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 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
28 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
29 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
30 The inverse ciphers use AES decryption for wrapping, and
31 AES encryption for unwrapping.
35 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
36 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
37 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
38 OpenSSL 3.0 these are replaced by the more generic functions
39 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
40 The old versions have been converted to deprecated macros that just call the
45 * The security callback, which can be customised by application code, supports
46 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
47 in the "other" parameter. In most places this is what is passed. All these
48 places occur server side. However there was one client side call of this
49 security operation and it passed a DH object instead. This is incorrect
50 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
51 of the other locations. Therefore this client side call has been changed to
52 pass an EVP_PKEY instead.
56 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
57 interface. Their functionality remains unchanged.
61 * Added new option for 'openssl list', '-providers', which will display the
62 list of loaded providers, their names, version and status. It optionally
63 displays their gettable parameters.
67 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
68 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
69 type is internally recognised so the workaround is no longer needed.
71 Functionality is still retained as it is, but will only work with
72 EVP_PKEYs with a legacy internal key.
76 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
77 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
81 * Changed all "STACK" functions to be macros instead of inline functions. Macro
82 parameters are still checked for type safety at compile time via helper
87 * Remove the RAND_DRBG API
89 The RAND_DRBG API did not fit well into the new provider concept as
90 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
91 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
92 and some of its API calls are rather low-level. This holds in particular
93 for the callback mechanism (RAND_DRBG_set_callbacks()).
95 Adding a compatibility layer to continue supporting the RAND_DRBG API as
96 a legacy API for a regular deprecation period turned out to come at the
97 price of complicating the new provider API unnecessarily. Since the
98 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
101 *Paul Dale and Matthias St. Pierre*
103 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
104 as well as actual hostnames.
108 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
109 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
110 conversely, silently ignore DTLS protocol version bounds when configuring
111 TLS-based contexts. The commands can be repeated to set bounds of both
112 types. The same applies with the corresponding "min_protocol" and
113 "max_protocol" command-line switches, in case some application uses both TLS
116 SSL_CTX instances that are created for a fixed protocol version (e.g.
117 TLSv1_server_method()) also silently ignore version bounds. Previously
118 attempts to apply bounds to these protocol versions would result in an
119 error. Now only the "version-flexible" SSL_CTX instances are subject to
120 limits in configuration files in command-line options.
124 * Deprecated the `ENGINE` API. Engines should be replaced with providers
129 * Reworked the recorded ERR codes to make better space for system errors.
130 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
131 given code is a system error (true) or an OpenSSL error (false).
135 * Reworked the test perl framework to better allow parallel testing.
137 *Nicola Tuveri and David von Oheimb*
139 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
140 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
144 * 'Configure' has been changed to figure out the configuration target if
145 none is given on the command line. Consequently, the 'config' script is
146 now only a mere wrapper. All documentation is changed to only mention
149 *Rich Salz and Richard Levitte*
151 * Added a library context `OSSL_LIB_CTX` that applications as well as
152 other libraries can use to form a separate context within which
153 libcrypto operations are performed.
155 There are two ways this can be used:
157 - Directly, by passing a library context to functions that take
158 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
160 - Indirectly, by creating a new library context and then assigning
161 it as the new default, with `OSSL_LIB_CTX_set0_default`.
163 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
164 apart from the functions directly related to `OSSL_LIB_CTX`, accept
165 NULL to indicate that the default library context should be used.
167 Library code that changes the default library context using
168 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
169 second call before returning to the caller.
171 _(Note: the library context was initially called `OPENSSL_CTX` and
172 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
176 * Handshake now fails if Extended Master Secret extension is dropped
181 * Dropped interactive mode from the 'openssl' program. From now on,
182 the `openssl` command without arguments is equivalent to `openssl
187 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
188 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
189 While the old function names have been retained for backward compatibility
190 they should not be used in new developments
191 because their return values are confusing: Unlike other `_cmp()` functions
192 they do not return 0 in case their arguments are equal.
196 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
197 EC_GROUP_get_field_type().
201 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
202 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
203 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
204 Applications should rely on the library automatically assigning a suitable
205 EC_METHOD internally upon EC_GROUP construction.
209 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
210 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
211 assigned internally without application intervention.
212 Users of EC_GROUP_new() should switch to a different suitable constructor.
216 * Add CAdES-BES signature verification support, mostly derived
217 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
219 *Filipe Raimundo da Silva*
221 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
225 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
226 functions are not widely used and now OpenSSL automatically perform this
227 conversion when needed.
231 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
232 EC_KEY_precompute_mult(). These functions are not widely used and
233 applications should instead switch to named curves which OpenSSL has
234 hardcoded lookup tables for.
238 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
239 should instead use the L<EC_POINT_mul(3)> function.
243 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
244 that are not applicable to the new provider model. Applications should
245 instead use EVP_default_properties_is_fips_enabled() and
246 EVP_default_properties_enable_fips().
250 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
251 is set, an unexpected EOF is ignored, it pretends a close notify was received
252 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
256 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
257 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
258 used and applications should instead use the
259 L<EC_POINT_set_affine_coordinates(3)> and
260 L<EC_POINT_get_affine_coordinates(3)> functions.
264 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
265 arrays to be more easily constructed via a series of utility functions.
266 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
267 the various push functions and finally convert to a passable OSSL_PARAM
268 array using OSSL_PARAM_BLD_to_param().
272 * The security strength of SHA1 and MD5 based signatures in TLS has been
273 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
274 working at the default security level of 1 and instead requires security
275 level 0. The security level can be changed either using the cipher string
276 with @SECLEVEL, or calling SSL_CTX_set_security_level().
280 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
281 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
282 internal keys, if they correspond to one of those built in types.
286 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
287 contain a provider side internal key.
291 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
292 They are old functions that we don't use, and that you could disable with
293 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
297 * Project text documents not yet having a proper file name extension
298 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
299 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
303 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
304 have been converted to Markdown with the goal to produce documents
305 which not only look pretty when viewed online in the browser, but
306 remain well readable inside a plain text editor.
308 To achieve this goal, a 'minimalistic' Markdown style has been applied
309 which avoids formatting elements that interfere too much with the
310 reading flow in the text file. For example, it
312 * avoids [ATX headings][] and uses [setext headings][] instead
313 (which works for `<h1>` and `<h2>` headings only).
314 * avoids [inline links][] and uses [reference links][] instead.
315 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
317 [ATX headings]: https://github.github.com/gfm/#atx-headings
318 [setext headings]: https://github.github.com/gfm/#setext-headings
319 [inline links]: https://github.github.com/gfm/#inline-link
320 [reference links]: https://github.github.com/gfm/#reference-link
321 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
322 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
324 *Matthias St. Pierre*
326 * The test suite is changed to preserve results of each test recipe.
327 A new directory test-runs/ with subdirectories named like the
328 test recipes are created in the build tree for this purpose.
332 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
333 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
334 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
336 *David von Oheimb, Martin Peylo*
338 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
339 The legacy OCSP-focused and only partly documented API is retained.
340 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
344 * Added `util/check-format.pl`, a tool for checking adherence to the
345 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
346 The checks performed are incomplete and yield some false positives.
347 Still the tool should be useful for detecting most typical glitches.
351 * BIO_do_connect and BIO_do_handshake have been extended:
352 If domain name resolution yields multiple IP addresses all of them are tried
353 after connect() failures.
357 * All of the low level RSA functions have been deprecated including:
359 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
360 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
361 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
362 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
363 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
364 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
365 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
366 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
367 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
368 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
369 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
370 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
371 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
372 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
373 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
374 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
375 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
376 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
377 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
378 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
379 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
380 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
381 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
382 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
383 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
384 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
385 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
386 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
388 Use of these low level functions has been informally discouraged for a long
389 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
390 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
391 L<EVP_PKEY_decrypt(3)>.
395 * X509 certificates signed using SHA1 are no longer allowed at security
397 In TLS/SSL the default security level is 1. It can be set either
398 using the cipher string with @SECLEVEL, or calling
399 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
400 a call to SSL_CTX_use_certificate() will fail if the security level is not
402 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
403 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
408 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
409 modified to use PKEY APIs. These commands are now in maintenance mode
410 and no new features will be added to them.
414 * The command line utility rsautl has been deprecated.
415 Instead use the pkeyutl program.
419 * The command line utilities genrsa and rsa have been modified to use PKEY
420 APIs. They now write PKCS#8 keys by default. These commands are now in
421 maintenance mode and no new features will be added to them.
425 * All of the low level DH functions have been deprecated including:
427 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
428 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
429 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
430 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
431 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
432 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
433 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
434 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
435 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
436 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
437 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
438 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
439 DH_meth_get_generate_params and DH_meth_set_generate_params.
441 Use of these low level functions has been informally discouraged for a long
442 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
443 and L<EVP_PKEY_derive(3)>.
447 * All of the low level DSA functions have been deprecated including:
449 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
450 DSA_get_default_method, DSA_set_method, DSA_get_method,
451 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
452 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
453 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
454 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
455 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
456 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
457 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
458 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
459 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
460 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
461 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
464 Use of these low level functions has been informally discouraged for a long
465 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
466 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
470 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
471 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
472 This means that applications don't have to look at the curve NID and
473 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
474 However, they still can, that EVP_PKEY_set_alias_type() call acts as
475 a no-op when the EVP_PKEY is already of the given type.
477 Parameter and key generation is also reworked to make it possible
478 to generate EVP_PKEY_SM2 parameters and keys without having to go
479 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
480 However, code that does the latter will still work as before.
484 * Deprecated low level ECDH and ECDSA functions. These include:
486 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
487 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
490 Use of these low level functions has been informally discouraged for a long
491 time. Instead applications should use the EVP_PKEY_derive(3),
492 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
496 * Deprecated the EC_KEY_METHOD functions. These include:
498 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
499 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
500 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
501 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
502 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
503 EC_KEY_METHOD_get_verify.
505 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
509 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
510 and EVP_PKEY_decrypt() instead.
511 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
512 and EVP_PKEY_encrypt() instead.
516 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
517 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
518 a new formulation to include all the things it can be used for,
519 as well as words of caution.
523 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
524 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
528 * All of the low level HMAC functions have been deprecated including:
530 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
531 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
534 Use of these low level functions has been informally discouraged for a long
535 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
536 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
537 and L<EVP_MAC_final(3)>.
541 * Over two thousand fixes were made to the documentation, including:
542 - Common options (such as -rand/-writerand, TLS version control, etc)
543 were refactored and point to newly-enhanced descriptions in openssl.pod.
544 - Added style conformance for all options (with help from Richard Levitte),
545 documented all reported missing options, added a CI build to check
546 that all options are documented and that no unimplemented options
548 - Documented some internals, such as all use of environment variables.
549 - Addressed all internal broken L<> references.
553 * All of the low level CMAC functions have been deprecated including:
555 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
556 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
558 Use of these low level functions has been informally discouraged for a long
559 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
560 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
561 and L<EVP_MAC_final(3)>.
565 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
566 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
569 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
570 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
571 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
572 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
573 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
574 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
575 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
576 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
577 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
578 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
580 Use of these low level functions has been informally discouraged
581 for a long time. Applications should use the EVP_DigestInit_ex(3),
582 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
586 * Corrected the documentation of the return values from the `EVP_DigestSign*`
587 set of functions. The documentation mentioned negative values for some
588 errors, but this was never the case, so the mention of negative values
591 Code that followed the documentation and thereby check with something
592 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
596 * All of the low level cipher functions have been deprecated including:
598 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
599 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
600 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
601 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
602 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
603 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
604 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
605 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
606 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
607 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
608 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
609 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
610 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
611 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
612 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
613 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
614 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
615 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
616 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
617 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
618 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
619 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
620 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
621 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
622 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
623 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
624 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
625 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
626 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
628 Use of these low level functions has been informally discouraged for
629 a long time. Applications should use the high level EVP APIs, e.g.
630 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
631 equivalently named decrypt functions instead.
633 *Matt Caswell and Paul Dale*
635 * Removed include/openssl/opensslconf.h.in and replaced it with
636 include/openssl/configuration.h.in, which differs in not including
637 <openssl/macros.h>. A short header include/openssl/opensslconf.h
638 was added to include both.
