of a `LHASH *`.
15202 To make it easier to use the new classes with the old `CONF_*` functions,
15203 the function CONF_set_default_method is provided.
15207 * Add '-tls1' option to 'openssl ciphers', which was already
15208 mentioned in the documentation but had not been implemented.
15209 (This option is not yet really useful because even the additional
15210 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15214 * Initial DSO code added into libcrypto for letting OpenSSL (and
15215 OpenSSL-based applications) load shared libraries and bind to
15216 them in a portable way.
15218 *Geoff Thorpe, with contributions from Richard Levitte*
15220 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15222 * Make sure _lrotl and _lrotr are only used with MSVC.
15224 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15225 (the default implementation of RAND_status).
15227 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15228 to '-clrext' (= clear extensions), as intended and documented.
15229 *Bodo Moeller; inconsistency pointed out by Michael Attili
15230 <attili@amaxo.com>*
15232 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15233 was larger than the MD block size.
15235 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15237 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15238 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15239 using the passed key: if the passed key was a private key the result
15240 of X509_print(), for example, would be to print out all the private key
15245 * des_quad_cksum() byte order bug fix.
15246 *Ulf Möller, using the problem description in krb4-0.9.7, where
15247 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15249 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15252 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15254 * For easily testing in shell scripts whether some command
15255 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15256 returns with exit code 0 iff no command of the given name is available.
15257 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15258 the output goes to stdout and nothing is printed to stderr.
15259 Additional arguments are always ignored.
15261 Since for each cipher there is a command of the same name,
15262 the 'no-cipher' compilation switches can be tested this way.
15264 ('openssl no-XXX' is not able to detect pseudo-commands such
15265 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15269 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15273 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15274 is set; it will be thrown away anyway because each handshake creates
15276 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15277 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15278 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15279 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15283 * New s_client option -ign_eof: EOF at stdin is ignored, and
15284 'Q' and 'R' lose their special meanings (quit/renegotiate).
15285 This is part of what -quiet does; unlike -quiet, -ign_eof
15286 does not suppress any output.
15290 * Add compatibility options to the purpose and trust code. The
15291 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15292 accepts a certificate or CA, this was the previous behaviour,
15293 with all the associated security issues.
15295 X509_TRUST_COMPAT is the old trust behaviour: only and
15296 automatically trust self signed roots in certificate store. A
15297 new trust setting X509_TRUST_DEFAULT is used to specify that
15298 a purpose has no associated trust setting and it should instead
15299 use the value in the default purpose.
15303 * Fix the PKCS#8 DSA private key code so it decodes keys again
15304 and fix a memory leak.
15308 * In util/mkerr.pl (which implements 'make errors'), preserve
15309 reason strings from the previous version of the .c file, as
15310 the default to have only downcase letters (and digits) in
15311 automatically generated reasons codes is not always appropriate.
15315 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15316 using strerror. Previously, ERR_reason_error_string() returned
15317 library names as reason strings for SYSerr; but SYSerr is a special
15318 case where small numbers are errno values, not library numbers.
15322 * Add '-dsaparam' option to 'openssl dhparam' application. This
15323 converts DSA parameters into DH parameters. (When creating parameters,
15324 DSA_generate_parameters is used.)
15328 * Include 'length' (recommended exponent length) in C code generated
15329 by 'openssl dhparam -C'.
15333 * The second argument to set_label in perlasm was already being used
15334 so couldn't be used as a "file scope" flag. Moved to third argument
15339 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15340 instead of RAND_bytes for encryption IVs and salts.
15344 * Include RAND_status() into RAND_METHOD instead of implementing
15345 it only for md_rand.c Otherwise replacing the PRNG by calling
15346 RAND_set_rand_method would be impossible.
15350 * Don't let DSA_generate_key() enter an infinite loop if the random
15351 number generation fails.
15355 * New 'rand' application for creating pseudo-random output.
15359 * Added configuration support for Linux/IA64
15361 *Rolf Haberrecker <rolf@suse.de>*
15363 * Assembler module support for Mingw32.
15367 * Shared library support for HPUX (in shlib/).
15369 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15371 * Shared library support for Solaris gcc.
15373 *Lutz Behnke <behnke@trustcenter.de>*
15375 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15377 * PKCS7_encrypt() was adding text MIME headers twice because they
15378 were added manually and by SMIME_crlf_copy().
15382 * In bntest.c don't call BN_rand with zero bits argument.
15384 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15386 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15387 case was implemented. This caused BN_div_recp() to fail occasionally.
15391 * Add an optional second argument to the set_label() in the perl
15392 assembly language builder. If this argument exists and is set
15393 to 1 it signals that the assembler should use a symbol whose
15394 scope is the entire file, not just the current function. This
15395 is needed with MASM which uses the format label:: for this scope.
15397 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15399 * Change the ASN1 types so they are typedefs by default. Before
15400 almost all types were #define'd to ASN1_STRING which was causing
15401 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15406 * Change names of new functions to the new get1/get0 naming
15407 convention: After 'get1', the caller owns a reference count
15408 and has to call `..._free`; 'get0' returns a pointer to some
15409 data structure without incrementing reference counters.
15410 (Some of the existing 'get' functions increment a reference
15411 counter, some don't.)
15412 Similarly, 'set1' and 'add1' functions increase reference
15413 counters or duplicate objects.
15417 * Allow for the possibility of temp RSA key generation failure:
15418 the code used to assume it always worked and crashed on failure.
15422 * Fix potential buffer overrun problem in BIO_printf().
15423 *Ulf Möller, using public domain code by Patrick Powell; problem
15424 pointed out by David Sacerdote <das33@cornell.edu>*
15426 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15427 RAND_egd() and RAND_status(). In the command line application,
15428 the EGD socket can be specified like a seed file using RANDFILE
15433 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15434 Some CAs (e.g. Verisign) distribute certificates in this form.
15438 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15439 list to exclude them. This means that no special compilation option
15440 is needed to use anonymous DH: it just needs to be included in the
15445 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15446 EVP_MD_type. The old functionality is available in a new macro called
15447 EVP_MD_md(). Change code that uses it and update docs.
15451 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15452 where the `void *` argument is replaced by a function pointer argument.
15453 Previously `void *` was abused to point to functions, which works on
15454 many platforms, but is not correct. As these functions are usually
15455 called by macros defined in OpenSSL header files, most source code
15456 should work without changes.
15460 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15461 sections with information on -D... compiler switches used for
15462 compiling the library so that applications can see them. To enable
15463 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15464 must be defined. E.g.,
15465 #define OPENSSL_ALGORITHM_DEFINES
15466 #include <openssl/opensslconf.h>
15467 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15469 *Richard Levitte, Ulf and Bodo Möller*
15471 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15476 * Change the 'other' type in certificate aux info to a STACK_OF
15477 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15478 the required ASN1 format: arbitrary types determined by an OID.
15482 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15483 argument to 'req'. This is not because the function is newer or
15484 better than others it just uses the work 'NEW' in the certificate
15485 request header lines. Some software needs this.
15489 * Reorganise password command line arguments: now passwords can be
15490 obtained from various sources. Delete the PEM_cb function and make
15491 it the default behaviour: i.e. if the callback is NULL and the
15492 usrdata argument is not NULL interpret it as a null terminated pass
15493 phrase. If usrdata and the callback are NULL then the pass phrase
15494 is prompted for as usual.
15498 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15499 the support is automatically enabled. The resulting binaries will
15500 autodetect the card and use it if present.
15502 *Ben Laurie and Compaq Inc.*
15504 * Work around for Netscape hang bug. This sends certificate request
15505 and server done in one record. Since this is perfectly legal in the
15506 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15507 the bugs/SSLv3 entry for more info.
15511 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15515 * Add -rand argument to smime and pkcs12 applications and read/write
15520 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15524 * Add command line password options to the remaining applications.
15528 * Bug fix for BN_div_recp() for numerators with an even number of
15533 * More tests in bntest.c, and changed test_bn output.
15537 * ./config recognizes MacOS X now.
15541 * Bug fix for BN_div() when the first words of num and divisor are
15542 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15546 * Add support for various broken PKCS#8 formats, and command line
15547 options to produce them.
15551 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15552 get temporary BIGNUMs from a BN_CTX.
15556 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15561 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15562 include a #define from the old name to the new. The original intent
15563 was that statically linked binaries could for example just call
15564 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15565 link with digests. This never worked because SSLeay_add_all_digests()
15566 and SSLeay_add_all_ciphers() were in the same source file so calling
15567 one would link with the other. They are now in separate source files.
15571 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15575 * Use a less unusual form of the Miller-Rabin primality test (it used
15576 a binary algorithm for exponentiation integrated into the Miller-Rabin
15577 loop, our standard modexp algorithms are faster).
15581 * Support for the EBCDIC character set completed.
15583 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15585 * Source code cleanups: use const where appropriate, eliminate casts,
15586 use `void *` instead of `char *` in lhash.
15590 * Bugfix: ssl3_send_server_key_exchange was not restartable
15591 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15592 this the server could overwrite ephemeral keys that the client
15597 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15598 using 50 iterations of the Rabin-Miller test.
15600 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15601 iterations of the Rabin-Miller test as required by the appendix
15602 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15603 As BN_is_prime_fasttest includes trial division, DSA parameter
15604 generation becomes much faster.
15606 This implies a change for the callback functions in DSA_is_prime
15607 and DSA_generate_parameters: The callback function is called once
15608 for each positive witness in the Rabin-Miller test, not just
15609 occasionally in the inner loop; and the parameters to the
15610 callback function now provide an iteration count for the outer
15611 loop rather than for the current invocation of the inner loop.
15612 DSA_generate_parameters additionally can call the callback
15613 function with an 'iteration count' of -1, meaning that a
15614 candidate has passed the trial division test (when q is generated
15615 from an application-provided seed, trial division is skipped).
15619 * New function BN_is_prime_fasttest that optionally does trial
15620 division before starting the Rabin-Miller test and has
15621 an additional BN_CTX * argument (whereas BN_is_prime always
15622 has to allocate at least one BN_CTX).
15623 'callback(1, -1, cb_arg)' is called when a number has passed the
15624 trial division stage.
15628 * Fix for bug in CRL encoding. The validity dates weren't being handled
15633 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15637 * New function BN_pseudo_rand().