640 This allows internal hacks where one might need to modify the set
641 of configured macros, for example this if deprecated symbols are
642 still supposed to be available internally:
644 #include <openssl/configuration.h>
646 #undef OPENSSL_NO_DEPRECATED
647 #define OPENSSL_SUPPRESS_DEPRECATED
649 #include <openssl/macros.h>
651 This should not be used by applications that use the exported
652 symbols, as that will lead to linking errors.
656 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
657 used in exponentiation with 512-bit moduli. No EC algorithms are
658 affected. Analysis suggests that attacks against 2-prime RSA1024,
659 3-prime RSA1536, and DSA1024 as a result of this defect would be very
660 difficult to perform and are not believed likely. Attacks against DH512
661 are considered just feasible. However, for an attack the target would
662 have to re-use the DH512 private key, which is not recommended anyway.
663 Also applications directly using the low level API BN_mod_exp may be
664 affected if they use BN_FLG_CONSTTIME.
669 * Most memory-debug features have been deprecated, and the functionality
670 replaced with no-ops.
674 * Added documentation for the STACK API.
678 * Introduced a new method type and API, OSSL_ENCODER, to
679 represent generic encoders. An implementation is expected to
680 be able to encode an object associated with a given name (such
681 as an algorithm name for an asymmetric key) into forms given by
682 implementation properties.
684 Encoders are primarily used from inside libcrypto, through
685 calls to functions like EVP_PKEY_print_private(),
686 PEM_write_bio_PrivateKey() and similar.
688 Encoders are specified in such a way that they can be made to
689 directly handle the provider side portion of an object, if this
690 provider side part comes from the same provider as the encoder
691 itself, but can also be made to handle objects in parametrized
692 form (as an OSSL_PARAM array of data). This allows a provider to
693 offer generic encoders as a service for any other provider.
697 * Added a .pragma directive to the syntax of configuration files, to
698 allow varying behavior in a supported and predictable manner.
699 Currently added pragma:
703 This allows dollar signs to be a keyword character unless it's
704 followed by a opening brace or parenthesis. This is useful for
705 platforms where dollar signs are commonly used in names, such as
706 volume names and system directory names on VMS.
710 * Added functionality to create an EVP_PKEY from user data. This
711 is effectively the same as creating a RSA, DH or DSA object and
712 then assigning them to an EVP_PKEY, but directly using algorithm
713 agnostic EVP functions. A benefit is that this should be future
714 proof for public key algorithms to come.
718 * Change the interpretation of the '--api' configuration option to
719 mean that this is a desired API compatibility level with no
720 further meaning. The previous interpretation, that this would
721 also mean to remove all deprecated symbols up to and including
722 the given version, no requires that 'no-deprecated' is also used
723 in the configuration.
725 When building applications, the desired API compatibility level
726 can be set with the OPENSSL_API_COMPAT macro like before. For
727 API compatibility version below 3.0, the old style numerical
728 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
729 For version 3.0 and on, the value is expected to be the decimal
730 value calculated from the major and minor version like this:
732 MAJOR * 10000 + MINOR * 100
736 -DOPENSSL_API_COMPAT=30000 For 3.0
737 -DOPENSSL_API_COMPAT=30200 For 3.2
739 To hide declarations that are deprecated up to and including the
740 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
741 given when building the application as well.
745 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
746 access to certificate and CRL stores via URIs and OSSL_STORE
749 This adds the following functions:
751 - X509_LOOKUP_store()
752 - X509_STORE_load_file()
753 - X509_STORE_load_path()
754 - X509_STORE_load_store()
755 - SSL_add_store_cert_subjects_to_stack()
756 - SSL_CTX_set_default_verify_store()
757 - SSL_CTX_load_verify_file()
758 - SSL_CTX_load_verify_dir()
759 - SSL_CTX_load_verify_store()
763 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
764 The presence of this system service is determined at run-time.
768 * Added functionality to create an EVP_PKEY context based on data
769 for methods from providers. This takes an algorithm name and a
770 property query string and simply stores them, with the intent
771 that any operation that uses this context will use those strings
772 to fetch the needed methods implicitly, thereby making the port
773 of application written for pre-3.0 OpenSSL easier.
777 * The undocumented function NCONF_WIN32() has been deprecated; for
778 conversion details see the HISTORY section of doc/man5/config.pod
782 * Introduced the new functions EVP_DigestSignInit_ex() and
783 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
784 EVP_DigestVerifyUpdate() have been converted to functions. See the man
785 pages for further details.
789 * Over two thousand fixes were made to the documentation, including:
790 adding missing command flags, better style conformance, documentation
793 *Rich Salz, Richard Levitte*
795 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
796 X25519, X448, Ed25519 and Ed448.
800 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
805 * Deprecated the public definition of ERR_STATE as well as the function
806 ERR_get_state(). This is done in preparation of making ERR_STATE an
811 * Added ERR functionality to give callers access to the stored function
812 names that have replaced the older function code based functions.
814 New functions are ERR_get_error_func(), ERR_peek_error_func(),
815 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
816 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
817 and ERR_peek_last_error_all().
819 These functions have become deprecated: ERR_get_error_line_data(),
820 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
821 ERR_func_error_string().
825 * Extended testing to be verbose for failing tests only. The make variables
826 VERBOSE_FAILURE or VF can be used to enable this:
828 $ make VF=1 test # Unix
829 $ mms /macro=(VF=1) test ! OpenVMS
830 $ nmake VF=1 test # Windows
834 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
835 used even when parsing explicit parameters, when loading a encoded key
836 or calling `EC_GROUP_new_from_ecpkparameters()`/
837 `EC_GROUP_new_from_ecparameters()`.
838 This prevents bypass of security hardening and performance gains,
839 especially for curves with specialized EC_METHODs.
840 By default, if a key encoded with explicit parameters is loaded and later
841 encoded, the output is still encoded with explicit parameters, even if
842 internally a "named" EC_GROUP is used for computation.
846 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
847 this change, EC_GROUP_set_generator would accept order and/or cofactor as
848 NULL. After this change, only the cofactor parameter can be NULL. It also
849 does some minimal sanity checks on the passed order.
854 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
855 An attack is simple, if the first CMS_recipientInfo is valid but the
856 second CMS_recipientInfo is chosen ciphertext. If the second
857 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
858 encryption key will be replaced by garbage, and the message cannot be
859 decoded, but if the RSA decryption fails, the correct encryption key is
860 used and the recipient will not notice the attack.
861 As a work around for this potential attack the length of the decrypted
862 key must be equal to the cipher default key length, in case the
863 certifiate is not given and all recipientInfo are tried out.
864 The old behaviour can be re-enabled in the CMS code by setting the
865 CMS_DEBUG_DECRYPT flag.
869 * Early start up entropy quality from the DEVRANDOM seed source has been
870 improved for older Linux systems. The RAND subsystem will wait for
871 /dev/random to be producing output before seeding from /dev/urandom.
872 The seeded state is stored for future library initialisations using
873 a system global shared memory segment. The shared memory identifier
874 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
875 the desired value. The default identifier is 114.
879 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
880 when primes for RSA keys are computed.
881 Since we previously always generated primes == 2 (mod 3) for RSA keys,
882 the 2-prime and 3-prime RSA modules were easy to distinguish, since
883 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
884 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
885 This avoids possible fingerprinting of newly generated RSA modules.
889 * Correct the extended master secret constant on EBCDIC systems. Without this
890 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
891 negotiate EMS will fail. Unfortunately this also means that TLS connections
892 between EBCDIC systems with this fix, and EBCDIC systems without this
893 fix will fail if they negotiate EMS.
897 * Changed the library initialisation so that the config file is now loaded
898 by default. This was already the case for libssl. It now occurs for both
899 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
900 OPENSSL_init_crypto() to suppress automatic loading of a config file.
904 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
905 where the former acts as a replacement for ERR_put_error(), and the
906 latter replaces the combination ERR_put_error()+ERR_add_error_data().
907 ERR_raise_data() adds more flexibility by taking a format string and
908 an arbitrary number of arguments following it, to be processed with
913 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
914 to check if a named provider is loaded and available. When called, it
915 will also activate all fallback providers if such are still present.
919 * Enforce a minimum DH modulus size of 512 bits.
923 * Changed DH parameters to generate the order q subgroup instead of 2q.
924 Previously generated DH parameters are still accepted by DH_check
925 but DH_generate_key works around that by clearing bit 0 of the
926 private key for those. This avoids leaking bit 0 of the private key.
930 * Significantly reduce secure memory usage by the randomness pools.
934 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
939 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
940 algorithms. An implementation of a key exchange algorithm can be obtained
941 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
942 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
943 the older EVP_PKEY_derive_init() function. See the man pages for the new
944 functions for further details.
948 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
952 * Removed the function names from error messages and deprecated the
955 * Removed NextStep support and the macro OPENSSL_UNISTD
959 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
960 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
961 Also removed "export var as function" capability; we do not export
962 variables, only functions.
966 * RC5_32_set_key has been changed to return an int type, with 0 indicating
967 an error and 1 indicating success. In previous versions of OpenSSL this
968 was a void type. If a key was set longer than the maximum possible this
973 * Support SM2 signing and verification schemes with X509 certificate.
977 * Use SHA256 as the default digest for TS query in the ts app.
981 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
982 This checks that the salt length is at least 128 bits, the derived key
983 length is at least 112 bits, and that the iteration count is at least 1000.
984 For backwards compatibility these checks are disabled by default in the
985 default provider, but are enabled by default in the fips provider.
986 To enable or disable these checks use the control
987 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
991 * Default cipher lists/suites are now available via a function, the
992 #defines are deprecated.
996 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
997 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
998 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1002 * Join the directories crypto/x509 and crypto/x509v3
1006 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1007 This changes the size when using the genpkey app when no size is given. It
1008 fixes an omission in earlier changes that changed all RSA, DSA and DH
1009 generation apps to use 2048 bits by default.
1013 * Added command 'openssl kdf' that uses the EVP_KDF API.
1017 * Added command 'openssl mac' that uses the EVP_MAC API.
1021 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1022 as default directories. Also added the command 'openssl info'
1023 for scripting purposes.
1027 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1028 deprecated. These undocumented functions were never integrated into the EVP
1029 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1030 Bi-directional IGE mode. These modes were never formally standardised and
1031 usage of these functions is believed to be very small. In particular
1032 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1033 is ever used. The security implications are believed to be minimal, but
1034 this issue was never fixed for backwards compatibility reasons. New code
1035 should not use these modes.
1039 * Add prediction resistance to the DRBG reseeding process.
1043 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1044 mandated by IEEE Std 1619-2018.
1048 * Added newline escaping functionality to a filename when using openssl dgst.
1049 This output format is to replicate the output format found in the `*sum`
1050 checksum programs. This aims to preserve backward compatibility.
1052 *Matt Eaton, Richard Levitte, and Paul Dale*
1054 * Removed the heartbeat message in DTLS feature, as it has very
1055 little usage and doesn't seem to fulfill a valuable purpose.
1056 The configuration option is now deprecated.
1060 * Changed the output of 'openssl {digestname} < file' to display the
1061 digest name in its output.
1065 * Added a new generic trace API which provides support for enabling
1066 instrumentation through trace output. This feature is mainly intended
1067 as an aid for developers and is disabled by default. To utilize it,
1068 OpenSSL needs to be configured with the `enable-trace` option.
1070 If the tracing API is enabled, the application can activate trace output
1071 by registering BIOs as trace channels for a number of tracing and debugging
1074 The 'openssl' application has been expanded to enable any of the types
1075 available via environment variables defined by the user, and serves as
1076 one possible example on how to use this functionality.
1078 *Richard Levitte & Matthias St. Pierre*
1080 * Added build tests for C++. These are generated files that only do one
1081 thing, to include one public OpenSSL head file each. This tests that
1082 the public header files can be usefully included in a C++ application.
1084 This test isn't enabled by default. It can be enabled with the option
1085 'enable-buildtest-c++'.
1089 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1093 * Add KMAC to EVP_MAC.
1097 * Added property based algorithm implementation selection framework to
1102 * Added SCA hardening for modular field inversion in EC_GROUP through
1103 a new dedicated field_inv() pointer in EC_METHOD.
1104 This also addresses a leakage affecting conversions from projective
1105 to affine coordinates.
1107 *Billy Bob Brumley, Nicola Tuveri*
1109 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1110 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1111 those algorithms that were already supported through the EVP_PKEY API
1112 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1113 and scrypt are now wrappers that call EVP_KDF.