15641 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15642 bignum version of BN_from_montgomery() with the working code from
15643 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15648 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15649 made it impossible to use the same SSL_SESSION data structure in
15650 SSL2 clients in multiple threads.
15654 * The return value of RAND_load_file() no longer counts bytes obtained
15655 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15656 to seed the PRNG (previously an explicit byte count was required).
15658 *Ulf Möller, Bodo Möller*
15660 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15661 used `char *` instead of `void *` and had casts all over the place.
15665 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15669 * Retain source code compatibility for BN_prime_checks macro:
15670 BN_is_prime(..., BN_prime_checks, ...) now uses
15671 BN_prime_checks_for_size to determine the appropriate number of
15672 Rabin-Miller iterations.
15676 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15677 DH_CHECK_P_NOT_SAFE_PRIME.
15678 (Check if this is true? OpenPGP calls them "strong".)
15682 * Merge the functionality of "dh" and "gendh" programs into a new program
15683 "dhparam". The old programs are retained for now but will handle DH keys
15684 (instead of parameters) in future.
15688 * Make the ciphers, s_server and s_client programs check the return values
15689 when a new cipher list is set.
15693 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15694 ciphers. Before when the 56bit ciphers were enabled the sorting was
15697 The syntax for the cipher sorting has been extended to support sorting by
15698 cipher-strength (using the strength_bits hard coded in the tables).
15699 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15701 Fix a bug in the cipher-command parser: when supplying a cipher command
15702 string with an "undefined" symbol (neither command nor alphanumeric
15703 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15704 an error is flagged.
15706 Due to the strength-sorting extension, the code of the
15707 ssl_create_cipher_list() function was completely rearranged. I hope that
15708 the readability was also increased :-)
15710 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15712 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15713 for the first serial number and places 2 in the serial number file. This
15714 avoids problems when the root CA is created with serial number zero and
15715 the first user certificate has the same issuer name and serial number
15720 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15721 the new code. Add documentation for this stuff.
15725 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15726 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15727 structures and behave in an analogous way to the X509v3 functions:
15728 they shouldn't be called directly but wrapper functions should be used
15731 So we also now have some wrapper functions that call the X509at functions
15732 when passed certificate requests. (TO DO: similar things can be done with
15733 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15734 things. Some of these need some d2i or i2d and print functionality
15735 because they handle more complex structures.)
15739 * Add missing #ifndefs that caused missing symbols when building libssl
15740 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15741 NO_RSA in `ssl/s2*.c`.
15743 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15745 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15746 has a return value which indicates the quality of the random data
15747 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15748 error queue. New function RAND_pseudo_bytes() generates output that is
15749 guaranteed to be unique but not unpredictable. RAND_add is like
15750 RAND_seed, but takes an extra argument for an entropy estimate
15751 (RAND_seed always assumes full entropy).
15755 * Do more iterations of Rabin-Miller probable prime test (specifically,
15756 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15757 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15758 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15759 false-positive rate of at most 2^-80 for random input.
15763 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15767 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15768 in the 0.9.5 release), this returns the chain
15769 from an X509_CTX structure with a dup of the stack and all
15770 the X509 reference counts upped: so the stack will exist
15771 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15774 Also make SSL_SESSION_print() print out the verify return
15779 * Add manpage for the pkcs12 command. Also change the default
15780 behaviour so MAC iteration counts are used unless the new
15781 -nomaciter option is used. This improves file security and
15782 only older versions of MSIE (4.0 for example) need it.
15786 * Honor the no-xxx Configure options when creating .DEF files.
15790 * Add PKCS#10 attributes to field table: challengePassword,
15791 unstructuredName and unstructuredAddress. These are taken from
15792 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15793 international characters are used.
15795 More changes to X509_ATTRIBUTE code: allow the setting of types
15796 based on strings. Remove the 'loc' parameter when adding
15797 attributes because these will be a SET OF encoding which is sorted
15802 * Initial changes to the 'req' utility to allow request generation
15803 automation. This will allow an application to just generate a template
15804 file containing all the field values and have req construct the
15807 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15808 used all over the place including certificate requests and PKCS#7
15809 structures. They are currently handled manually where necessary with
15810 some primitive wrappers for PKCS#7. The new functions behave in a
15811 manner analogous to the X509 extension functions: they allow
15812 attributes to be looked up by NID and added.
15814 Later something similar to the X509V3 code would be desirable to
15815 automatically handle the encoding, decoding and printing of the
15816 more complex types. The string types like challengePassword can
15817 be handled by the string table functions.
15819 Also modified the multi byte string table handling. Now there is
15820 a 'global mask' which masks out certain types. The table itself
15821 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15822 is useful when for example there is only one permissible type
15823 (as in countryName) and using the mask might result in no valid
15828 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15829 SSL_get_peer_finished to allow applications to obtain the latest
15830 Finished messages sent to the peer or expected from the peer,
15831 respectively. (SSL_get_peer_finished is usually the Finished message
15832 actually received from the peer, otherwise the protocol will be aborted.)
15834 As the Finished message are message digests of the complete handshake
15835 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15836 be used for external authentication procedures when the authentication
15837 provided by SSL/TLS is not desired or is not enough.
15841 * Enhanced support for Alpha Linux is added. Now ./config checks if
15842 the host supports BWX extension and if Compaq C is present on the
15843 $PATH. Just exploiting of the BWX extension results in 20-30%
15844 performance kick for some algorithms, e.g. DES and RC4 to mention
15845 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15850 * Add support for MS "fast SGC". This is arguably a violation of the
15851 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15852 weak crypto and after checking the certificate is SGC a second one
15853 with strong crypto. MS SGC stops the first handshake after receiving
15854 the server certificate message and sends a second client hello. Since
15855 a server will typically do all the time consuming operations before
15856 expecting any further messages from the client (server key exchange
15857 is the most expensive) there is little difference between the two.
15859 To get OpenSSL to support MS SGC we have to permit a second client
15860 hello message after we have sent server done. In addition we have to
15861 reset the MAC if we do get this second client hello.
15865 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15866 if a DER encoded private key is RSA or DSA traditional format. Changed
15867 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15868 format DER encoded private key. Newer code should use PKCS#8 format which
15869 has the key type encoded in the ASN1 structure. Added DER private key
15870 support to pkcs8 application.
15874 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15875 ciphersuites has been selected (as required by the SSL 3/TLS 1
15876 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15877 is set, we interpret this as a request to violate the specification
15878 (the worst that can happen is a handshake failure, and 'correct'
15879 behaviour would result in a handshake failure anyway).
15883 * In SSL_CTX_add_session, take into account that there might be multiple
15884 SSL_SESSION structures with the same session ID (e.g. when two threads
15885 concurrently obtain them from an external cache).
15886 The internal cache can handle only one SSL_SESSION with a given ID,
15887 so if there's a conflict, we now throw out the old one to achieve
15892 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15893 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15894 some routines that use cipher OIDs: some ciphers do not have OIDs
15895 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15900 * Simplify the trust setting structure and code. Now we just have
15901 two sequences of OIDs for trusted and rejected settings. These will
15902 typically have values the same as the extended key usage extension
15903 and any application specific purposes.
15905 The trust checking code now has a default behaviour: it will just
15906 check for an object with the same NID as the passed id. Functions can
15907 be provided to override either the default behaviour or the behaviour
15908 for a given id. SSL client, server and email already have functions
15909 in place for compatibility: they check the NID and also return "trusted"
15910 if the certificate is self signed.
15914 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15915 traditional format into an EVP_PKEY structure.
15919 * Add a password callback function PEM_cb() which either prompts for
15920 a password if usr_data is NULL or otherwise assumes it is a null
15921 terminated password. Allow passwords to be passed on command line
15922 environment or config files in a few more utilities.
15926 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15927 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15928 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15929 Update documentation.
15933 * Support for ASN1 "NULL" type. This could be handled before by using
15934 ASN1_TYPE but there wasn't any function that would try to read a NULL
15935 and produce an error if it couldn't. For compatibility we also have
15936 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15937 don't allocate anything because they don't need to.
15941 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15944 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15946 * Rebuild of the memory allocation routines used by OpenSSL code and
15947 possibly others as well. The purpose is to make an interface that
15948 provide hooks so anyone can build a separate set of allocation and
15949 deallocation routines to be used by OpenSSL, for example memory
15950 pool implementations, or something else, which was previously hard
15951 since Malloc(), Realloc() and Free() were defined as macros having
15952 the values malloc, realloc and free, respectively (except for Win32
15953 compilations). The same is provided for memory debugging code.
15954 OpenSSL already comes with functionality to find memory leaks, but
15955 this gives people a chance to debug other memory problems.
15957 With these changes, a new set of functions and macros have appeared:
15959 CRYPTO_set_mem_debug_functions() [F]
15960 CRYPTO_get_mem_debug_functions() [F]
15961 CRYPTO_dbg_set_options() [F]
15962 CRYPTO_dbg_get_options() [F]
15963 CRYPTO_malloc_debug_init() [M]
15965 The memory debug functions are NULL by default, unless the library
15966 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15967 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15968 gives the standard debugging functions that come with OpenSSL) or
15969 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15970 provided by the library user) must be used. When the standard
15971 debugging functions are used, CRYPTO_dbg_set_options can be used to
15972 request additional information:
15973 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15974 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15976 Also, things like CRYPTO_set_mem_functions will always give the
15977 expected result (the new set of functions is used for allocation
15978 and deallocation) at all times, regardless of platform and compiler
15981 To finish it up, some functions that were never use in any other
15982 way than through macros have a new API and new semantic:
15984 CRYPTO_dbg_malloc()
15985 CRYPTO_dbg_realloc()
15988 All macros of value have retained their old syntax.
15990 *Richard Levitte and Bodo Moeller*
15992 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15993 ordering of SMIMECapabilities wasn't in "strength order" and there
15994 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15999 * Some ASN1 types with illegal zero length encoding (INTEGER,
16000 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16002 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16004 * Merge in my S/MIME library for OpenSSL. This provides a simple
16005 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16006 functionality to handle multipart/signed properly) and a utility
16007 called 'smime' to call all this stuff. This is based on code I
16008 originally wrote for Celo who have kindly allowed it to be
16009 included in OpenSSL.
16013 * Add variants des_set_key_checked and des_set_key_unchecked of
16014 des_set_key (aka des_key_sched). Global variable des_check_key
16015 decides which of these is called by des_set_key; this way
16016 des_check_key behaves as it always did, but applications and
16017 the library itself, which was buggy for des_check_key == 1,
16018 have a cleaner way to pick the version they need.
16022 * New function PKCS12_newpass() which changes the password of a
16027 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16028 dynamic mix. In both cases the ids can be used as an index into the
16029 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16030 functions so they accept a list of the field values and the
16031 application doesn't need to directly manipulate the X509_TRUST
16036 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16041 * Modify the way the V3 extension code looks up extensions. This now
16042 works in a similar way to the object code: we have some "standard"
16043 extensions in a static table which is searched with OBJ_bsearch()
16044 and the application can add dynamic ones if needed. The file
16045 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16046 updated whenever a new extension is added to the core code and kept
16047 in ext_nid order. There is a simple program 'tabtest.c' which checks
16048 this. New extensions are not added too often so this file can readily
16049 be maintained manually.
16051 There are two big advantages in doing things this way. The extensions
16052 can be looked up immediately and no longer need to be "added" using
16053 X509V3_add_standard_extensions(): this function now does nothing.
16054 Side note: I get *lots* of email saying the extension code doesn't
16055 work because people forget to call this function.
16056 Also no dynamic allocation is done unless new extensions are added:
16057 so if we don't add custom extensions there is no need to call
16058 X509V3_EXT_cleanup().
16062 * Modify enc utility's salting as follows: make salting the default. Add a
16063 magic header, so unsalted files fail gracefully instead of just decrypting
16064 to garbage. This is because not salting is a big security hole, so people
16065 should be discouraged from doing it.
16069 * Fixes and enhancements to the 'x509' utility. It allowed a message
16070 digest to be passed on the command line but it only used this
16071 parameter when signing a certificate. Modified so all relevant
16072 operations are affected by the digest parameter including the
16073 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16074 DSA key was used because it didn't fix the digest.
16078 * Initial certificate chain verify code. Currently tests the untrusted
16079 certificates for consistency with the verify purpose (which is set
16080 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16082 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16083 this is because it will reject chains with invalid extensions whereas
16084 every previous version of OpenSSL and SSLeay made no checks at all.
16086 Trust code: checks the root CA for the relevant trust settings. Trust
16087 settings have an initial value consistent with the verify purpose: e.g.
16088 if the verify purpose is for SSL client use it expects the CA to be
16089 trusted for SSL client use. However the default value can be changed to
16090 permit custom trust settings: one example of this would be to only trust
16091 certificates from a specific "secure" set of CAs.
16093 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16094 which should be used for version portability: especially since the
16095 verify structure is likely to change more often now.
16097 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16098 to set them. If not set then assume SSL clients will verify SSL servers
16101 Two new options to the verify program: -untrusted allows a set of
16102 untrusted certificates to be passed in and -purpose which sets the
16103 intended purpose of the certificate. If a purpose is set then the
16104 new chain verify code is used to check extension consistency.
16108 * Support for the authority information access extension.
16112 * Modify RSA and DSA PEM read routines to transparently handle
16113 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16114 public keys in a format compatible with certificate
16115 SubjectPublicKeyInfo structures. Unfortunately there were already
16116 functions called *_PublicKey_* which used various odd formats so
16117 these are retained for compatibility: however the DSA variants were
16118 never in a public release so they have been deleted. Changed dsa/rsa
16119 utilities to handle the new format: note no releases ever handled public
16120 keys so we should be OK.
16122 The primary motivation for this change is to avoid the same fiasco
16123 that dogs private keys: there are several incompatible private key
16124 formats some of which are standard and some OpenSSL specific and
16125 require various evil hacks to allow partial transparent handling and
16126 even then it doesn't work with DER formats. Given the option anything
16127 other than PKCS#8 should be dumped: but the other formats have to
16128 stay in the name of compatibility.
16130 With public keys and the benefit of hindsight one standard format
16131 is used which works with EVP_PKEY, RSA or DSA structures: though
16132 it clearly returns an error if you try to read the wrong kind of key.
16134 Added a -pubkey option to the 'x509' utility to output the public key.
16135 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16136 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16137 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16138 that do the same as the `EVP_PKEY_assign_*()` except they up the
16139 reference count of the added key (they don't "swallow" the
16144 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16145 CRLs would fail if the file contained no certificates or no CRLs:
16146 added a new function to read in both types and return the number
16147 read: this means that if none are read it will be an error. The
16148 DER versions of the certificate and CRL reader would always fail
16149 because it isn't possible to mix certificates and CRLs in DER format
16150 without choking one or the other routine. Changed this to just read
16151 a certificate: this is the best we can do. Also modified the code
16152 in apps/verify.c to take notice of return codes: it was previously
16153 attempting to read in certificates from NULL pointers and ignoring
16154 any errors: this is one reason why the cert and CRL reader seemed
16155 to work. It doesn't check return codes from the default certificate
16156 routines: these may well fail if the certificates aren't installed.
16160 * Code to support otherName option in GeneralName.
16164 * First update to verify code. Change the verify utility
16165 so it warns if it is passed a self signed certificate:
16166 for consistency with the normal behaviour. X509_verify
16167 has been modified to it will now verify a self signed
16168 certificate if *exactly* the same certificate appears
16169 in the store: it was previously impossible to trust a
16170 single self signed certificate. This means that:
16171 openssl verify ss.pem
16172 now gives a warning about a self signed certificate but
16173 openssl verify -CAfile ss.pem ss.pem
16178 * For servers, store verify_result in SSL_SESSION data structure
16179 (and add it to external session representation).
16180 This is needed when client certificate verifications fails,
16181 but an application-provided verification callback (set by
16182 SSL_CTX_set_cert_verify_callback) allows accepting the session
16183 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16184 but returns 1): When the session is reused, we have to set
16185 ssl->verify_result to the appropriate error code to avoid
16188 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16190 * Fix a bug in the new PKCS#7 code: it didn't consider the
16191 case in PKCS7_dataInit() where the signed PKCS7 structure
16192 didn't contain any existing data because it was being created.
16194 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16196 * Add a salt to the key derivation routines in enc.c. This
16197 forms the first 8 bytes of the encrypted file. Also add a
16198 -S option to allow a salt to be input on the command line.
16202 * New function X509_cmp(). Oddly enough there wasn't a function
16203 to compare two certificates. We do this by working out the SHA1
16204 hash and comparing that. X509_cmp() will be needed by the trust
16209 * SSL_get1_session() is like SSL_get_session(), but increments
16210 the reference count in the SSL_SESSION returned.
16212 *Geoff Thorpe <geoff@eu.c2.net>*
16214 * Fix for 'req': it was adding a null to request attributes.
16215 Also change the X509_LOOKUP and X509_INFO code to handle
16216 certificate auxiliary information.
16220 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16225 * Add the possibility to add extra information to the memory leak
16226 detecting output, to form tracebacks, showing from where each
16227 allocation was originated: CRYPTO_push_info("constant string") adds
16228 the string plus current file name and line number to a per-thread
16229 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16230 is like calling CYRPTO_pop_info() until the stack is empty.
16231 Also updated memory leak detection code to be multi-thread-safe.
16235 * Add options -text and -noout to pkcs7 utility and delete the
16236 encryption options which never did anything. Update docs.
16240 * Add options to some of the utilities to allow the pass phrase
16241 to be included on either the command line (not recommended on
16242 OSes like Unix) or read from the environment. Update the
16243 manpages and fix a few bugs.
16247 * Add a few manpages for some of the openssl commands.
16251 * Fix the -revoke option in ca. It was freeing up memory twice,
16252 leaking and not finding already revoked certificates.
16256 * Extensive changes to support certificate auxiliary information.
16257 This involves the use of X509_CERT_AUX structure and X509_AUX
16258 functions. An X509_AUX function such as PEM_read_X509_AUX()
16259 can still read in a certificate file in the usual way but it
16260 will also read in any additional "auxiliary information". By
16261 doing things this way a fair degree of compatibility can be
16262 retained: existing certificates can have this information added
16263 using the new 'x509' options.
16265 Current auxiliary information includes an "alias" and some trust
16266 settings. The trust settings will ultimately be used in enhanced
16267 certificate chain verification routines: currently a certificate
16268 can only be trusted if it is self signed and then it is trusted
16273 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16274 The problem was that one of the replacement routines had not been working
16275 since SSLeay releases. For now the offending routine has been replaced
16276 with non-optimised assembler. Even so, this now gives around 95%
16277 performance improvement for 1024 bit RSA signs.
16281 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16282 handling. Most clients have the effective key size in bits equal to
16283 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16284 A few however don't do this and instead use the size of the decrypted key
16285 to determine the RC2 key length and the AlgorithmIdentifier to determine
16286 the effective key length. In this case the effective key length can still
16287 be 40 bits but the key length can be 168 bits for example. This is fixed
16288 by manually forcing an RC2 key into the EVP_PKEY structure because the
16289 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16290 the key length and effective key length are equal.
16294 * Add a bunch of functions that should simplify the creation of
16295 X509_NAME structures. Now you should be able to do:
16296 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16297 and have it automatically work out the correct field type and fill in
16298 the structures. The more adventurous can try:
16299 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16300 and it will (hopefully) work out the correct multibyte encoding.
16304 * Change the 'req' utility to use the new field handling and multibyte
16305 copy routines. Before the DN field creation was handled in an ad hoc
16306 way in req, ca, and x509 which was rather broken and didn't support
16307 BMPStrings or UTF8Strings. Since some software doesn't implement
16308 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16309 using the dirstring_type option. See the new comment in the default
16310 openssl.cnf for more info.
16314 * Make crypto/rand/md_rand.c more robust:
16315 - Assure unique random numbers after fork().
16316 - Make sure that concurrent threads access the global counter and
16317 md serializably so that we never lose entropy in them
16318 or use exactly the same state in multiple threads.
16319 Access to the large state is not always serializable because
16320 the additional locking could be a performance killer, and
16321 md should be large enough anyway.
16325 * New file apps/app_rand.c with commonly needed functionality
16326 for handling the random seed file.
16328 Use the random seed file in some applications that previously did not:
16330 dsaparam -genkey (which also ignored its '-rand' option),
16333 x509 (when signing).