1117 * Build devcrypto engine as a dynamic engine.
1119 *Eneas U de Queiroz*
1121 * Add keyed BLAKE2 to EVP_MAC.
1125 * Fix a bug in the computation of the endpoint-pair shared secret used
1126 by DTLS over SCTP. This breaks interoperability with older versions
1127 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1128 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1129 interoperability with such broken implementations. However, enabling
1130 this switch breaks interoperability with correct implementations.
1132 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1133 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1137 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1141 * Change the license to the Apache License v2.0.
1145 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1147 - Major releases (indicated by incrementing the MAJOR release number)
1148 may introduce incompatible API/ABI changes.
1149 - Minor releases (indicated by incrementing the MINOR release number)
1150 may introduce new features but retain API/ABI compatibility.
1151 - Patch releases (indicated by incrementing the PATCH number)
1152 are intended for bug fixes and other improvements of existing
1153 features only (like improving performance or adding documentation)
1154 and retain API/ABI compatibility.
1158 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1162 * Remove the 'dist' target and add a tarball building script. The
1163 'dist' target has fallen out of use, and it shouldn't be
1164 necessary to configure just to create a source distribution.
1168 * Recreate the OS390-Unix config target. It no longer relies on a
1169 special script like it did for OpenSSL pre-1.1.0.
1173 * Instead of having the source directories listed in Configure, add
1174 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1179 * Add GMAC to EVP_MAC.
1183 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1187 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1188 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1189 to facilitate the continued use of MACs through raw private keys in
1190 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1194 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1195 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1199 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1200 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1201 are retained for backwards compatibility.
1205 * AES-XTS mode now enforces that its two keys are different to mitigate
1206 the attacked described in "Efficient Instantiations of Tweakable
1207 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1208 Details of this attack can be obtained from:
1209 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1213 * Rename the object files, i.e. give them other names than in previous
1214 versions. Their names now include the name of the final product, as
1215 well as its type mnemonic (bin, lib, shlib).
1219 * Added new option for 'openssl list', '-objects', which will display the
1220 list of built in objects, i.e. OIDs with names.
1224 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1225 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1230 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1231 improves application performance by removing data copies and providing
1232 applications with zero-copy system calls such as sendfile and splice.
1236 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1237 option is set, openssl cleanses (zeroize) plaintext bytes from
1238 internal buffers after delivering them to the application. Note,
1239 the application is still responsible for cleansing other copies
1240 (e.g.: data received by SSL_read(3)).
1244 * `PKCS12_parse` now maintains the order of the parsed certificates
1245 when outputting them via `*ca` (rather than reversing it).
1252 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1256 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1258 * Certificates with explicit curve parameters are now disallowed in
1259 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1263 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1264 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1265 conversely, silently ignore DTLS protocol version bounds when configuring
1266 TLS-based contexts. The commands can be repeated to set bounds of both
1267 types. The same applies with the corresponding "min_protocol" and
1268 "max_protocol" command-line switches, in case some application uses both TLS
1271 SSL_CTX instances that are created for a fixed protocol version (e.g.
1272 TLSv1_server_method()) also silently ignore version bounds. Previously
1273 attempts to apply bounds to these protocol versions would result in an
1274 error. Now only the "version-flexible" SSL_CTX instances are subject to
1275 limits in configuration files in command-line options.
1279 * Handshake now fails if Extended Master Secret extension is dropped
1284 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1286 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1288 * Fixed segmentation fault in SSL_check_chain()
1289 Server or client applications that call the SSL_check_chain() function
1290 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1291 dereference as a result of incorrect handling of the
1292 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1293 or unrecognised signature algorithm is received from the peer. This could
1294 be exploited by a malicious peer in a Denial of Service attack.
1299 * Added AES consttime code for no-asm configurations
1300 an optional constant time support for AES was added
1301 when building openssl for no-asm.
1302 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1303 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1304 At this time this feature is by default disabled.
1305 It will be enabled by default in 3.0.
1309 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1311 * Revert the change of EOF detection while reading in libssl to avoid
1312 regressions in applications depending on the current way of reporting
1313 the EOF. As the existing method is not fully accurate the change to
1314 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1315 branch and will be present in the 3.0 release.
1319 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1320 when primes for RSA keys are computed.
1321 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1322 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1323 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1324 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1325 This avoids possible fingerprinting of newly generated RSA modules.
1329 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1331 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1332 while reading in libssl then we would report an error back to the
1333 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1334 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1335 therefore give a hint as to what went wrong.
1339 * Check that ed25519 and ed448 are allowed by the security level. Previously
1340 signature algorithms not using an MD were not being checked that they were
1341 allowed by the security level.
1345 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1346 was not quite right. The behaviour was not consistent between resumption
1347 and normal handshakes, and also not quite consistent with historical
1348 behaviour. The behaviour in various scenarios has been clarified and
1349 it has been updated to make it match historical behaviour as closely as
1354 * *[VMS only]* The header files that the VMS compilers include automatically,
1355 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1356 that the C++ compiler doesn't understand. This is a shortcoming in the
1357 compiler, but can be worked around with `__cplusplus` guards.
1359 C++ applications that use OpenSSL libraries must be compiled using the
1360 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1361 functions. Otherwise, only functions with symbols of less than 31
1362 characters can be used, as the linker will not be able to successfully
1363 resolve symbols with longer names.
1367 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1368 The presence of this system service is determined at run-time.
1372 * Added newline escaping functionality to a filename when using openssl dgst.
1373 This output format is to replicate the output format found in the `*sum`
1374 checksum programs. This aims to preserve backward compatibility.
1376 *Matt Eaton, Richard Levitte, and Paul Dale*
1378 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1383 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1385 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1386 number generator (RNG). This was intended to include protection in the
1387 event of a fork() system call in order to ensure that the parent and child
1388 processes did not share the same RNG state. However this protection was not
1389 being used in the default case.
1391 A partial mitigation for this issue is that the output from a high
1392 precision timer is mixed into the RNG state so the likelihood of a parent
1393 and child process sharing state is significantly reduced.
1395 If an application already calls OPENSSL_init_crypto() explicitly using
1396 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1399 *Matthias St. Pierre*
1401 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1402 used even when parsing explicit parameters, when loading a encoded key
1403 or calling `EC_GROUP_new_from_ecpkparameters()`/
1404 `EC_GROUP_new_from_ecparameters()`.
1405 This prevents bypass of security hardening and performance gains,
1406 especially for curves with specialized EC_METHODs.
1407 By default, if a key encoded with explicit parameters is loaded and later
1408 encoded, the output is still encoded with explicit parameters, even if
1409 internally a "named" EC_GROUP is used for computation.
1413 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1414 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1415 NULL. After this change, only the cofactor parameter can be NULL. It also
1416 does some minimal sanity checks on the passed order.
1421 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1422 An attack is simple, if the first CMS_recipientInfo is valid but the
1423 second CMS_recipientInfo is chosen ciphertext. If the second
1424 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1425 encryption key will be replaced by garbage, and the message cannot be
1426 decoded, but if the RSA decryption fails, the correct encryption key is
1427 used and the recipient will not notice the attack.
1428 As a work around for this potential attack the length of the decrypted
1429 key must be equal to the cipher default key length, in case the
1430 certifiate is not given and all recipientInfo are tried out.
1431 The old behaviour can be re-enabled in the CMS code by setting the
1432 CMS_DEBUG_DECRYPT flag.
1437 * Early start up entropy quality from the DEVRANDOM seed source has been
1438 improved for older Linux systems. The RAND subsystem will wait for
1439 /dev/random to be producing output before seeding from /dev/urandom.
1440 The seeded state is stored for future library initialisations using
1441 a system global shared memory segment. The shared memory identifier
1442 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1443 the desired value. The default identifier is 114.
1447 * Correct the extended master secret constant on EBCDIC systems. Without this
1448 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1449 negotiate EMS will fail. Unfortunately this also means that TLS connections
1450 between EBCDIC systems with this fix, and EBCDIC systems without this
1451 fix will fail if they negotiate EMS.
1455 * Use Windows installation paths in the mingw builds
1457 Mingw isn't a POSIX environment per se, which means that Windows
1458 paths should be used for installation.
1463 * Changed DH_check to accept parameters with order q and 2q subgroups.
1464 With order 2q subgroups the bit 0 of the private key is not secret
1465 but DH_generate_key works around that by clearing bit 0 of the
1466 private key for those. This avoids leaking bit 0 of the private key.
1470 * Significantly reduce secure memory usage by the randomness pools.
1474 * Revert the DEVRANDOM_WAIT feature for Linux systems
1476 The DEVRANDOM_WAIT feature added a select() call to wait for the
1477 /dev/random device to become readable before reading from the
1478 /dev/urandom device.
1480 It turned out that this change had negative side effects on
1481 performance which were not acceptable. After some discussion it
1482 was decided to revert this feature and leave it up to the OS
1483 resp. the platform maintainer to ensure a proper initialization
1484 during early boot time.
1486 *Matthias St. Pierre*
1488 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1490 * Add build tests for C++. These are generated files that only do one
1491 thing, to include one public OpenSSL head file each. This tests that
1492 the public header files can be usefully included in a C++ application.
1494 This test isn't enabled by default. It can be enabled with the option
1495 'enable-buildtest-c++'.
1499 * Enable SHA3 pre-hashing for ECDSA and DSA.
1503 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1504 This changes the size when using the genpkey app when no size is given. It
1505 fixes an omission in earlier changes that changed all RSA, DSA and DH
1506 generation apps to use 2048 bits by default.
1510 * Reorganize the manual pages to consistently have RETURN VALUES,
1511 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1512 util/fix-doc-nits accordingly.
1514 *Paul Yang, Joshua Lock*
1516 * Add the missing accessor EVP_PKEY_get0_engine()
1520 * Have apps like 's_client' and 's_server' output the signature scheme
1521 along with other cipher suite parameters when debugging.
1525 * Make OPENSSL_config() error agnostic again.
1529 * Do the error handling in RSA decryption constant time.
1533 * Prevent over long nonces in ChaCha20-Poly1305.
1535 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1536 for every encryption operation. RFC 7539 specifies that the nonce value
1537 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1538 and front pads the nonce with 0 bytes if it is less than 12
1539 bytes. However it also incorrectly allows a nonce to be set of up to 16
1540 bytes. In this case only the last 12 bytes are significant and any
1541 additional leading bytes are ignored.
1543 It is a requirement of using this cipher that nonce values are
1544 unique. Messages encrypted using a reused nonce value are susceptible to
1545 serious confidentiality and integrity attacks. If an application changes
1546 the default nonce length to be longer than 12 bytes and then makes a
1547 change to the leading bytes of the nonce expecting the new value to be a
1548 new unique nonce then such an application could inadvertently encrypt
1549 messages with a reused nonce.
1551 Additionally the ignored bytes in a long nonce are not covered by the
1552 integrity guarantee of this cipher. Any application that relies on the
1553 integrity of these ignored leading bytes of a long nonce may be further
1554 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1555 is safe because no such use sets such a long nonce value. However user
1556 applications that use this cipher directly and set a non-default nonce
1557 length to be longer than 12 bytes may be vulnerable.
1559 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1565 * Add DEVRANDOM_WAIT feature for Linux systems
1567 On older Linux systems where the getrandom() system call is not available,
1568 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1569 Contrary to getrandom(), the /dev/urandom device will not block during
1570 early boot when the kernel CSPRNG has not been seeded yet.
1572 To mitigate this known weakness, use select() to wait for /dev/random to
1573 become readable before reading from /dev/urandom.
1575 * Ensure that SM2 only uses SM3 as digest algorithm
1579 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1581 * Change the info callback signals for the start and end of a post-handshake
1582 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1583 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1584 confused by this and assume that a TLSv1.2 renegotiation has started. This
1585 can break KeyUpdate handling. Instead we no longer signal the start and end
1586 of a post handshake message exchange (although the messages themselves are
1587 still signalled). This could break some applications that were expecting
1588 the old signals. However without this KeyUpdate is not usable for many
1593 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1595 * Timing vulnerability in DSA signature generation
1597 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1598 timing side channel attack. An attacker could use variations in the signing
1599 algorithm to recover the private key.
1601 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1606 * Timing vulnerability in ECDSA signature generation
1608 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1609 timing side channel attack. An attacker could use variations in the signing
1610 algorithm to recover the private key.