16334 Except on systems with /dev/urandom, it is crucial to have a random
16335 seed file at least for key creation, DSA signing, and for DH exchanges;
16336 for RSA signatures we could do without one.
16338 gendh and gendsa (unlike genrsa) used to read only the first byte
16339 of each file listed in the '-rand' option. The function as previously
16340 found in genrsa is now in app_rand.c and is used by all programs
16341 that support '-rand'.
16345 * In RAND_write_file, use mode 0600 for creating files;
16346 don't just chmod when it may be too late.
16350 * Report an error from X509_STORE_load_locations
16351 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16355 * New function ASN1_mbstring_copy() this copies a string in either
16356 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16357 into an ASN1_STRING type. A mask of permissible types is passed
16358 and it chooses the "minimal" type to use or an error if not type
16363 * Add function equivalents to the various macros in asn1.h. The old
16364 macros are retained with an `M_` prefix. Code inside the library can
16365 use the `M_` macros. External code (including the openssl utility)
16366 should *NOT* in order to be "shared library friendly".
16370 * Add various functions that can check a certificate's extensions
16371 to see if it usable for various purposes such as SSL client,
16372 server or S/MIME and CAs of these types. This is currently
16373 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16374 verification. Also added a -purpose flag to x509 utility to
16375 print out all the purposes.
16379 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16384 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16385 for, obtain and decode and extension and obtain its critical flag.
16386 This allows all the necessary extension code to be handled in a
16387 single function call.
16391 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16392 platforms. See crypto/rc4/rc4_enc.c for further details.
16396 * New -noout option to asn1parse. This causes no output to be produced
16397 its main use is when combined with -strparse and -out to extract data
16398 from a file (which may not be in ASN.1 format).
16402 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16403 when producing the local key id.
16405 *Richard Levitte <levitte@stacken.kth.se>*
16407 * New option -dhparam in s_server. This allows a DH parameter file to be
16408 stated explicitly. If it is not stated then it tries the first server
16409 certificate file. The previous behaviour hard coded the filename
16414 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16415 a public key to be input or output. For example:
16416 openssl rsa -in key.pem -pubout -out pubkey.pem
16417 Also added necessary DSA public key functions to handle this.
16421 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16422 in the message. This was handled by allowing
16423 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16425 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16427 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16428 to the end of the strings whereas this didn't. This would cause problems
16429 if strings read with d2i_ASN1_bytes() were later modified.
16431 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16433 * Fix for base64 decode bug. When a base64 bio reads only one line of
16434 data and it contains EOF it will end up returning an error. This is
16435 caused by input 46 bytes long. The cause is due to the way base64
16436 BIOs find the start of base64 encoded data. They do this by trying a
16437 trial decode on each line until they find one that works. When they
16438 do a flag is set and it starts again knowing it can pass all the
16439 data directly through the decoder. Unfortunately it doesn't reset
16440 the context it uses. This means that if EOF is reached an attempt
16441 is made to pass two EOFs through the context and this causes the
16442 resulting error. This can also cause other problems as well. As is
16443 usual with these problems it takes *ages* to find and the fix is
16444 trivial: move one line.
16446 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16448 * Ugly workaround to get s_client and s_server working under Windows. The
16449 old code wouldn't work because it needed to select() on sockets and the
16450 tty (for keypresses and to see if data could be written). Win32 only
16451 supports select() on sockets so we select() with a 1s timeout on the
16452 sockets and then see if any characters are waiting to be read, if none
16453 are present then we retry, we also assume we can always write data to
16454 the tty. This isn't nice because the code then blocks until we've
16455 received a complete line of data and it is effectively polling the
16456 keyboard at 1s intervals: however it's quite a bit better than not
16457 working at all :-) A dedicated Windows application might handle this
16458 with an event loop for example.
16462 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16463 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16464 will be called when RSA_sign() and RSA_verify() are used. This is useful
16465 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16466 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16467 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16468 This necessitated the support of an extra signature type NID_md5_sha1
16469 for SSL signatures and modifications to the SSL library to use it instead
16470 of calling RSA_public_decrypt() and RSA_private_encrypt().
16474 * Add new -verify -CAfile and -CApath options to the crl program, these
16475 will lookup a CRL issuers certificate and verify the signature in a
16476 similar way to the verify program. Tidy up the crl program so it
16477 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16478 less strict. It will now permit CRL extensions even if it is not
16479 a V2 CRL: this will allow it to tolerate some broken CRLs.
16483 * Initialize all non-automatic variables each time one of the openssl
16484 sub-programs is started (this is necessary as they may be started
16485 multiple times from the "OpenSSL>" prompt).
16487 *Lennart Bang, Bodo Moeller*
16489 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16490 removing all other RSA functionality (this is what NO_RSA does). This
16491 is so (for example) those in the US can disable those operations covered
16492 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16497 * Non-copying interface to BIO pairs.
16498 (still largely untested)
16502 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16503 ASCII string. This was handled independently in various places before.
16507 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16508 UTF8 strings a character at a time.
16512 * Use client_version from client hello to select the protocol
16513 (s23_srvr.c) and for RSA client key exchange verification
16514 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16518 * Add various utility functions to handle SPKACs, these were previously
16519 handled by poking round in the structure internals. Added new function
16520 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16521 print, verify and generate SPKACs. Based on an original idea from
16522 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16526 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16530 * Allow the config file extension section to be overwritten on the
16531 command line. Based on an original idea from Massimiliano Pala
16532 <madwolf@comune.modena.it>. The new option is called -extensions
16533 and can be applied to ca, req and x509. Also -reqexts to override
16534 the request extensions in req and -crlexts to override the crl extensions
16539 * Add new feature to the SPKAC handling in ca. Now you can include
16540 the same field multiple times by preceding it by "XXXX." for example:
16543 this is the same syntax as used in the req config file.
16547 * Allow certificate extensions to be added to certificate requests. These
16548 are specified in a 'req_extensions' option of the req section of the
16549 config file. They can be printed out with the -text option to req but
16550 are otherwise ignored at present.
16554 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16555 data read consists of only the final block it would not decrypted because
16556 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16557 A misplaced 'break' also meant the decrypted final block might not be
16558 copied until the next read.
16562 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16563 a few extra parameters to the DH structure: these will be useful if
16564 for example we want the value of 'q' or implement X9.42 DH.
16568 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16569 provides hooks that allow the default DSA functions or functions on a
16570 "per key" basis to be replaced. This allows hardware acceleration and
16571 hardware key storage to be handled without major modification to the
16572 library. Also added low level modexp hooks and CRYPTO_EX structure and
16573 associated functions.
16577 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16578 as "read only": it can't be written to and the buffer it points to will
16579 not be freed. Reading from a read only BIO is much more efficient than
16580 a normal memory BIO. This was added because there are several times when
16581 an area of memory needs to be read from a BIO. The previous method was
16582 to create a memory BIO and write the data to it, this results in two
16583 copies of the data and an O(n^2) reading algorithm. There is a new
16584 function BIO_new_mem_buf() which creates a read only memory BIO from
16585 an area of memory. Also modified the PKCS#7 routines to use read only
16590 * Bugfix: ssl23_get_client_hello did not work properly when called in
16591 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16592 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16593 but a retry condition occurred while trying to read the rest.
16597 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16598 NID_pkcs7_encrypted by default: this was wrong since this should almost
16599 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16600 the encrypted data type: this is a more sensible place to put it and it
16601 allows the PKCS#12 code to be tidied up that duplicated this
16606 * Changed obj_dat.pl script so it takes its input and output files on
16607 the command line. This should avoid shell escape redirection problems
16612 * Initial support for certificate extension requests, these are included
16613 in things like Xenroll certificate requests. Included functions to allow
16614 extensions to be obtained and added.
16618 * -crlf option to s_client and s_server for sending newlines as
16619 CRLF (as required by many protocols).
16623 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16625 * Install libRSAglue.a when OpenSSL is built with RSAref.
16627 *Ralf S. Engelschall*
16629 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16631 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16633 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16638 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16639 DH parameters/keys (q is lost during that conversion, but the resulting
16640 DH parameters contain its length).
16642 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16643 much faster than DH_generate_parameters (which creates parameters
16644 where `p = 2*q + 1`), and also the smaller q makes DH computations
16645 much more efficient (160-bit exponentiation instead of 1024-bit
16646 exponentiation); so this provides a convenient way to support DHE
16647 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16648 utter importance to use
16649 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16651 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16652 when such DH parameters are used, because otherwise small subgroup
16653 attacks may become possible!
16657 * Avoid memory leak in i2d_DHparams.
16661 * Allow the -k option to be used more than once in the enc program:
16662 this allows the same encrypted message to be read by multiple recipients.
16666 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16667 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16668 it will always use the numerical form of the OID, even if it has a short
16673 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16674 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16675 otherwise bn_mod_exp was called. In the case of hardware keys for example
16676 no private key components need be present and it might store extra data
16677 in the RSA structure, which cannot be accessed from bn_mod_exp.
16678 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16679 private key operations.
16683 * Added support for SPARC Linux.
16687 * pem_password_cb function type incompatibly changed from
16688 typedef int pem_password_cb(char *buf, int size, int rwflag);
16690 ....(char *buf, int size, int rwflag, void *userdata);
16691 so that applications can pass data to their callbacks:
16692 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16693 additional void * argument, which is just handed through whenever
16694 the password callback is called.
16696 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16698 New function SSL_CTX_set_default_passwd_cb_userdata.
16700 Compatibility note: As many C implementations push function arguments
16701 onto the stack in reverse order, the new library version is likely to
16702 interoperate with programs that have been compiled with the old
16703 pem_password_cb definition (PEM_whatever takes some data that
16704 happens to be on the stack as its last argument, and the callback
16705 just ignores this garbage); but there is no guarantee whatsoever that
16708 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16709 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16710 problems not only on Windows, but also on some Unix platforms.
16711 To avoid problematic command lines, these definitions are now in an
16712 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16713 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16717 * MIPS III/IV assembler module is reimplemented.
16721 * More DES library cleanups: remove references to srand/rand and
16722 delete an unused file.
16726 * Add support for the free Netwide assembler (NASM) under Win32,
16727 since not many people have MASM (ml) and it can be hard to obtain.
16728 This is currently experimental but it seems to work OK and pass all
16729 the tests. Check out INSTALL.W32 for info.
16733 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16734 without temporary keys kept an extra copy of the server key,
16735 and connections with temporary keys did not free everything in case
16740 * New function RSA_check_key and new openssl rsa option -check
16741 for verifying the consistency of RSA keys.
16743 *Ulf Moeller, Bodo Moeller*
16745 * Various changes to make Win32 compile work:
16746 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16747 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16748 comparison" warnings.
16749 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16753 * Add a debugging option to PKCS#5 v2 key generation function: when
16754 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16755 derived keys are printed to stderr.
16759 * Copy the flags in ASN1_STRING_dup().
16761 *Roman E. Pavlov <pre@mo.msk.ru>*
16763 * The x509 application mishandled signing requests containing DSA
16764 keys when the signing key was also DSA and the parameters didn't match.
16766 It was supposed to omit the parameters when they matched the signing key:
16767 the verifying software was then supposed to automatically use the CA's
16768 parameters if they were absent from the end user certificate.
16770 Omitting parameters is no longer recommended. The test was also
16771 the wrong way round! This was probably due to unusual behaviour in
16772 EVP_cmp_parameters() which returns 1 if the parameters match.
16773 This meant that parameters were omitted when they *didn't* match and
16774 the certificate was useless. Certificates signed with 'ca' didn't have
16777 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16779 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16780 The interface is as follows:
16781 Applications can use
16782 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16783 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16784 "off" is now the default.
16785 The library internally uses
16786 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16787 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16788 to disable memory-checking temporarily.
16790 Some inconsistent states that previously were possible (and were
16791 even the default) are now avoided.
16793 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16794 with each memory chunk allocated; this is occasionally more helpful
16795 than just having a counter.
16797 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16799 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16804 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16805 which largely parallels "options", but is for changing API behaviour,
16806 whereas "options" are about protocol behaviour.
16807 Initial "mode" flags are:
16809 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16810 a single record has been written.
16811 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16812 retries use the same buffer location.
16813 (But all of the contents must be
16818 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16821 * Fix problems with no-hmac etc.
16823 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16825 * New functions RSA_get_default_method(), RSA_set_method() and
16826 RSA_get_method(). These allows replacement of RSA_METHODs without having
16827 to mess around with the internals of an RSA structure.
16831 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16832 Also really enable memory leak checks in openssl.c and in some
16835 *Chad C. Mulligan, Bodo Moeller*
16837 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16838 up the length of negative integers. This has now been simplified to just
16839 store the length when it is first determined and use it later, rather
16840 than trying to keep track of where data is copied and updating it to
16842 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16844 * Add a new function PKCS7_signatureVerify. This allows the verification
16845 of a PKCS#7 signature but with the signing certificate passed to the
16846 function itself. This contrasts with PKCS7_dataVerify which assumes the
16847 certificate is present in the PKCS#7 structure. This isn't always the
16848 case: certificates can be omitted from a PKCS#7 structure and be
16849 distributed by "out of band" means (such as a certificate database).
16853 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16854 function prototypes in pem.h, also change util/mkdef.pl to add the
16855 necessary function names.
16859 * mk1mf.pl (used by Windows builds) did not properly read the
16860 options set by Configure in the top level Makefile, and Configure
16861 was not even able to write more than one option correctly.
16862 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16866 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16867 file to be loaded from a BIO or FILE pointer. The BIO version will
16868 for example allow memory BIOs to contain config info.
16872 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16873 Whoever hopes to achieve shared-library compatibility across versions
16874 must use this, not the compile-time macro.
16875 (Exercise 0.9.4: Which is the minimum library version required by
16877 Note: All this applies only to multi-threaded programs, others don't
16882 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16883 through a BIO pair triggered the default case, i.e.
16884 SSLerr(...,SSL_R_UNKNOWN_STATE).
16888 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16889 can use the SSL library even if none of the specific BIOs is
16894 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16895 for the encoded length.
16897 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16899 * Add initial documentation of the X509V3 functions.
16903 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16904 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16905 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16906 secure PKCS#8 private key format with a high iteration count.
16910 * Fix determination of Perl interpreter: A perl or perl5
16911 *directory* in $PATH was also accepted as the interpreter.
16913 *Ralf S. Engelschall*
16915 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16916 wrong with it but it was very old and did things like calling
16917 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16918 unusual formatting.
16922 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16923 to use the new extension code.
16927 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16928 with macros. This should make it easier to change their form, add extra
16929 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16934 * Add to configuration table a new entry that can specify an alternative
16935 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16936 according to Mark Crispin <MRC@Panda.COM>.
16940 * DES CBC did not update the IV. Weird.
16944 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16945 Changing the behaviour of the former might break existing programs --
16946 where IV updating is needed, des_ncbc_encrypt can be used.
16949 * When bntest is run from "make test" it drives bc to check its
16950 calculations, as well as internally checking them. If an internal check
16951 fails, it needs to cause bc to give a non-zero result or make test carries
16952 on without noticing the failure. Fixed.
16956 * DES library cleanups.
16960 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16961 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16962 ciphers. NOTE: although the key derivation function has been verified
16963 against some published test vectors it has not been extensively tested
16964 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16969 * Instead of "mkdir -p", which is not fully portable, use new
16970 Perl script "util/mkdir-p.pl".
16974 * Rewrite the way password based encryption (PBE) is handled. It used to
16975 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16976 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16977 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16978 the 'parameter' field of the AlgorithmIdentifier is passed to the
16979 underlying key generation function so it must do its own ASN1 parsing.
16980 This has also changed the EVP_PBE_CipherInit() function which now has a
16981 'parameter' argument instead of literal salt and iteration count values
16982 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16986 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16987 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16988 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16989 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16990 value was just used as a "magic string" and not used directly its
16991 value doesn't matter.
16995 * Introduce some semblance of const correctness to BN. Shame C doesn't
17000 * "linux-sparc64" configuration (ultrapenguin).
17002 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17003 "linux-sparc" configuration.
17005 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17007 * config now generates no-xxx options for missing ciphers.
17011 * Support the EBCDIC character set (work in progress).
17012 File ebcdic.c not yet included because it has a different license.
17014 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17016 * Support BS2000/OSD-POSIX.
17018 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17020 * Make callbacks for key generation use `void *` instead of `char *`.
17024 * Make S/MIME samples compile (not yet tested).
17028 * Additional typesafe stacks.
17032 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17036 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17038 * New configuration variant "sco5-gcc".
17040 * Updated some demos.
17042 *Sean O Riordain, Wade Scholine*
17044 * Add missing BIO_free at exit of pkcs12 application.
17048 * Fix memory leak in conf.c.
17052 * Updates for Win32 to assembler version of MD5.
17056 * Set #! path to perl in apps/der_chop to where we found it
17057 instead of using a fixed path.
17061 * SHA library changes for irix64-mips4-cc.
17065 * Improvements for VMS support.
17069 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17071 * Bignum library bug fix. IRIX 6 passes "make test" now!
17072 This also avoids the problems with SC4.2 and unpatched SC5.
17074 *Andy Polyakov <appro@fy.chalmers.se>*
17076 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17077 These are required because of the typesafe stack would otherwise break
17078 existing code. If old code used a structure member which used to be STACK
17079 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17080 sk_num or sk_value it would produce an error because the num, data members
17081 are not present in STACK_OF. Now it just produces a warning. sk_set
17082 replaces the old method of assigning a value to sk_value
17083 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17084 that does this will no longer work (and should use sk_set instead) but
17085 this could be regarded as a "questionable" behaviour anyway.
17089 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17090 correctly handle encrypted S/MIME data.
17094 * Change type of various DES function arguments from des_cblock
17095 (which means, in function argument declarations, pointer to char)
17096 to des_cblock * (meaning pointer to array with 8 char elements),
17097 which allows the compiler to do more typechecking; it was like
17098 that back in SSLeay, but with lots of ugly casts.
17100 Introduce new type const_des_cblock.
17104 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17105 problems: find RecipientInfo structure that matches recipient certificate
17106 and initialise the ASN1 structures properly based on passed cipher.
17110 * Belatedly make the BN tests actually check the results.
17114 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17115 to and from BNs: it was completely broken. New compilation option
17116 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17117 key elements as negative integers.
17121 * Reorganize and speed up MD5.
17123 *Andy Polyakov <appro@fy.chalmers.se>*
17127 *Richard Levitte <richard@levitte.org>*
17129 * New option -out to asn1parse to allow the parsed structure to be
17130 output to a file. This is most useful when combined with the -strparse
17131 option to examine the output of things like OCTET STRINGS.
17135 * Make SSL library a little more fool-proof by not requiring any longer
17136 that `SSL_set_{accept,connect}_state` be called before
17137 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17138 in many applications because usually everything *appeared* to work as
17139 intended anyway -- now it really works as intended).
17143 * Move openssl.cnf out of lib/.
17147 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17148 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17149 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17151 *Ralf S. Engelschall*
17153 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17154 handle PKCS#7 enveloped data properly.
17156 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17158 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17159 copying pointers. The cert_st handling is changed by this in
17160 various ways (and thus what used to be known as ctx->default_cert
17161 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17162 any longer when s->cert does not give us what we need).
17163 ssl_cert_instantiate becomes obsolete by this change.
17164 As soon as we've got the new code right (possibly it already is?),
17165 we have solved a couple of bugs of the earlier code where s->cert
17166 was used as if it could not have been shared with other SSL structures.
17168 Note that using the SSL API in certain dirty ways now will result
17169 in different behaviour than observed with earlier library versions:
17170 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17171 does not influence s as it used to.
17173 In order to clean up things more thoroughly, inside SSL_SESSION
17174 we don't use CERT any longer, but a new structure SESS_CERT
17175 that holds per-session data (if available); currently, this is
17176 the peer's certificate chain and, for clients, the server's certificate
17177 and temporary key. CERT holds only those values that can have
17178 meaningful defaults in an SSL_CTX.
17182 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17183 from the internal representation. Various PKCS#7 fixes: remove some
17184 evil casts and set the enc_dig_alg field properly based on the signing
17189 * Allow PKCS#12 password to be set from the command line or the
17190 environment. Let 'ca' get its config file name from the environment
17191 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17196 * Allow certificate policies extension to use an IA5STRING for the
17197 organization field. This is contrary to the PKIX definition but
17198 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17203 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17204 without disallowing inline assembler and the like for non-pedantic builds.