1612 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1617 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1618 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1619 of two gigabytes and the error handling improved.
1621 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1622 categorized as a normal bug, not a security issue, because the DRBG reseeds
1623 automatically and is fully functional even without additional randomness
1624 provided by the application.
1626 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1628 * Add a new ClientHello callback. Provides a callback interface that gives
1629 the application the ability to adjust the nascent SSL object at the
1630 earliest stage of ClientHello processing, immediately after extensions have
1631 been collected but before they have been processed. In particular, this
1632 callback can adjust the supported TLS versions in response to the contents
1637 * Add SM2 base algorithm support.
1641 * s390x assembly pack: add (improved) hardware-support for the following
1642 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1643 aes-cfb/cfb8, aes-ecb.
1647 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1648 parameter is no longer accepted, as it leads to a corrupt table. NULL
1649 pem_str is reserved for alias entries only.
1653 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1654 step for prime curves. The new implementation is based on formulae from
1655 differential addition-and-doubling in homogeneous projective coordinates
1656 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1657 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1658 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1659 to work in projective coordinates.
1661 *Billy Bob Brumley, Nicola Tuveri*
1663 * Change generating and checking of primes so that the error rate of not
1664 being prime depends on the intended use based on the size of the input.
1665 For larger primes this will result in more rounds of Miller-Rabin.
1666 The maximal error rate for primes with more than 1080 bits is lowered
1669 *Kurt Roeckx, Annie Yousar*
1671 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1675 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1676 moving between systems, and to avoid confusion when a Windows build is
1677 done with mingw vs with MSVC. For POSIX installs, there's still a
1678 symlink or copy named 'tsget' to avoid that confusion as well.
1682 * Revert blinding in ECDSA sign and instead make problematic addition
1683 length-invariant. Switch even to fixed-length Montgomery multiplication.
1687 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1688 step for binary curves. The new implementation is based on formulae from
1689 differential addition-and-doubling in mixed Lopez-Dahab projective
1690 coordinates, modified to independently blind the operands.
1692 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1694 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1695 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1696 EC_METHODs to implement their own specialized "ladder step", to take
1697 advantage of more favorable coordinate systems or more efficient
1698 differential addition-and-doubling algorithms.
1700 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1702 * Modified the random device based seed sources to keep the relevant
1703 file descriptors open rather than reopening them on each access.
1704 This allows such sources to operate in a chroot() jail without
1705 the associated device nodes being available. This behaviour can be
1706 controlled using RAND_keep_random_devices_open().
1710 * Numerous side-channel attack mitigations have been applied. This may have
1711 performance impacts for some algorithms for the benefit of improved
1712 security. Specific changes are noted in this change log by their respective
1717 * AIX shared library support overhaul. Switch to AIX "natural" way of
1718 handling shared libraries, which means collecting shared objects of
1719 different versions and bitnesses in one common archive. This allows to
1720 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1721 doesn't affect the way 3rd party applications are linked, only how
1722 multi-version installation is managed.
1726 * Make ec_group_do_inverse_ord() more robust and available to other
1727 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1728 mitigations are applied to the fallback BN_mod_inverse().
1729 When using this function rather than BN_mod_inverse() directly, new
1730 EC cryptosystem implementations are then safer-by-default.
1734 * Add coordinate blinding for EC_POINT and implement projective
1735 coordinate blinding for generic prime curves as a countermeasure to
1736 chosen point SCA attacks.
1738 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1740 * Add blinding to ECDSA and DSA signatures to protect against side channel
1741 attacks discovered by Keegan Ryan (NCC Group).
1745 * Enforce checking in the pkeyutl command line app to ensure that the input
1746 length does not exceed the maximum supported digest length when performing
1747 a sign, verify or verifyrecover operation.
1751 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1752 I/O in combination with something like select() or poll() will hang. This
1753 can be turned off again using SSL_CTX_clear_mode().
1754 Many applications do not properly handle non-application data records, and
1755 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1756 around the problems in those applications, but can also break some.
1757 It's recommended to read the manpages about SSL_read(), SSL_write(),
1758 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1759 SSL_CTX_set_read_ahead() again.
1763 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1764 now allow empty (zero character) pass phrases.
1768 * Apply blinding to binary field modular inversion and remove patent
1769 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1773 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1774 binary and prime elliptic curves.
1778 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1779 constant time fixed point multiplication.
1783 * Revise elliptic curve scalar multiplication with timing attack
1784 defenses: ec_wNAF_mul redirects to a constant time implementation
1785 when computing fixed point and variable point multiplication (which
1786 in OpenSSL are mostly used with secret scalars in keygen, sign,
1787 ECDH derive operations).
1788 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1791 * Updated CONTRIBUTING
1795 * Updated DRBG / RAND to request nonce and additional low entropy
1796 randomness from the system.
1798 *Matthias St. Pierre*
1800 * Updated 'openssl rehash' to use OpenSSL consistent default.
1804 * Moved the load of the ssl_conf module to libcrypto, which helps
1805 loading engines that libssl uses before libssl is initialised.
1809 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1813 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1815 *Ingo Schwarze, Rich Salz*
1817 * Added output of accepting IP address and port for 'openssl s_server'
1821 * Added a new API for TLSv1.3 ciphersuites:
1822 SSL_CTX_set_ciphersuites()
1823 SSL_set_ciphersuites()
1827 * Memory allocation failures consistently add an error to the error
1832 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1833 in libcrypto when run as setuid/setgid.
1837 * Load any config file by default when libssl is used.
1841 * Added new public header file <openssl/rand_drbg.h> and documentation
1842 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1844 *Matthias St. Pierre*
1846 * QNX support removed (cannot find contributors to get their approval
1847 for the license change).
1851 * TLSv1.3 replay protection for early data has been implemented. See the
1852 SSL_read_early_data() man page for further details.
1856 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1857 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1858 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1859 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1860 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1861 configuration has been separated out. See the ciphers man page or the
1862 SSL_CTX_set_ciphersuites() man page for more information.
1866 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1867 in responder mode now supports the new "-multi" option, which
1868 spawns the specified number of child processes to handle OCSP
1869 requests. The "-timeout" option now also limits the OCSP
1870 responder's patience to wait to receive the full client request
1871 on a newly accepted connection. Child processes are respawned
1872 as needed, and the CA index file is automatically reloaded
1873 when changed. This makes it possible to run the "ocsp" responder
1874 as a long-running service, making the OpenSSL CA somewhat more
1875 feature-complete. In this mode, most diagnostic messages logged
1876 after entering the event loop are logged via syslog(3) rather than
1881 * Added support for X448 and Ed448. Heavily based on original work by
1886 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1887 objects loaded. This adds the functions OSSL_STORE_expect() and
1888 OSSL_STORE_find() as well as needed tools to construct searches and
1889 get the search data out of them.
1893 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1894 version of OpenSSL should review their configuration settings to ensure
1895 that they are still appropriate for TLSv1.3. For further information see:
1896 <https://wiki.openssl.org/index.php/TLS1.3>
1900 * Grand redesign of the OpenSSL random generator
1902 The default RAND method now utilizes an AES-CTR DRBG according to
1903 NIST standard SP 800-90Ar1. The new random generator is essentially
1904 a port of the default random generator from the OpenSSL FIPS 2.0
1905 object module. It is a hybrid deterministic random bit generator
1906 using an AES-CTR bit stream and which seeds and reseeds itself
1907 automatically using trusted system entropy sources.
1909 Some of its new features are:
1910 - Support for multiple DRBG instances with seed chaining.
1911 - The default RAND method makes use of a DRBG.
1912 - There is a public and private DRBG instance.
1913 - The DRBG instances are fork-safe.
1914 - Keep all global DRBG instances on the secure heap if it is enabled.
1915 - The public and private DRBG instance are per thread for lock free
1918 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1920 * Changed Configure so it only says what it does and doesn't dump
1921 so much data. Instead, ./configdata.pm should be used as a script
1922 to display all sorts of configuration data.
1926 * Added processing of "make variables" to Configure.
1930 * Added SHA512/224 and SHA512/256 algorithm support.
1934 * The last traces of Netware support, first removed in 1.1.0, have
1939 * Get rid of Makefile.shared, and in the process, make the processing
1940 of certain files (rc.obj, or the .def/.map/.opt files produced from
1941 the ordinal files) more visible and hopefully easier to trace and
1942 debug (or make silent).
1946 * Make it possible to have environment variable assignments as
1947 arguments to config / Configure.
1951 * Add multi-prime RSA (RFC 8017) support.
1955 * Add SM3 implemented according to GB/T 32905-2016
1956 *Jack Lloyd <jack.lloyd@ribose.com>,*
1957 *Ronald Tse <ronald.tse@ribose.com>,*
1958 *Erick Borsboom <erick.borsboom@ribose.com>*
1960 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1961 as documented in RFC6066.
1962 Based on a patch from Tomasz Moń
1964 *Filipe Raimundo da Silva*
1966 * Add SM4 implemented according to GB/T 32907-2016.
1967 *Jack Lloyd <jack.lloyd@ribose.com>,*
1968 *Ronald Tse <ronald.tse@ribose.com>,*
1969 *Erick Borsboom <erick.borsboom@ribose.com>*
1971 * Reimplement -newreq-nodes and ERR_error_string_n; the
1972 original author does not agree with the license change.
1976 * Add ARIA AEAD TLS support.
1980 * Some macro definitions to support VS6 have been removed. Visual
1981 Studio 6 has not worked since 1.1.0
1985 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1986 without clearing the errors.
1990 * Add "atfork" functions. If building on a system that without
1991 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1992 requirements. The RAND facility now uses/requires this.
2000 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2001 not possible to disable entirely. However, it's still possible to
2002 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2005 To disable, configure with 'no-ui-console'. 'no-ui' is still
2006 possible to use as an alias. Check at compile time with the
2007 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2008 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2012 * Add a STORE module, which implements a uniform and URI based reader of
2013 stores that can contain keys, certificates, CRLs and numerous other
2014 objects. The main API is loosely based on a few stdio functions,
2015 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2016 OSSL_STORE_error and OSSL_STORE_close.
2017 The implementation uses backends called "loaders" to implement arbitrary
2018 URI schemes. There is one built in "loader" for the 'file' scheme.
2022 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2023 then adjusted to work on FreeBSD 8.4 as well.
2024 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2025 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2029 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2030 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2031 error code calls like this:
2033 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2035 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2036 that can be encoded in C. For the foreseeable future, this will only
2039 *Richard Levitte and Tim Hudson*
2041 * Removed BSD cryptodev engine.
2045 * Add a build target 'build_all_generated', to build all generated files
2046 and only that. This can be used to prepare everything that requires
2047 things like perl for a system that lacks perl and then move everything
2048 to that system and do the rest of the build there.
2052 * In the UI interface, make it possible to duplicate the user data. This
2053 can be used by engines that need to retain the data for a longer time
2054 than just the call where this user data is passed.
2058 * Ignore the '-named_curve auto' value for compatibility of applications
2061 *Tomas Mraz <tmraz@fedoraproject.org>*
2063 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2064 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2065 alerts across multiple records (some of which could be empty). In practice
2066 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2067 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2068 support this at all. Supporting it adds significant complexity to the
2069 record layer, and its removal is unlikely to cause interoperability
2074 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2075 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2076 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2081 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2082 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2084 *Richard Levitte, Andy Polyakov*
2086 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2091 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2092 platform rather than 'mingw'.
2096 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2097 success if they are asked to add an object which already exists
2098 in the store. This change cascades to other functions which load
2099 certificates and CRLs.
2103 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2104 facilitate stack unwinding even from assembly subroutines.
2108 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2109 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2113 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2114 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2115 which is the minimum version we support.
2119 * Certificate time validation (X509_cmp_time) enforces stricter
2120 compliance with RFC 5280. Fractional seconds and timezone offsets
2121 are no longer allowed.
2125 * Add support for ARIA
2129 * s_client will now send the Server Name Indication (SNI) extension by
2130 default unless the new "-noservername" option is used. The server name is
2131 based on the host provided to the "-connect" option unless overridden by
2132 using "-servername".
2136 * Add support for SipHash
2140 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2141 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2142 prevent issues where no progress is being made and the peer continually
2143 sends unrecognised record types, using up resources processing them.
2147 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2148 using the algorithm defined in
2149 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2153 * Heartbeat support has been removed; the ABI is changed for now.