17208 * Support Borland C++ builder.
17210 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17216 * SHA-1 cleanups and performance enhancements.
17218 *Andy Polyakov <appro@fy.chalmers.se>*
17220 * Sparc v8plus assembler for the bignum library.
17222 *Andy Polyakov <appro@fy.chalmers.se>*
17224 * Accept any -xxx and +xxx compiler options in Configure.
17228 * Update HPUX configuration.
17232 * Add missing `sk_<type>_unshift()` function to safestack.h
17234 *Ralf S. Engelschall*
17236 * New function SSL_CTX_use_certificate_chain_file that sets the
17237 "extra_cert"s in addition to the certificate. (This makes sense
17238 only for "PEM" format files, as chains as a whole are not
17243 * Support verify_depth from the SSL API.
17244 x509_vfy.c had what can be considered an off-by-one-error:
17245 Its depth (which was not part of the external interface)
17246 was actually counting the number of certificates in a chain;
17247 now it really counts the depth.
17251 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17252 instead of X509err, which often resulted in confusing error
17253 messages since the error codes are not globally unique
17254 (e.g. an alleged error in ssl3_accept when a certificate
17255 didn't match the private key).
17257 * New function SSL_CTX_set_session_id_context that allows to set a default
17258 value (so that you don't need SSL_set_session_id_context for each
17259 connection using the SSL_CTX).
17263 * OAEP decoding bug fix.
17267 * Support INSTALL_PREFIX for package builders, as proposed by
17272 * New Configure options "threads" and "no-threads". For systems
17273 where the proper compiler options are known (currently Solaris
17274 and Linux), "threads" is the default.
17278 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17282 * Install various scripts to $(OPENSSLDIR)/misc, not to
17283 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17284 such as /usr/local/bin.
17288 * "make linux-shared" to build shared libraries.
17290 *Niels Poppe <niels@netbox.org>*
17292 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17296 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17297 extension adding in x509 utility.
17301 * Remove NOPROTO sections and error code comments.
17305 * Partial rewrite of the DEF file generator to now parse the ANSI
17310 * New Configure options --prefix=DIR and --openssldir=DIR.
17314 * Complete rewrite of the error code script(s). It is all now handled
17315 by one script at the top level which handles error code gathering,
17316 header rewriting and C source file generation. It should be much better
17317 than the old method: it now uses a modified version of Ulf's parser to
17318 read the ANSI prototypes in all header files (thus the old K&R definitions
17319 aren't needed for error creation any more) and do a better job of
17320 translating function codes into names. The old 'ASN1 error code embedded
17321 in a comment' is no longer necessary and it doesn't use .err files which
17322 have now been deleted. Also the error code call doesn't have to appear all
17323 on one line (which resulted in some large lines...).
17327 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17331 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17332 0 (which usually indicates a closed connection), but continue reading.
17336 * Fix some race conditions.
17340 * Add support for CRL distribution points extension. Add Certificate
17341 Policies and CRL distribution points documentation.
17345 * Move the autogenerated header file parts to crypto/opensslconf.h.
17349 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17350 8 of keying material. Merlin has also confirmed interop with this fix
17351 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17353 *Merlin Hughes <merlin@baltimore.ie>*
17355 * Fix lots of warnings.
17357 *Richard Levitte <levitte@stacken.kth.se>*
17359 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17360 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17362 *Richard Levitte <levitte@stacken.kth.se>*
17364 * Fix problems with sizeof(long) == 8.
17366 *Andy Polyakov <appro@fy.chalmers.se>*
17368 * Change functions to ANSI C.
17372 * Fix typos in error codes.
17374 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17376 * Remove defunct assembler files from Configure.
17380 * SPARC v8 assembler BIGNUM implementation.
17382 *Andy Polyakov <appro@fy.chalmers.se>*
17384 * Support for Certificate Policies extension: both print and set.
17385 Various additions to support the r2i method this uses.
17389 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17390 return a const string when you are expecting an allocated buffer.
17394 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17395 types DirectoryString and DisplayText.
17399 * Add code to allow r2i extensions to access the configuration database,
17400 add an LHASH database driver and add several ctx helper functions.
17404 * Fix an evil bug in bn_expand2() which caused various BN functions to
17405 fail when they extended the size of a BIGNUM.
17409 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17410 support typesafe stack.
17414 * Fix typo in SSL_[gs]et_options().
17416 *Nils Frostberg <nils@medcom.se>*
17418 * Delete various functions and files that belonged to the (now obsolete)
17419 old X509V3 handling code.
17423 * New Configure option "rsaref".
17427 * Don't auto-generate pem.h.
17431 * Introduce type-safe ASN.1 SETs.
17435 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17437 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17439 * Introduce type-safe STACKs. This will almost certainly break lots of code
17440 that links with OpenSSL (well at least cause lots of warnings), but fear
17441 not: the conversion is trivial, and it eliminates loads of evil casts. A
17442 few STACKed things have been converted already. Feel free to convert more.
17443 In the fullness of time, I'll do away with the STACK type altogether.
17447 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17448 specified in `<certfile>` by updating the entry in the index.txt file.
17449 This way one no longer has to edit the index.txt file manually for
17450 revoking a certificate. The -revoke option does the gory details now.
17452 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17454 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17455 `-text` option at all and this way the `-noout -text` combination was
17456 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17458 *Ralf S. Engelschall*
17460 * Make sure a corresponding plain text error message exists for the
17461 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17462 verify callback function determined that a certificate was revoked.
17464 *Ralf S. Engelschall*
17466 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17467 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17468 all available ciphers including rc5, which was forgotten until now.
17469 In order to let the testing shell script know which algorithms
17470 are available, a new (up to now undocumented) command
17471 `openssl list-cipher-commands` is used.
17475 * Bugfix: s_client occasionally would sleep in select() when
17476 it should have checked SSL_pending() first.
17480 * New functions DSA_do_sign and DSA_do_verify to provide access to
17481 the raw DSA values prior to ASN.1 encoding.
17485 * Tweaks to Configure
17487 *Niels Poppe <niels@netbox.org>*
17489 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17494 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17498 * New config option to avoid instructions that are illegal on the 80386.
17499 The default code is faster, but requires at least a 486.
17503 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17504 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17505 same as SSL2_VERSION anyway.
17509 * New "-showcerts" option for s_client.
17513 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17514 application. Various cleanups and fixes.
17518 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17519 modify error routines to work internally. Add error codes and PBE init
17520 to library startup routines.
17524 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17525 packing functions to asn1 and evp. Changed function names and error
17526 codes along the way.
17530 * PKCS12 integration: and so it begins... First of several patches to
17531 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17532 objects to objects.h
17536 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17537 and display support for Thawte strong extranet extension.
17541 * Add LinuxPPC support.
17543 *Jeff Dubrule <igor@pobox.org>*
17545 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17546 bn_div_words in alpha.s.
17548 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17550 * Make sure the RSA OAEP test is skipped under -DRSAref because
17551 OAEP isn't supported when OpenSSL is built with RSAref.
17553 *Ulf Moeller <ulf@fitug.de>*
17555 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17556 so they no longer are missing under -DNOPROTO.
17558 *Soren S. Jorvang <soren@t.dk>*
17560 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17562 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17563 doesn't work when the session is reused. Coming soon!
17567 * Fix a security hole, that allows sessions to be reused in the wrong
17568 context thus bypassing client cert protection! All software that uses
17569 client certs and session caches in multiple contexts NEEDS PATCHING to
17570 allow session reuse! A fuller solution is in the works.
17572 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17574 * Some more source tree cleanups (removed obsolete files
17575 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17576 permission on "config" script to be executable) and a fix for the INSTALL
17579 *Ulf Moeller <ulf@fitug.de>*
17581 * Remove some legacy and erroneous uses of malloc, free instead of
17584 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17586 * Make rsa_oaep_test return non-zero on error.
17588 *Ulf Moeller <ulf@fitug.de>*
17590 * Add support for native Solaris shared libraries. Configure
17591 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17592 if someone would make that last step automatic.
17594 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17596 * ctx_size was not built with the right compiler during "make links". Fixed.
17600 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17601 except NULL ciphers". This means the default cipher list will no longer
17602 enable NULL ciphers. They need to be specifically enabled e.g. with
17603 the string "DEFAULT:eNULL".
17607 * Fix to RSA private encryption routines: if p < q then it would
17608 occasionally produce an invalid result. This will only happen with
17609 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17613 * Be less restrictive and allow also `perl util/perlpath.pl
17614 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17615 because this way one can also use an interpreter named `perl5` (which is
17616 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17617 installed as `perl`).
17619 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17621 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17623 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17625 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17626 advapi32.lib to Win32 build and change the pem test comparison
17627 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17628 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17629 and crypto/des/ede_cbcm_enc.c.
17633 * DES quad checksum was broken on big-endian architectures. Fixed.
17637 * Comment out two functions in bio.h that aren't implemented. Fix up the
17638 Win32 test batch file so it (might) work again. The Win32 test batch file
17639 is horrible: I feel ill....
17643 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17644 in e_os.h. Audit of header files to check ANSI and non ANSI
17645 sections: 10 functions were absent from non ANSI section and not exported
17646 from Windows DLLs. Fixed up libeay.num for new functions.
17650 * Make `openssl version` output lines consistent.
17652 *Ralf S. Engelschall*
17654 * Fix Win32 symbol export lists for BIO functions: Added
17655 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17656 to ms/libeay{16,32}.def.
17658 *Ralf S. Engelschall*
17660 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17661 fine under Unix and passes some trivial tests I've now added. But the
17662 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17663 added to make sure no one expects that this stuff really works in the
17664 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17665 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17668 *Ralf S. Engelschall*
17670 * Fix the generation of two part addresses in perl.
17672 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17674 * Add config entry for Linux on MIPS.
17676 *John Tobey <jtobey@channel1.com>*
17678 * Make links whenever Configure is run, unless we are on Windoze.