2155 *Richard Levitte, Rich Salz*
2157 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2161 * The RSA "null" method, which was partially supported to avoid patent
2162 issues, has been replaced to always returns NULL.
2169 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2171 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2172 used even when parsing explicit parameters, when loading a encoded key
2173 or calling `EC_GROUP_new_from_ecpkparameters()`/
2174 `EC_GROUP_new_from_ecparameters()`.
2175 This prevents bypass of security hardening and performance gains,
2176 especially for curves with specialized EC_METHODs.
2177 By default, if a key encoded with explicit parameters is loaded and later
2178 encoded, the output is still encoded with explicit parameters, even if
2179 internally a "named" EC_GROUP is used for computation.
2183 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2184 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2185 NULL. After this change, only the cofactor parameter can be NULL. It also
2186 does some minimal sanity checks on the passed order.
2191 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2192 An attack is simple, if the first CMS_recipientInfo is valid but the
2193 second CMS_recipientInfo is chosen ciphertext. If the second
2194 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2195 encryption key will be replaced by garbage, and the message cannot be
2196 decoded, but if the RSA decryption fails, the correct encryption key is
2197 used and the recipient will not notice the attack.
2198 As a work around for this potential attack the length of the decrypted
2199 key must be equal to the cipher default key length, in case the
2200 certifiate is not given and all recipientInfo are tried out.
2201 The old behaviour can be re-enabled in the CMS code by setting the
2202 CMS_DEBUG_DECRYPT flag.
2207 * Use Windows installation paths in the mingw builds
2209 Mingw isn't a POSIX environment per se, which means that Windows
2210 paths should be used for installation.
2215 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2217 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2218 This changes the size when using the genpkey app when no size is given. It
2219 fixes an omission in earlier changes that changed all RSA, DSA and DH
2220 generation apps to use 2048 bits by default.
2224 * Prevent over long nonces in ChaCha20-Poly1305.
2226 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2227 for every encryption operation. RFC 7539 specifies that the nonce value
2228 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2229 and front pads the nonce with 0 bytes if it is less than 12
2230 bytes. However it also incorrectly allows a nonce to be set of up to 16
2231 bytes. In this case only the last 12 bytes are significant and any
2232 additional leading bytes are ignored.
2234 It is a requirement of using this cipher that nonce values are
2235 unique. Messages encrypted using a reused nonce value are susceptible to
2236 serious confidentiality and integrity attacks. If an application changes
2237 the default nonce length to be longer than 12 bytes and then makes a
2238 change to the leading bytes of the nonce expecting the new value to be a
2239 new unique nonce then such an application could inadvertently encrypt
2240 messages with a reused nonce.
2242 Additionally the ignored bytes in a long nonce are not covered by the
2243 integrity guarantee of this cipher. Any application that relies on the
2244 integrity of these ignored leading bytes of a long nonce may be further
2245 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2246 is safe because no such use sets such a long nonce value. However user
2247 applications that use this cipher directly and set a non-default nonce
2248 length to be longer than 12 bytes may be vulnerable.
2250 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2256 * Added SCA hardening for modular field inversion in EC_GROUP through
2257 a new dedicated field_inv() pointer in EC_METHOD.
2258 This also addresses a leakage affecting conversions from projective
2259 to affine coordinates.
2261 *Billy Bob Brumley, Nicola Tuveri*
2263 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2264 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2268 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2272 * Remove the 'dist' target and add a tarball building script. The
2273 'dist' target has fallen out of use, and it shouldn't be
2274 necessary to configure just to create a source distribution.
2278 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2280 * Timing vulnerability in DSA signature generation
2282 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2283 timing side channel attack. An attacker could use variations in the signing
2284 algorithm to recover the private key.
2286 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2291 * Timing vulnerability in ECDSA signature generation
2293 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2294 timing side channel attack. An attacker could use variations in the signing
2295 algorithm to recover the private key.
2297 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2302 * Add coordinate blinding for EC_POINT and implement projective
2303 coordinate blinding for generic prime curves as a countermeasure to
2304 chosen point SCA attacks.
2306 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2308 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2310 * Client DoS due to large DH parameter
2312 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2313 malicious server can send a very large prime value to the client. This will
2314 cause the client to spend an unreasonably long period of time generating a
2315 key for this prime resulting in a hang until the client has finished. This
2316 could be exploited in a Denial Of Service attack.
2318 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2323 * Cache timing vulnerability in RSA Key Generation
2325 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2326 a cache timing side channel attack. An attacker with sufficient access to
2327 mount cache timing attacks during the RSA key generation process could
2328 recover the private key.
2330 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2331 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2336 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2337 parameter is no longer accepted, as it leads to a corrupt table. NULL
2338 pem_str is reserved for alias entries only.
2342 * Revert blinding in ECDSA sign and instead make problematic addition
2343 length-invariant. Switch even to fixed-length Montgomery multiplication.
2347 * Change generating and checking of primes so that the error rate of not
2348 being prime depends on the intended use based on the size of the input.
2349 For larger primes this will result in more rounds of Miller-Rabin.
2350 The maximal error rate for primes with more than 1080 bits is lowered
2353 *Kurt Roeckx, Annie Yousar*
2355 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2359 * Add blinding to ECDSA and DSA signatures to protect against side channel
2360 attacks discovered by Keegan Ryan (NCC Group).
2364 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2365 now allow empty (zero character) pass phrases.
2369 * Certificate time validation (X509_cmp_time) enforces stricter
2370 compliance with RFC 5280. Fractional seconds and timezone offsets
2371 are no longer allowed.
2375 * Fixed a text canonicalisation bug in CMS
2377 Where a CMS detached signature is used with text content the text goes
2378 through a canonicalisation process first prior to signing or verifying a
2379 signature. This process strips trailing space at the end of lines, converts
2380 line terminators to CRLF and removes additional trailing line terminators
2381 at the end of a file. A bug in the canonicalisation process meant that
2382 some characters, such as form-feed, were incorrectly treated as whitespace
2383 and removed. This is contrary to the specification (RFC5485). This fix
2384 could mean that detached text data signed with an earlier version of
2385 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2386 signed with a fixed OpenSSL may fail to verify with an earlier version of
2387 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2388 and use the "-binary" flag (for the "cms" command line application) or set
2389 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2393 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2395 * Constructed ASN.1 types with a recursive definition could exceed the stack
2397 Constructed ASN.1 types with a recursive definition (such as can be found
2398 in PKCS7) could eventually exceed the stack given malicious input with
2399 excessive recursion. This could result in a Denial Of Service attack. There
2400 are no such structures used within SSL/TLS that come from untrusted sources
2401 so this is considered safe.
2403 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2409 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2411 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2412 effectively reduced to only comparing the least significant bit of each
2413 byte. This allows an attacker to forge messages that would be considered as
2414 authenticated in an amount of tries lower than that guaranteed by the
2415 security claims of the scheme. The module can only be compiled by the
2416 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2418 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2424 * Add a build target 'build_all_generated', to build all generated files
2425 and only that. This can be used to prepare everything that requires
2426 things like perl for a system that lacks perl and then move everything
2427 to that system and do the rest of the build there.
2431 * Backport SSL_OP_NO_RENGOTIATION
2433 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2434 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2435 changes this is no longer possible in 1.1.0. Therefore the new
2436 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2437 1.1.0 to provide equivalent functionality.
2439 Note that if an application built against 1.1.0h headers (or above) is run
2440 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2441 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2445 * Removed the OS390-Unix config target. It relied on a script that doesn't
2450 * rsaz_1024_mul_avx2 overflow bug on x86_64
2452 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2453 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2454 Analysis suggests that attacks against RSA and DSA as a result of this
2455 defect would be very difficult to perform and are not believed likely.
2456 Attacks against DH1024 are considered just feasible, because most of the
2457 work necessary to deduce information about a private key may be performed
2458 offline. The amount of resources required for such an attack would be
2459 significant. However, for an attack on TLS to be meaningful, the server
2460 would have to share the DH1024 private key among multiple clients, which is
2461 no longer an option since CVE-2016-0701.
2463 This only affects processors that support the AVX2 but not ADX extensions
2464 like Intel Haswell (4th generation).
2466 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2467 was originally found via the OSS-Fuzz project.
2472 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2474 * bn_sqrx8x_internal carry bug on x86_64
2476 There is a carry propagating bug in the x86_64 Montgomery squaring
2477 procedure. No EC algorithms are affected. Analysis suggests that attacks
2478 against RSA and DSA as a result of this defect would be very difficult to
2479 perform and are not believed likely. Attacks against DH are considered just
2480 feasible (although very difficult) because most of the work necessary to
2481 deduce information about a private key may be performed offline. The amount
2482 of resources required for such an attack would be very significant and
2483 likely only accessible to a limited number of attackers. An attacker would
2484 additionally need online access to an unpatched system using the target
2485 private key in a scenario with persistent DH parameters and a private
2486 key that is shared between multiple clients.
2488 This only affects processors that support the BMI1, BMI2 and ADX extensions
2489 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2491 This issue was reported to OpenSSL by the OSS-Fuzz project.
2496 * Malformed X.509 IPAddressFamily could cause OOB read
2498 If an X.509 certificate has a malformed IPAddressFamily extension,
2499 OpenSSL could do a one-byte buffer overread. The most likely result
2500 would be an erroneous display of the certificate in text format.
2502 This issue was reported to OpenSSL by the OSS-Fuzz project.
2507 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2509 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2510 platform rather than 'mingw'.
2514 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2515 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2516 which is the minimum version we support.
2520 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2522 * Encrypt-Then-Mac renegotiation crash
2524 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2525 negotiated where it was not in the original handshake (or vice-versa) then
2526 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2527 and servers are affected.
2529 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2534 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2536 * Truncated packet could crash via OOB read
2538 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2539 cipher is being used, then a truncated packet can cause that host to
2540 perform an out-of-bounds read, usually resulting in a crash.
2542 This issue was reported to OpenSSL by Robert Święcki of Google.
2547 * Bad (EC)DHE parameters cause a client crash
2549 If a malicious server supplies bad parameters for a DHE or ECDHE key
2550 exchange then this can result in the client attempting to dereference a
2551 NULL pointer leading to a client crash. This could be exploited in a Denial
2554 This issue was reported to OpenSSL by Guido Vranken.
2559 * BN_mod_exp may produce incorrect results on x86_64
2561 There is a carry propagating bug in the x86_64 Montgomery squaring
2562 procedure. No EC algorithms are affected. Analysis suggests that attacks
2563 against RSA and DSA as a result of this defect would be very difficult to
2564 perform and are not believed likely. Attacks against DH are considered just
2565 feasible (although very difficult) because most of the work necessary to
2566 deduce information about a private key may be performed offline. The amount
2567 of resources required for such an attack would be very significant and
2568 likely only accessible to a limited number of attackers. An attacker would
2569 additionally need online access to an unpatched system using the target
2570 private key in a scenario with persistent DH parameters and a private
2571 key that is shared between multiple clients. For example this can occur by
2572 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2573 similar to CVE-2015-3193 but must be treated as a separate problem.
2575 This issue was reported to OpenSSL by the OSS-Fuzz project.
2580 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2582 * ChaCha20/Poly1305 heap-buffer-overflow
2584 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2585 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2586 crash. This issue is not considered to be exploitable beyond a DoS.
2588 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2593 * CMS Null dereference
2595 Applications parsing invalid CMS structures can crash with a NULL pointer
2596 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2597 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2598 structure callback if an attempt is made to free certain invalid encodings.
2599 Only CHOICE structures using a callback which do not handle NULL value are
2602 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2607 * Montgomery multiplication may produce incorrect results
2609 There is a carry propagating bug in the Broadwell-specific Montgomery
2610 multiplication procedure that handles input lengths divisible by, but
2611 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2612 and DH private keys are impossible. This is because the subroutine in
2613 question is not used in operations with the private key itself and an input
2614 of the attacker's direct choice. Otherwise the bug can manifest itself as
2615 transient authentication and key negotiation failures or reproducible
2616 erroneous outcome of public-key operations with specially crafted input.
2617 Among EC algorithms only Brainpool P-512 curves are affected and one
2618 presumably can attack ECDH key negotiation. Impact was not analyzed in
2619 detail, because pre-requisites for attack are considered unlikely. Namely
2620 multiple clients have to choose the curve in question and the server has to
2621 share the private key among them, neither of which is default behaviour.