17682 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17683 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17688 * Add a useful kludge to allow package maintainers to specify compiler and
17689 other platforms details on the command line without having to patch the
17690 Configure script every time: One now can use
17691 `perl Configure <id>:<details>`,
17692 i.e. platform ids are allowed to have details appended
17693 to them (separated by colons). This is treated as there would be a static
17694 pre-configured entry in Configure's %table under key `<id>` with value
17695 `<details>` and `perl Configure <id>` is called. So, when you want to
17696 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17697 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17698 now, which overrides the FreeBSD-elf entry on-the-fly.
17700 *Ralf S. Engelschall*
17702 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17706 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17707 on the `perl Configure ...` command line. This way one can compile
17708 OpenSSL libraries with Position Independent Code (PIC) which is needed
17709 for linking it into DSOs.
17711 *Ralf S. Engelschall*
17713 * Remarkably, export ciphers were totally broken and no-one had noticed!
17718 * Cleaned up the LICENSE document: The official contact for any license
17719 questions now is the OpenSSL core team under openssl-core@openssl.org.
17720 And add a paragraph about the dual-license situation to make sure people
17721 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17722 to the OpenSSL toolkit.
17724 *Ralf S. Engelschall*
17726 * General source tree makefile cleanups: Made `making xxx in yyy...`
17727 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17728 Additionally cleaned up the `make links` target: Remove unnecessary
17729 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17730 to speed processing and no longer clutter the display with confusing
17731 stuff. Instead only the actually done links are displayed.
17733 *Ralf S. Engelschall*
17735 * Permit null encryption ciphersuites, used for authentication only. It used
17736 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17737 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17742 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17743 signed attributes when verifying signatures (this would break them),
17744 the detached data encoding was wrong and public keys obtained using
17745 X509_get_pubkey() weren't freed.
17749 * Add text documentation for the BUFFER functions. Also added a work around
17750 to a Win95 console bug. This was triggered by the password read stuff: the
17751 last character typed gets carried over to the next fread(). If you were
17752 generating a new cert request using 'req' for example then the last
17753 character of the passphrase would be CR which would then enter the first
17758 * Added the new 'Includes OpenSSL Cryptography Software' button as
17759 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17760 button and can be used by applications based on OpenSSL to show the
17761 relationship to the OpenSSL project.
17763 *Ralf S. Engelschall*
17765 * Remove confusing variables in function signatures in files
17766 ssl/ssl_lib.c and ssl/ssl.h.
17768 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17770 * Don't install bss_file.c under PREFIX/include/
17772 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17774 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17775 functions that return function pointers and has support for NT specific
17776 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17777 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17778 unsigned to signed types: this was killing the Win32 compile.
17782 * Add new certificate file to stack functions,
17783 SSL_add_dir_cert_subjects_to_stack() and
17784 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17785 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17786 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17787 This means that Apache-SSL and similar packages don't have to mess around
17788 to add as many CAs as they want to the preferred list.
17792 * Experiment with doxygen documentation. Currently only partially applied to
17794 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17795 openssl.doxy as the configuration file.
17799 * Get rid of remaining C++-style comments which strict C compilers hate.
17801 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17803 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17804 compiled in by default: it has problems with large keys.
17808 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17809 DH private keys and/or callback functions which directly correspond to
17810 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17811 is needed for applications which have to configure certificates on a
17812 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17814 For the RSA certificate situation is makes no difference, but
17815 for the DSA certificate situation this fixes the "no shared cipher"
17816 problem where the OpenSSL cipher selection procedure failed because the
17817 temporary keys were not overtaken from the context and the API provided
17818 no way to reconfigure them.
17819 The new functions now let applications reconfigure the stuff and they
17820 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17821 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17822 non-public-API function ssl_cert_instantiate() is used as a helper
17823 function and also to reduce code redundancy inside ssl_rsa.c.
17825 *Ralf S. Engelschall*
17827 * Move s_server -dcert and -dkey options out of the undocumented feature
17828 area because they are useful for the DSA situation and should be
17829 recognized by the users.
17831 *Ralf S. Engelschall*
17833 * Fix the cipher decision scheme for export ciphers: the export bits are
17834 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17835 SSL_EXP_MASK. So, the original variable has to be used instead of the
17836 already masked variable.
17838 *Richard Levitte <levitte@stacken.kth.se>*
17840 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17842 *Richard Levitte <levitte@stacken.kth.se>*
17844 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17845 from `int` to `unsigned int` because it is a length and initialized by
17846 EVP_DigestFinal() which expects an `unsigned int *`.
17848 *Richard Levitte <levitte@stacken.kth.se>*
17850 * Don't hard-code path to Perl interpreter on shebang line of Configure
17851 script. Instead use the usual Shell->Perl transition trick.
17853 *Ralf S. Engelschall*
17855 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17856 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17857 -noout -modulus` as it's already the case for `openssl rsa -noout
17858 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17859 currently the public key is printed (a decision which was already done by
17860 `openssl dsa -modulus` in the past) which serves a similar purpose.
17861 Additionally the NO_RSA no longer completely removes the whole -modulus
17862 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17865 *Ralf S. Engelschall*
17867 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17868 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17870 *Arne Ansper <arne@ats.cyber.ee>*
17872 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17873 to be added. Now both 'req' and 'ca' can use new objects defined in the
17878 * Add cool BIO that does syslog (or event log on NT).
17880 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17882 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17883 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17884 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17885 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17889 * Add preliminary config info for new extension code.
17893 * Make RSA_NO_PADDING really use no padding.
17895 *Ulf Moeller <ulf@fitug.de>*
17897 * Generate errors when private/public key check is done.
17901 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17902 for some CRL extensions and new objects added.
17906 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17907 key usage extension and fuller support for authority key id.
17911 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17912 padding method for RSA, which is recommended for new applications in PKCS
17913 #1 v2.0 (RFC 2437, October 1998).
17914 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17915 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17916 against Bleichbacher's attack on RSA.
17917 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17920 * Updates to the new SSL compression code
17922 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17924 * Fix so that the version number in the master secret, when passed
17925 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17926 (because the server will not accept higher), that the version number
17927 is 0x03,0x01, not 0x03,0x00
17929 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17931 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17932 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17933 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17937 * Support for RAW extensions where an arbitrary extension can be
17938 created by including its DER encoding. See apps/openssl.cnf for
17943 * Make sure latest Perl versions don't interpret some generated C array
17944 code as Perl array code in the crypto/err/err_genc.pl script.
17946 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17948 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17949 not many people have the assembler. Various Win32 compilation fixes and
17950 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17951 build instructions.
17955 * Modify configure script 'Configure' to automatically create crypto/date.h
17956 file under Win32 and also build pem.h from pem.org. New script
17957 util/mkfiles.pl to create the MINFO file on environments that can't do a
17958 'make files': perl util/mkfiles.pl >MINFO should work.
17962 * Major rework of DES function declarations, in the pursuit of correctness
17963 and purity. As a result, many evil casts evaporated, and some weirdness,
17964 too. You may find this causes warnings in your code. Zapping your evil
17965 casts will probably fix them. Mostly.
17969 * Fix for a typo in asn1.h. Bug fix to object creation script
17970 obj_dat.pl. It considered a zero in an object definition to mean
17971 "end of object": none of the objects in objects.h have any zeros
17972 so it wasn't spotted.
17974 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17976 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17977 Masking (CBCM). In the absence of test vectors, the best I have been able
17978 to do is check that the decrypt undoes the encrypt, so far. Send me test
17979 vectors if you have them.
17983 * Correct calculation of key length for export ciphers (too much space was
17984 allocated for null ciphers). This has not been tested!
17988 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17989 message is now correct (it understands "crypto" and "ssl" on its
17990 command line). There is also now an "update" option. This will update
17991 the util/ssleay.num and util/libeay.num files with any new functions.
17993 perl util/mkdef.pl crypto ssl update
17994 it will update them.
17998 * Overhauled the Perl interface:
17999 - ported BN stuff to OpenSSL's different BN library
18000 - made the perl/ source tree CVS-aware
18001 - renamed the package from SSLeay to OpenSSL (the files still contain
18002 their history because I've copied them in the repository)
18003 - removed obsolete files (the test scripts will be replaced
18004 by better Test::Harness variants in the future)
18006 *Ralf S. Engelschall*
18008 * First cut for a very conservative source tree cleanup:
18009 1. merge various obsolete readme texts into doc/ssleay.txt
18010 where we collect the old documents and readme texts.
18011 2. remove the first part of files where I'm already sure that we no
18012 longer need them because of three reasons: either they are just temporary
18013 files which were left by Eric or they are preserved original files where
18014 I've verified that the diff is also available in the CVS via "cvs diff
18015 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18016 the crypto/md/ stuff).
18018 *Ralf S. Engelschall*
18020 * More extension code. Incomplete support for subject and issuer alt
18021 name, issuer and authority key id. Change the i2v function parameters
18022 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18023 what that's for :-) Fix to ASN1 macro which messed up
18024 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18028 * Preliminary support for ENUMERATED type. This is largely copied from the
18033 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18035 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18037 * Make sure `make rehash` target really finds the `openssl` program.
18039 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18041 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18042 like to hear about it if this slows down other processors.
18046 * Add CygWin32 platform information to Configure script.
18048 *Alan Batie <batie@aahz.jf.intel.com>*
18050 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18052 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18054 * New program nseq to manipulate netscape certificate sequences
18058 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18063 * Fixes to BN code. Previously the default was to define BN_RECURSION
18064 but the BN code had some problems that would cause failures when
18065 doing certificate verification and some other functions.
18067 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18069 * Add ASN1 and PEM code to support netscape certificate sequences.
18073 * Add ASN1 and PEM code to support netscape certificate sequences.
18077 * Add several PKIX and private extended key usage OIDs.
18081 * Modify the 'ca' program to handle the new extension code. Modify
18082 openssl.cnf for new extension format, add comments.
18086 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18087 and add a sample to openssl.cnf so req -x509 now adds appropriate
18092 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18093 error code, add initial support to X509_print() and x509 application.
18097 * Takes a deep breath and start adding X509 V3 extension support code. Add
18098 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18099 stuff is currently isolated and isn't even compiled yet.
18103 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18104 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18105 Removed the versions check from X509 routines when loading extensions:
18106 this allows certain broken certificates that don't set the version
18107 properly to be processed.
18111 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18112 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18113 can still be regenerated with "make depend".
18117 * Spelling mistake in C version of CAST-128.