2622 Even then only clients that chose the curve will be affected.
2624 This issue was publicly reported as transient failures and was not
2625 initially recognized as a security issue. Thanks to Richard Morgan for
2626 providing reproducible case.
2631 * Removed automatic addition of RPATH in shared libraries and executables,
2632 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2636 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2638 * Fix Use After Free for large message sizes
2640 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2641 message larger than approx 16k is received then the underlying buffer to
2642 store the incoming message is reallocated and moved. Unfortunately a
2643 dangling pointer to the old location is left which results in an attempt to
2644 write to the previously freed location. This is likely to result in a
2645 crash, however it could potentially lead to execution of arbitrary code.
2647 This issue only affects OpenSSL 1.1.0a.
2649 This issue was reported to OpenSSL by Robert Święcki.
2654 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2656 * OCSP Status Request extension unbounded memory growth
2658 A malicious client can send an excessively large OCSP Status Request
2659 extension. If that client continually requests renegotiation, sending a
2660 large OCSP Status Request extension each time, then there will be unbounded
2661 memory growth on the server. This will eventually lead to a Denial Of
2662 Service attack through memory exhaustion. Servers with a default
2663 configuration are vulnerable even if they do not support OCSP. Builds using
2664 the "no-ocsp" build time option are not affected.
2666 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2671 * SSL_peek() hang on empty record
2673 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2674 sends an empty record. This could be exploited by a malicious peer in a
2675 Denial Of Service attack.
2677 This issue was reported to OpenSSL by Alex Gaynor.
2682 * Excessive allocation of memory in tls_get_message_header() and
2683 dtls1_preprocess_fragment()
2685 A (D)TLS message includes 3 bytes for its length in the header for the
2686 message. This would allow for messages up to 16Mb in length. Messages of
2687 this length are excessive and OpenSSL includes a check to ensure that a
2688 peer is sending reasonably sized messages in order to avoid too much memory
2689 being consumed to service a connection. A flaw in the logic of version
2690 1.1.0 means that memory for the message is allocated too early, prior to
2691 the excessive message length check. Due to way memory is allocated in
2692 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2693 to service a connection. This could lead to a Denial of Service through
2694 memory exhaustion. However, the excessive message length check still takes
2695 place, and this would cause the connection to immediately fail. Assuming
2696 that the application calls SSL_free() on the failed connection in a timely
2697 manner then the 21Mb of allocated memory will then be immediately freed
2698 again. Therefore the excessive memory allocation will be transitory in
2699 nature. This then means that there is only a security impact if:
2701 1) The application does not call SSL_free() in a timely manner in the event
2702 that the connection fails
2704 2) The application is working in a constrained environment where there is
2705 very little free memory
2707 3) The attacker initiates multiple connection attempts such that there are
2708 multiple connections in a state where memory has been allocated for the
2709 connection; SSL_free() has not yet been called; and there is insufficient
2710 memory to service the multiple requests.
2712 Except in the instance of (1) above any Denial Of Service is likely to be
2713 transitory because as soon as the connection fails the memory is
2714 subsequently freed again in the SSL_free() call. However there is an
2715 increased risk during this period of application crashes due to the lack of
2716 memory - which would then mean a more serious Denial of Service.
2718 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2719 (CVE-2016-6307 and CVE-2016-6308)
2723 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2724 had to be removed. Primary reason is that vendor assembler can't
2725 assemble our modules with -KPIC flag. As result it, assembly
2726 support, was not even available as option. But its lack means
2727 lack of side-channel resistant code, which is incompatible with
2728 security by todays standards. Fortunately gcc is readily available
2729 prepackaged option, which we firmly point at...
2733 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2735 * Windows command-line tool supports UTF-8 opt-in option for arguments
2736 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2737 (to any value) allows Windows user to access PKCS#12 file generated
2738 with Windows CryptoAPI and protected with non-ASCII password, as well
2739 as files generated under UTF-8 locale on Linux also protected with
2744 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2745 have been disabled by default and removed from DEFAULT, just like RC4.
2746 See the RC4 item below to re-enable both.
2750 * The method for finding the storage location for the Windows RAND seed file
2751 has changed. First we check %RANDFILE%. If that is not set then we check
2752 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2753 all else fails we fall back to C:\.
2757 * The EVP_EncryptUpdate() function has had its return type changed from void
2758 to int. A return of 0 indicates and error while a return of 1 indicates
2763 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2764 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2765 off the constant time implementation for RSA, DSA and DH have been made
2766 no-ops and deprecated.
2770 * Windows RAND implementation was simplified to only get entropy by
2771 calling CryptGenRandom(). Various other RAND-related tickets
2774 *Joseph Wylie Yandle, Rich Salz*
2776 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2777 and `OPENSSL_LH_`, respectively. The old names are available
2778 with API compatibility. They new names are now completely documented.
2782 * Unify TYPE_up_ref(obj) methods signature.
2783 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2784 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2785 int (instead of void) like all others TYPE_up_ref() methods.
2786 So now these methods also check the return value of CRYPTO_atomic_add(),
2787 and the validity of object reference counter.
2789 *fdasilvayy@gmail.com*
2791 * With Windows Visual Studio builds, the .pdb files are installed
2792 alongside the installed libraries and executables. For a static
2793 library installation, ossl_static.pdb is the associate compiler
2794 generated .pdb file to be used when linking programs.
2798 * Remove openssl.spec. Packaging files belong with the packagers.
2802 * Automatic Darwin/OSX configuration has had a refresh, it will now
2803 recognise x86_64 architectures automatically. You can still decide
2804 to build for a different bitness with the environment variable
2805 KERNEL_BITS (can be 32 or 64), for example:
2807 KERNEL_BITS=32 ./config
2811 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2812 256 bit AES and HMAC with SHA256.
2816 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2820 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2824 * To enable users to have their own config files and build file templates,
2825 Configure looks in the directory indicated by the environment variable
2826 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2827 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2828 name and is used as is.
2832 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2833 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2834 X509_CERT_FILE_CTX was removed.
2838 * "shared" builds are now the default. To create only static libraries use
2839 the "no-shared" Configure option.
2843 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2844 All of these option have not worked for some while and are fundamental
2849 * Make various cleanup routines no-ops and mark them as deprecated. Most
2850 global cleanup functions are no longer required because they are handled
2851 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2852 Explicitly de-initing can cause problems (e.g. where a library that uses
2853 OpenSSL de-inits, but an application is still using it). The affected
2854 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2855 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2856 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2857 COMP_zlib_cleanup().
2861 * --strict-warnings no longer enables runtime debugging options
2862 such as REF_DEBUG. Instead, debug options are automatically
2863 enabled with '--debug' builds.
2865 *Andy Polyakov, Emilia Käsper*
2867 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2868 have been moved out of the public header files. New functions for managing
2869 these have been added.
2873 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2874 objects have been moved out of the public header files. New
2875 functions for managing these have been added.
2879 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2880 have been moved out of the public header files. New functions for managing
2881 these have been added.
2885 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2886 moved out of the public header files. New functions for managing these
2891 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2895 * Removed the mk1mf build scripts.
2899 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2900 it is always safe to #include a header now.
2904 * Removed the aged BC-32 config and all its supporting scripts
2908 * Removed support for Ultrix, Netware, and OS/2.
2912 * Add support for HKDF.
2914 *Alessandro Ghedini*
2916 * Add support for blake2b and blake2s
2920 * Added support for "pipelining". Ciphers that have the
2921 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2922 encryptions/decryptions simultaneously. There are currently no built-in
2923 ciphers with this property but the expectation is that engines will be able
2924 to offer it to significantly improve throughput. Support has been extended
2925 into libssl so that multiple records for a single connection can be
2926 processed in one go (for >=TLS 1.1).
2930 * Added the AFALG engine. This is an async capable engine which is able to
2931 offload work to the Linux kernel. In this initial version it only supports
2932 AES128-CBC. The kernel must be version 4.1.0 or greater.
2936 * OpenSSL now uses a new threading API. It is no longer necessary to
2937 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2938 are two supported threading models: pthreads and windows threads. It is
2939 also possible to configure OpenSSL at compile time for "no-threads". The
2940 old threading API should no longer be used. The functions have been
2941 replaced with "no-op" compatibility macros.
2943 *Alessandro Ghedini, Matt Caswell*
2945 * Modify behavior of ALPN to invoke callback after SNI/servername
2946 callback, such that updates to the SSL_CTX affect ALPN.
2950 * Add SSL_CIPHER queries for authentication and key-exchange.
2954 * Changes to the DEFAULT cipherlist:
2955 - Prefer (EC)DHE handshakes over plain RSA.
2956 - Prefer AEAD ciphers over legacy ciphers.
2957 - Prefer ECDSA over RSA when both certificates are available.
2958 - Prefer TLSv1.2 ciphers/PRF.
2959 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2964 * Change the ECC default curve list to be this, in order: x25519,
2965 secp256r1, secp521r1, secp384r1.
2969 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2970 disabled by default. They can be re-enabled using the
2971 enable-weak-ssl-ciphers option to Configure.
2975 * If the server has ALPN configured, but supports no protocols that the
2976 client advertises, send a fatal "no_application_protocol" alert.
2977 This behaviour is SHALL in RFC 7301, though it isn't universally
2978 implemented by other servers.
2982 * Add X25519 support.
2983 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2984 for public and private key encoding using the format documented in
2985 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2986 key generation and key derivation.
2988 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2993 * Deprecate SRP_VBASE_get_by_user.
2994 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2995 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
2996 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2997 seed, even if the seed is configured.
2999 Users should use SRP_VBASE_get1_by_user instead. Note that in
3000 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3001 also that even though configuring the SRP seed attempts to hide
3002 invalid usernames by continuing the handshake with fake
3003 credentials, this behaviour is not constant time and no strong
3004 guarantees are made that the handshake is indistinguishable from
3005 that of a valid user.
3009 * Configuration change; it's now possible to build dynamic engines
3010 without having to build shared libraries and vice versa. This
3011 only applies to the engines in engines/, those in crypto/engine/
3012 will always be built into libcrypto (i.e. "static").
3014 Building dynamic engines is enabled by default; to disable, use
3015 the configuration option "disable-dynamic-engine".
3017 The only requirements for building dynamic engines are the
3018 presence of the DSO module and building with position independent
3019 code, so they will also automatically be disabled if configuring
3020 with "disable-dso" or "disable-pic".
3022 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3023 are also taken away from openssl/opensslconf.h, as they are
3028 * Configuration change; if there is a known flag to compile
3029 position independent code, it will always be applied on the
3030 libcrypto and libssl object files, and never on the application
3031 object files. This means other libraries that use routines from
3032 libcrypto / libssl can be made into shared libraries regardless
3033 of how OpenSSL was configured.
3035 If this isn't desirable, the configuration options "disable-pic"
3036 or "no-pic" can be used to disable the use of PIC. This will
3037 also disable building shared libraries and dynamic engines.
3041 * Removed JPAKE code. It was experimental and has no wide use.
3045 * The INSTALL_PREFIX Makefile variable has been renamed to
3046 DESTDIR. That makes for less confusion on what this variable
3047 is for. Also, the configuration option --install_prefix is
3052 * Heartbeat for TLS has been removed and is disabled by default
3053 for DTLS; configure with enable-heartbeats. Code that uses the
3054 old #define's might need to be updated.
3056 *Emilia Käsper, Rich Salz*
3058 * Rename REF_CHECK to REF_DEBUG.
3062 * New "unified" build system
3064 The "unified" build system is aimed to be a common system for all
3065 platforms we support. With it comes new support for VMS.
3067 This system builds supports building in a different directory tree
3068 than the source tree. It produces one Makefile (for unix family
3069 or lookalikes), or one descrip.mms (for VMS).
3071 The source of information to make the Makefile / descrip.mms is
3072 small files called 'build.info', holding the necessary
3073 information for each directory with source to compile, and a
3074 template in Configurations, like unix-Makefile.tmpl or
3077 With this change, the library names were also renamed on Windows
3078 and on VMS. They now have names that are closer to the standard
3079 on Unix, and include the major version number, and in certain
3080 cases, the architecture they are built for. See "Notes on shared
3081 libraries" in INSTALL.
3083 We rely heavily on the perl module Text::Template.
3087 * Added support for auto-initialisation and de-initialisation of the library.
3088 OpenSSL no longer requires explicit init or deinit routines to be called,
3089 except in certain circumstances. See the OPENSSL_init_crypto() and
3090 OPENSSL_init_ssl() man pages for further information.
3094 * The arguments to the DTLSv1_listen function have changed. Specifically the
3095 "peer" argument is now expected to be a BIO_ADDR object.
3097 * Rewrite of BIO networking library. The BIO library lacked consistent
3098 support of IPv6, and adding it required some more extensive
3099 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3100 which hold all types of addresses and chains of address information.
3101 It also introduces a new API, with functions like BIO_socket,
3102 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3103 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3104 have been adapted accordingly.
3108 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3113 * CRIME protection: disable compression by default, even if OpenSSL is
3114 compiled with zlib enabled. Applications can still enable compression
3115 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3116 using the SSL_CONF library to configure compression.
3120 * The signature of the session callback configured with
3121 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3122 was explicitly marked as `const unsigned char*` instead of
3127 * Always DPURIFY. Remove the use of uninitialized memory in the
3128 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3132 * Removed many obsolete configuration items, including
3133 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3134 MD2_CHAR, MD2_INT, MD2_LONG
3136 IDEA_SHORT, IDEA_LONG
3137 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3139 *Rich Salz, with advice from Andy Polyakov*
3141 * Many BN internals have been moved to an internal header file.
3143 *Rich Salz with help from Andy Polyakov*
3145 * Configuration and writing out the results from it has changed.
3146 Files such as Makefile include/openssl/opensslconf.h and are now
3147 produced through general templates, such as Makefile.in and
3148 crypto/opensslconf.h.in and some help from the perl module
3151 Also, the center of configuration information is no longer
3152 Makefile. Instead, Configure produces a perl module in
3153 configdata.pm which holds most of the config data (in the hash
3154 table %config), the target data that comes from the target
3155 configuration in one of the `Configurations/*.conf` files (in
3160 * To clarify their intended purposes, the Configure options
3161 --prefix and --openssldir change their semantics, and become more
3162 straightforward and less interdependent.
3164 --prefix shall be used exclusively to give the location INSTALLTOP
3165 where programs, scripts, libraries, include files and manuals are
3166 going to be installed. The default is now /usr/local.
3168 --openssldir shall be used exclusively to give the default
3169 location OPENSSLDIR where certificates, private keys, CRLs are
3170 managed. This is also where the default openssl.cnf gets
3172 If the directory given with this option is a relative path, the
3173 values of both the --prefix value and the --openssldir value will
3174 be combined to become OPENSSLDIR.
3175 The default for --openssldir is INSTALLTOP/ssl.
3177 Anyone who uses --openssldir to specify where OpenSSL is to be
3178 installed MUST change to use --prefix instead.
3182 * The GOST engine was out of date and therefore it has been removed. An up
3183 to date GOST engine is now being maintained in an external repository.
3184 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3185 support for GOST ciphersuites (these are only activated if a GOST engine
3190 * EGD is no longer supported by default; use enable-egd when
3193 *Ben Kaduk and Rich Salz*
3195 * The distribution now has Makefile.in files, which are used to
3196 create Makefile's when Configure is run. *Configure must be run
3197 before trying to build now.*
3201 * The return value for SSL_CIPHER_description() for error conditions
3206 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3208 Obtaining and performing DNSSEC validation of TLSA records is
3209 the application's responsibility. The application provides
3210 the TLSA records of its choice to OpenSSL, and these are then
3211 used to authenticate the peer.
3213 The TLSA records need not even come from DNS. They can, for
3214 example, be used to implement local end-entity certificate or
3215 trust-anchor "pinning", where the "pin" data takes the form
3216 of TLSA records, which can augment or replace verification
3217 based on the usual WebPKI public certification authorities.
3221 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3222 continues to support deprecated interfaces in default builds.
3223 However, applications are strongly advised to compile their
3224 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3225 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3226 or the 1.1.0 releases.
3228 In environments in which all applications have been ported to
3229 not use any deprecated interfaces OpenSSL's Configure script
3230 should be used with the --api=1.1.0 option to entirely remove
3231 support for the deprecated features from the library and
3232 unconditionally disable them in the installed headers.
3233 Essentially the same effect can be achieved with the "no-deprecated"
3234 argument to Configure, except that this will always restrict
3235 the build to just the latest API, rather than a fixed API
3238 As applications are ported to future revisions of the API,
3239 they should update their compile-time OPENSSL_API_COMPAT define
3240 accordingly, but in most cases should be able to continue to
3241 compile with later releases.
3243 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3244 0x10000000L and 0x00908000L, respectively. However those
3245 versions did not support the OPENSSL_API_COMPAT feature, and
3246 so applications are not typically tested for explicit support
3247 of just the undeprecated features of either release.
3251 * Add support for setting the minimum and maximum supported protocol.
3252 It can bet set via the SSL_set_min_proto_version() and
3253 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3254 MaxProtocol. It's recommended to use the new APIs to disable
3255 protocols instead of disabling individual protocols using
3256 SSL_set_options() or SSL_CONF's Protocol. This change also
3257 removes support for disabling TLS 1.2 in the OpenSSL TLS
3258 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3262 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3266 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3267 and integrates ECDSA and ECDH functionality into EC. Implementations can
3268 now redirect key generation and no longer need to convert to or from
3271 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3272 include the ec.h header file instead.
3276 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3277 ciphers who are no longer supported and drops support the ephemeral RSA key
3278 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3282 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3283 opaque. For HMAC_CTX, the following constructors and destructors
3286 HMAC_CTX *HMAC_CTX_new(void);
3287 void HMAC_CTX_free(HMAC_CTX *ctx);
3289 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3290 destroy such methods has been added. See EVP_MD_meth_new(3) and
3291 EVP_CIPHER_meth_new(3) for documentation.
3294 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3295 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3296 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3297 an already created structure.
3298 2) For consistency with the majority of our object creators and
3299 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3300 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3301 for deprecated builds.
3305 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3306 cryptographic operations to be performed asynchronously as long as an
3307 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3308 further details. Libssl has also had this capability integrated with the
3309 introduction of the new mode SSL_MODE_ASYNC and associated error
3310 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3311 pages. This work was developed in partnership with Intel Corp.
3315 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3316 always enabled now. If you want to disable the support you should
3317 exclude it using the list of supported ciphers. This also means that the
3318 "-no_ecdhe" option has been removed from s_server.
3322 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3323 SSL_{CTX_}set1_curves() which can set a list.
3327 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3328 curve you want to support using SSL_{CTX_}set1_curves().
3332 * State machine rewrite. The state machine code has been significantly
3333 refactored in order to remove much duplication of code and solve issues
3334 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3335 further details). This change does have some associated API changes.
3336 Notably the SSL_state() function has been removed and replaced by
3337 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3338 SSL_set_state() has been removed altogether. The previous handshake states
3339 defined in ssl.h and ssl3.h have also been removed.
3343 * All instances of the string "ssleay" in the public API were replaced
3344 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3345 Some error codes related to internal RSA_eay API's were renamed.
3349 * The demo files in crypto/threads were moved to demo/threads.
3353 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3356 *Matt Caswell, Rich Salz*
3358 * New ASN.1 embed macro.
3360 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3361 structure is not allocated: it is part of the parent. That is instead of
3369 This reduces memory fragmentation and make it impossible to accidentally
3370 set a mandatory field to NULL.
3372 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3373 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3374 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3379 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3383 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3384 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3385 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3386 DES and RC4 ciphersuites.
3390 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3391 This changes the decoding behaviour for some invalid messages,
3392 though the change is mostly in the more lenient direction, and
3393 legacy behaviour is preserved as much as possible.
3397 * Fix no-stdio build.
3398 *David Woodhouse <David.Woodhouse@intel.com> and also*
3399 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3401 * New testing framework
3402 The testing framework has been largely rewritten and is now using
3403 perl and the perl modules Test::Harness and an extended variant of
3404 Test::More called OpenSSL::Test to do its work. All test scripts in
3405 test/ have been rewritten into test recipes, and all direct calls to
3406 executables in test/Makefile have become individual recipes using the
3407 simplified testing OpenSSL::Test::Simple.
3409 For documentation on our testing modules, do:
3411 perldoc test/testlib/OpenSSL/Test/Simple.pm
3412 perldoc test/testlib/OpenSSL/Test.pm
3416 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3417 are used; the latter aborts on memory leaks (usually checked on exit).
3418 Some undocumented "set malloc, etc., hooks" functions were removed
3419 and others were changed. All are now documented.
3423 * In DSA_generate_parameters_ex, if the provided seed is too short,
3426 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3428 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3429 from RFC4279, RFC4785, RFC5487, RFC5489.
3431 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3432 original RSA_PSK patch.
3436 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3437 era flag was never set throughout the codebase (only read). Also removed
3438 SSL3_FLAGS_POP_BUFFER which was only used if
3439 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3443 * Changed the default name options in the "ca", "crl", "req" and "x509"
3444 to be "oneline" instead of "compat".
3448 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3449 not aware of clients that still exhibit this bug, and the workaround
3450 hasn't been working properly for a while.
3454 * The return type of BIO_number_read() and BIO_number_written() as well as
3455 the corresponding num_read and num_write members in the BIO structure has
3456 changed from unsigned long to uint64_t. On platforms where an unsigned
3457 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3462 * Given the pervasive nature of TLS extensions it is inadvisable to run
3463 OpenSSL without support for them. It also means that maintaining
3464 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3465 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3469 * Removed support for the two export grade static DH ciphersuites
3470 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3471 were newly added (along with a number of other static DH ciphersuites) to
3472 1.0.2. However the two export ones have *never* worked since they were
3473 introduced. It seems strange in any case to be adding new export
3474 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3478 * Version negotiation has been rewritten. In particular SSLv23_method(),
3479 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3480 and turned into macros which simply call the new preferred function names
3481 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3482 should use the new names instead. Also as part of this change the ssl23.h
3483 header file has been removed.
3487 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3488 code and the associated standard is no longer considered fit-for-purpose.
3492 * RT2547 was closed. When generating a private key, try to make the
3493 output file readable only by the owner. This behavior change might
3494 be noticeable when interacting with other software.
3496 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3501 * Added HTTP GET support to the ocsp command.
3505 * Changed default digest for the dgst and enc commands from MD5 to
3510 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3514 * Added support for TLS extended master secret from
3515 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3516 initial patch which was a great help during development.
3520 * All libssl internal structures have been removed from the public header
3521 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3522 now redundant). Users should not attempt to access internal structures
3523 directly. Instead they should use the provided API functions.
3527 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3528 Access to deprecated functions can be re-enabled by running config with
3529 "enable-deprecated". In addition applications wishing to use deprecated
3530 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3531 will, by default, disable some transitive includes that previously existed
3532 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3536 * Added support for OCB mode. OpenSSL has been granted a patent license
3537 compatible with the OpenSSL license for use of OCB. Details are available
3538 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3539 for OCB can be removed by calling config with no-ocb.
3543 * SSLv2 support has been removed. It still supports receiving a SSLv2
3544 compatible client hello.
3548 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3549 done while fixing the error code for the key-too-small case.
3551 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3553 * CA.sh has been removed; use CA.pl instead.
3557 * Removed old DES API.
3561 * Remove various unsupported platforms:
3567 Sinix/ReliantUNIX RM400
3572 16-bit platforms such as WIN16
3576 * Clean up OPENSSL_NO_xxx #define's
3577 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3578 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3579 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3580 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3581 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3582 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3583 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3584 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3585 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3586 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3590 * Cleaned up dead code
3591 Remove all but one '#ifdef undef' which is to be looked at.
3595 * Clean up calling of xxx_free routines.
3596 Just like free(), fix most of the xxx_free routines to accept
3597 NULL. Remove the non-null checks from callers. Save much code.
3601 * Add secure heap for storage of private keys (when possible).
3602 Add BIO_s_secmem(), CBIGNUM, etc.
3603 Contributed by Akamai Technologies under our Corporate CLA.
3607 * Experimental support for a new, fast, unbiased prime candidate generator,
3608 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3610 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3612 * New output format NSS in the sess_id command line tool. This allows
3613 exporting the session id and the master key in NSS keylog format.