18119 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18121 * Changes to the error generation code. The perl script err-code.pl
18122 now reads in the old error codes and retains the old numbers, only
18123 adding new ones if necessary. It also only changes the .err files if new
18124 codes are added. The makefiles have been modified to only insert errors
18125 when needed (to avoid needlessly modifying header files). This is done
18126 by only inserting errors if the .err file is newer than the auto generated
18127 C file. To rebuild all the error codes from scratch (the old behaviour)
18128 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18129 or delete all the .err files.
18133 * CAST-128 was incorrectly implemented for short keys. The C version has
18134 been fixed, but is untested. The assembler versions are also fixed, but
18135 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18136 to regenerate it if needed.
18137 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18138 Hagino <itojun@kame.net>*
18140 * File was opened incorrectly in randfile.c.
18142 *Ulf Möller <ulf@fitug.de>*
18144 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18145 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18146 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18147 al: it's just almost always a UTCTime. Note this patch adds new error
18148 codes so do a "make errors" if there are problems.
18152 * Correct Linux 1 recognition in config.
18154 *Ulf Möller <ulf@fitug.de>*
18156 * Remove pointless MD5 hash when using DSA keys in ca.
18158 *Anonymous <nobody@replay.com>*
18160 * Generate an error if given an empty string as a cert directory. Also
18161 generate an error if handed NULL (previously returned 0 to indicate an
18162 error, but didn't set one).
18164 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18166 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18170 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18171 parameters. This was causing a warning which killed off the Win32 compile.
18175 * Remove C++ style comments from crypto/bn/bn_local.h.
18177 *Neil Costigan <neil.costigan@celocom.com>*
18179 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18180 based on a text string, looking up short and long names and finally
18181 "dot" format. The "dot" format stuff didn't work. Added new function
18182 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18183 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18184 OID is not part of the table.
18188 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18189 X509_LOOKUP_by_alias().
18193 * Sort openssl functions by name.
18197 * Get the gendsa program working (hopefully) and add it to app list. Remove
18198 encryption from sample DSA keys (in case anyone is interested the password
18203 * Make *all* `*_free` functions accept a NULL pointer.
18205 *Frans Heymans <fheymans@isaserver.be>*
18207 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18210 *Anonymous <nobody@replay.com>*
18212 * s_server should send the CAfile as acceptable CAs, not its own cert.
18214 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18216 * Don't blow it for numeric -newkey arguments to apps/req.
18218 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18220 * Temp key "for export" tests were wrong in s3_srvr.c.
18222 *Anonymous <nobody@replay.com>*
18224 * Add prototype for temp key callback functions
18225 SSL_CTX_set_tmp_{rsa,dh}_callback().
18229 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18230 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18234 * X509_name_add_entry() freed the wrong thing after an error.
18236 *Arne Ansper <arne@ats.cyber.ee>*
18238 * rsa_eay.c would attempt to free a NULL context.
18240 *Arne Ansper <arne@ats.cyber.ee>*
18242 * BIO_s_socket() had a broken should_retry() on Windoze.
18244 *Arne Ansper <arne@ats.cyber.ee>*
18246 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18248 *Arne Ansper <arne@ats.cyber.ee>*
18250 * Make sure the already existing X509_STORE->depth variable is initialized
18251 in X509_STORE_new(), but document the fact that this variable is still
18252 unused in the certificate verification process.
18254 *Ralf S. Engelschall*
18256 * Fix the various library and apps files to free up pkeys obtained from
18257 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18261 * Fix reference counting in X509_PUBKEY_get(). This makes
18262 demos/maurice/example2.c work, amongst others, probably.
18264 *Steve Henson and Ben Laurie*
18266 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18267 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18268 are no longer created. This way we have a single and consistent command
18269 line interface `openssl <command>`, similar to `cvs <command>`.
18271 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18273 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18274 BIT STRING wrapper always have zero unused bits.
18278 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18282 * Make the top-level INSTALL documentation easier to understand.
18286 * Makefiles updated to exit if an error occurs in a sub-directory
18287 make (including if user presses ^C) [Paul Sutton]
18289 * Make Montgomery context stuff explicit in RSA data structure.
18293 * Fix build order of pem and err to allow for generated pem.h.
18297 * Fix renumbering bug in X509_NAME_delete_entry().
18301 * Enhanced the err-ins.pl script so it makes the error library number
18302 global and can add a library name. This is needed for external ASN1 and
18303 other error libraries.
18307 * Fixed sk_insert which never worked properly.
18311 * Fix ASN1 macros so they can handle indefinite length constructed
18312 EXPLICIT tags. Some non standard certificates use these: they can now
18317 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18318 into a single doc/ssleay.txt bundle. This way the information is still
18319 preserved but no longer messes up this directory. Now it's new room for
18320 the new set of documentation files.
18322 *Ralf S. Engelschall*
18324 * SETs were incorrectly DER encoded. This was a major pain, because they
18325 shared code with SEQUENCEs, which aren't coded the same. This means that
18326 almost everything to do with SETs or SEQUENCEs has either changed name or
18327 number of arguments.
18329 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18331 * Fix test data to work with the above.
18335 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18336 was already fixed by Eric for 0.9.1 it seems.
18338 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18340 * Autodetect FreeBSD3.
18344 * Fix various bugs in Configure. This affects the following platforms:
18348 unixware-2.0-pentium
18353 * Eliminate generated files from CVS. Reorder tests to regenerate files
18354 before they are needed.
18358 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18362 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18364 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18365 changed SSLeay to OpenSSL in version strings.
18367 *Ralf S. Engelschall*
18369 * Some fixups to the top-level documents.
18373 * Fixed the nasty bug where rsaref.h was not found under compile-time
18374 because the symlink to include/ was missing.
18376 *Ralf S. Engelschall*
18378 * Incorporated the popular no-RSA/DSA-only patches
18379 which allow to compile a RSA-free SSLeay.
18381 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18383 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18384 when "ssleay" is still not found.
18386 *Ralf S. Engelschall*
18388 * Added more platforms to Configure: Cray T3E, HPUX 11,
18390 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18392 * Updated the README file.
18394 *Ralf S. Engelschall*
18396 * Added various .cvsignore files in the CVS repository subdirs
18397 to make a "cvs update" really silent.
18399 *Ralf S. Engelschall*
18401 * Recompiled the error-definition header files and added
18402 missing symbols to the Win32 linker tables.
18404 *Ralf S. Engelschall*
18406 * Cleaned up the top-level documents;
18407 o new files: CHANGES and LICENSE
18408 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18409 o merged COPYRIGHT into LICENSE
18410 o removed obsolete TODO file
18411 o renamed MICROSOFT to INSTALL.W32
18413 *Ralf S. Engelschall*
18415 * Removed dummy files from the 0.9.1b source tree:
18416 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18417 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18418 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18419 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18420 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18422 *Ralf S. Engelschall*
18424 * Added various platform portability fixes.
18428 * The Genesis of the OpenSSL rpject:
18429 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18430 Young and Tim J. Hudson created while they were working for C2Net until
18433 *The OpenSSL Project*
18435 ### Changes between 0.9.0b and 0.9.1b [not released]
18437 * Updated a few CA certificates under certs/
18441 * Changed some BIGNUM api stuff.
18445 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18446 DGUX x86, Linux Alpha, etc.
18450 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18451 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18456 * Add -strparse option to asn1pars program which parses nested
18459 *Dr Stephen Henson <shenson@bigfoot.com>*
18461 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18465 * DSA fix for "ca" program.
18469 * Added "-genkey" option to "dsaparam" program.
18473 * Added RIPE MD160 (rmd160) message digest.
18477 * Added -a (all) option to "ssleay version" command.
18481 * Added PLATFORM define which is the id given to Configure.
18485 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18489 * Extended the ASN.1 parser routines.
18493 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18497 * Added a BN_CTX to the BN library.
18501 * Fixed the weak key values in DES library
18505 * Changed API in EVP library for cipher aliases.
18509 * Added support for RC2/64bit cipher.
18513 * Converted the lhash library to the crypto/mem.c functions.
18517 * Added more recognized ASN.1 object ids.
18521 * Added more RSA padding checks for SSL/TLS.
18525 * Added BIO proxy/filter functionality.
18529 * Added extra_certs to SSL_CTX which can be used
18530 send extra CA certificates to the client in the CA cert chain sending
18531 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18535 * Now Fortezza is denied in the authentication phase because
18536 this is key exchange mechanism is not supported by SSLeay at all.
18540 * Additional PKCS1 checks.
18544 * Support the string "TLSv1" for all TLS v1 ciphers.
18548 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18549 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18553 * Fixed a few memory leaks.
18557 * Fixed various code and comment typos.
18561 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18562 bytes sent in the client random.
18564 *Edward Bishop <ebishop@spyglass.com>*
18568 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18569 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18570 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18571 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18572 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18573 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18574 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18575 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18576 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18577 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18578 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18579 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18580 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18581 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18582 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18583 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18584 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18585 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18586 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18587 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18588 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18589 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18590 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18591 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18592 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18593 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18594 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18595 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18596 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18597 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18598 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18599 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18600 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18601 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18602 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18603 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18604 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18605 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18606 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18607 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18608 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18609 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18610 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18611 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18612 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18613 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18614 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18615 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18616 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18617 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18618 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18619 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18620 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18621 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18622 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18623 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18624 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18625 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18626 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18627 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18628 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18629 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18630 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18631 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18632 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18633 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18634 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18635 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18636 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18637 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18638 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18639 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18640 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18641 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18642 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18643 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18644 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18645 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18646 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18647 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18648 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18649 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18650 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18651 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18652 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18653 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18654 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18655 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18656 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18657 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18658 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18659 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18660 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18661 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18662 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18663 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18664 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18665 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18666 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18667 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18668 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18669 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18670 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18671 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18672 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18673 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18674 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18675 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18676 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18677 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18678 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18679 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18680 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18681 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18682 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18683 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18684 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18685 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18686 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18687 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18688 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18689 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18690 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18691 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18692 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18693 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18694 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18695 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18696 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18697 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18698 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18699 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18700 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18701 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18702 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18703 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18704 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18705 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18706 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18707 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18708 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18709 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18710 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18711 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18712 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18713 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18714 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18715 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18716 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18717 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18718 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18719 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18720 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18721 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18722 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18723 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18724 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18725 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18726 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18727 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18728 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655