3615 *Martin Kaiser <martin@kaiser.cx>*
3617 * Harmonize version and its documentation. -f flag is used to display
3620 *mancha <mancha1@zoho.com>*
3622 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3623 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3625 *mancha <mancha1@zoho.com>*
3627 * Fix some double frees. These are not thought to be exploitable.
3629 *mancha <mancha1@zoho.com>*
3631 * A missing bounds check in the handling of the TLS heartbeat extension
3632 can be used to reveal up to 64k of memory to a connected client or
3635 Thanks for Neel Mehta of Google Security for discovering this bug and to
3636 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3637 preparing the fix ([CVE-2014-0160])
3639 *Adam Langley, Bodo Moeller*
3641 * Fix for the attack described in the paper "Recovering OpenSSL
3642 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3643 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3644 <http://eprint.iacr.org/2014/140>
3646 Thanks to Yuval Yarom and Naomi Benger for discovering this
3647 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3649 *Yuval Yarom and Naomi Benger*
3651 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3652 this fixes a limitation in previous versions of OpenSSL.
3656 * Experimental encrypt-then-mac support.
3658 Experimental support for encrypt then mac from
3659 draft-gutmann-tls-encrypt-then-mac-02.txt
3661 To enable it set the appropriate extension number (0x42 for the test
3662 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3664 For non-compliant peers (i.e. just about everything) this should have no
3667 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3671 * Add EVP support for key wrapping algorithms, to avoid problems with
3672 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3673 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3674 algorithms and include tests cases.
3678 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3683 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3684 MGF1 digest and OAEP label.
3688 * Make openssl verify return errors.
3690 *Chris Palmer <palmer@google.com> and Ben Laurie*
3692 * New function ASN1_TIME_diff to calculate the difference between two
3693 ASN1_TIME structures or one structure and the current time.
3697 * Update fips_test_suite to support multiple command line options. New
3698 test to induce all self test errors in sequence and check expected
3703 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3704 sign or verify all in one operation.
3708 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3709 test programs and fips_test_suite. Includes functionality to parse
3710 the minimal script output of fipsalgest.pl directly.
3714 * Add authorisation parameter to FIPS_module_mode_set().
3718 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3722 * Use separate DRBG fields for internal and external flags. New function
3723 FIPS_drbg_health_check() to perform on demand health checking. Add
3724 generation tests to fips_test_suite with reduced health check interval to
3725 demonstrate periodic health checking. Add "nodh" option to
3726 fips_test_suite to skip very slow DH test.
3730 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3735 * More extensive health check for DRBG checking many more failure modes.
3736 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3737 combination: call this in fips_test_suite.
3741 * Add support for canonical generation of DSA parameter 'g'. See
3744 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3745 POST to handle HMAC cases.
3749 * Add functions FIPS_module_version() and FIPS_module_version_text()
3750 to return numerical and string versions of the FIPS module number.
3754 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3755 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3756 outside the validated module in the FIPS capable OpenSSL.
3760 * Minor change to DRBG entropy callback semantics. In some cases
3761 there is no multiple of the block length between min_len and
3762 max_len. Allow the callback to return more than max_len bytes
3763 of entropy but discard any extra: it is the callback's responsibility
3764 to ensure that the extra data discarded does not impact the
3765 requested amount of entropy.
3769 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3770 information in FIPS186-3, SP800-57 and SP800-131A.
3774 * CCM support via EVP. Interface is very similar to GCM case except we
3775 must supply all data in one chunk (i.e. no update, final) and the
3776 message length must be supplied if AAD is used. Add algorithm test
3781 * Initial version of POST overhaul. Add POST callback to allow the status
3782 of POST to be monitored and/or failures induced. Modify fips_test_suite
3783 to use callback. Always run all selftests even if one fails.
3787 * XTS support including algorithm test driver in the fips_gcmtest program.
3788 Note: this does increase the maximum key length from 32 to 64 bytes but
3789 there should be no binary compatibility issues as existing applications
3790 will never use XTS mode.
3794 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3795 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3796 performs algorithm blocking for unapproved PRNG types. Also do not
3797 set PRNG type in FIPS_mode_set(): leave this to the application.
3798 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3799 the standard OpenSSL PRNG: set additional data to a date time vector.
3803 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3804 This shouldn't present any incompatibility problems because applications
3805 shouldn't be using these directly and any that are will need to rethink
3806 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3810 * Extensive self tests and health checking required by SP800-90 DRBG.
3811 Remove strength parameter from FIPS_drbg_instantiate and always
3812 instantiate at maximum supported strength.
3816 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3820 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3824 * New function DH_compute_key_padded() to compute a DH key and pad with
3825 leading zeroes if needed: this complies with SP800-56A et al.
3829 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3830 anything, incomplete, subject to change and largely untested at present.
3834 * Modify fipscanisteronly build option to only build the necessary object
3835 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3839 * Add experimental option FIPSSYMS to give all symbols in
3840 fipscanister.o and FIPS or fips prefix. This will avoid
3841 conflicts with future versions of OpenSSL. Add perl script
3842 util/fipsas.pl to preprocess assembly language source files
3843 and rename any affected symbols.
3847 * Add selftest checks and algorithm block of non-fips algorithms in
3848 FIPS mode. Remove DES2 from selftests.
3852 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3853 return internal method without any ENGINE dependencies. Add new
3854 tiny fips sign and verify functions.
3858 * New build option no-ec2m to disable characteristic 2 code.
3862 * New build option "fipscanisteronly". This only builds fipscanister.o
3863 and (currently) associated fips utilities. Uses the file Makefile.fips
3864 instead of Makefile.org as the prototype.
3868 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3869 Update fips_gcmtest to use IV generator.
3873 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3874 setting output buffer to NULL. The `*Final` function must be
3875 called although it will not retrieve any additional data. The tag
3876 can be set or retrieved with a ctrl. The IV length is by default 12
3877 bytes (96 bits) but can be set to an alternative value. If the IV
3878 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3883 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3884 underlying do_cipher function handles all cipher semantics itself
3885 including padding and finalisation. This is useful if (for example)
3886 an ENGINE cipher handles block padding itself. The behaviour of
3887 do_cipher is subtly changed if this flag is set: the return value
3888 is the number of characters written to the output buffer (zero is
3889 no longer an error code) or a negative error code. Also if the
3890 input buffer is NULL and length 0 finalisation should be performed.
3894 * If a candidate issuer certificate is already part of the constructed
3895 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3899 * Improve forward-security support: add functions
3901 void SSL_CTX_set_not_resumable_session_callback(
3902 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3903 void SSL_set_not_resumable_session_callback(
3904 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3906 for use by SSL/TLS servers; the callback function will be called whenever a
3907 new session is created, and gets to decide whether the session may be
3908 cached to make it resumable (return 0) or not (return 1). (As by the
3909 SSL/TLS protocol specifications, the session_id sent by the server will be
3910 empty to indicate that the session is not resumable; also, the server will
3911 not generate RFC 4507 (RFC 5077) session tickets.)
3913 A simple reasonable callback implementation is to return is_forward_secure.
3914 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3915 by the SSL/TLS server library, indicating whether it can provide forward
3918 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3920 * New -verify_name option in command line utilities to set verification
3925 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3926 Add CMAC pkey methods.
3930 * Experimental renegotiation in s_server -www mode. If the client
3931 browses /reneg connection is renegotiated. If /renegcert it is
3932 renegotiated requesting a certificate.
3936 * Add an "external" session cache for debugging purposes to s_server. This
3937 should help trace issues which normally are only apparent in deployed
3938 multi-process servers.
3942 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3943 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3944 BIO_set_cipher() and some obscure PEM functions were changed so they
3945 can now return an error. The RAND changes required a change to the
3946 RAND_METHOD structure.
3950 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3951 a gcc attribute to warn if the result of a function is ignored. This
3952 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3953 whose return value is often ignored.
3957 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3958 These allow SCTs (signed certificate timestamps) to be requested and
3959 validated when establishing a connection.
3961 *Rob Percival <robpercival@google.com>*
3966 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3968 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3969 used even when parsing explicit parameters, when loading a encoded key
3970 or calling `EC_GROUP_new_from_ecpkparameters()`/
3971 `EC_GROUP_new_from_ecparameters()`.
3972 This prevents bypass of security hardening and performance gains,
3973 especially for curves with specialized EC_METHODs.
3974 By default, if a key encoded with explicit parameters is loaded and later
3975 encoded, the output is still encoded with explicit parameters, even if
3976 internally a "named" EC_GROUP is used for computation.
3980 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3981 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3982 NULL. After this change, only the cofactor parameter can be NULL. It also
3983 does some minimal sanity checks on the passed order.
3988 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3989 An attack is simple, if the first CMS_recipientInfo is valid but the
3990 second CMS_recipientInfo is chosen ciphertext. If the second
3991 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3992 encryption key will be replaced by garbage, and the message cannot be
3993 decoded, but if the RSA decryption fails, the correct encryption key is
3994 used and the recipient will not notice the attack.
3995 As a work around for this potential attack the length of the decrypted
3996 key must be equal to the cipher default key length, in case the
3997 certifiate is not given and all recipientInfo are tried out.
3998 The old behaviour can be re-enabled in the CMS code by setting the
3999 CMS_DEBUG_DECRYPT flag.
4004 * Document issue with installation paths in diverse Windows builds
4006 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4007 binaries and run-time config file.
4012 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4014 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4015 This changes the size when using the genpkey app when no size is given. It
4016 fixes an omission in earlier changes that changed all RSA, DSA and DH
4017 generation apps to use 2048 bits by default.
4021 * Add FIPS support for Android Arm 64-bit
4023 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4024 Module in Version 2.0.10. For some reason, the corresponding target
4025 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4026 built with FIPS support on Android Arm 64-bit. This omission has been
4029 *Matthias St. Pierre*
4031 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4033 * 0-byte record padding oracle
4035 If an application encounters a fatal protocol error and then calls
4036 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4037 then OpenSSL can respond differently to the calling application if a 0 byte
4038 record is received with invalid padding compared to if a 0 byte record is
4039 received with an invalid MAC. If the application then behaves differently
4040 based on that in a way that is detectable to the remote peer, then this
4041 amounts to a padding oracle that could be used to decrypt data.
4043 In order for this to be exploitable "non-stitched" ciphersuites must be in
4044 use. Stitched ciphersuites are optimised implementations of certain
4045 commonly used ciphersuites. Also the application must call SSL_shutdown()
4046 twice even if a protocol error has occurred (applications should not do
4047 this but some do anyway).
4049 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4050 Aviram, with additional investigation by Steven Collison and Andrew
4051 Hourselt. It was reported to OpenSSL on 10th December 2018.
4056 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4060 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4062 * Microarchitecture timing vulnerability in ECC scalar multiplication
4064 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4065 shown to be vulnerable to a microarchitecture timing side channel attack.
4066 An attacker with sufficient access to mount local timing attacks during
4067 ECDSA signature generation could recover the private key.
4069 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4070 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4076 * Timing vulnerability in DSA signature generation
4078 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4079 timing side channel attack. An attacker could use variations in the signing
4080 algorithm to recover the private key.
4082 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4087 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4088 Module, accidentally introduced while backporting security fixes from the
4089 development branch and hindering the use of ECC in FIPS mode.
4093 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4095 * Client DoS due to large DH parameter
4097 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4098 malicious server can send a very large prime value to the client. This will
4099 cause the client to spend an unreasonably long period of time generating a
4100 key for this prime resulting in a hang until the client has finished. This
4101 could be exploited in a Denial Of Service attack.
4103 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4108 * Cache timing vulnerability in RSA Key Generation
4110 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4111 a cache timing side channel attack. An attacker with sufficient access to
4112 mount cache timing attacks during the RSA key generation process could
4113 recover the private key.
4115 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4116 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4121 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4122 parameter is no longer accepted, as it leads to a corrupt table. NULL
4123 pem_str is reserved for alias entries only.
4127 * Revert blinding in ECDSA sign and instead make problematic addition
4128 length-invariant. Switch even to fixed-length Montgomery multiplication.
4132 * Change generating and checking of primes so that the error rate of not
4133 being prime depends on the intended use based on the size of the input.
4134 For larger primes this will result in more rounds of Miller-Rabin.
4135 The maximal error rate for primes with more than 1080 bits is lowered
4138 *Kurt Roeckx, Annie Yousar*
4140 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4144 * Add blinding to ECDSA and DSA signatures to protect against side channel
4145 attacks discovered by Keegan Ryan (NCC Group).
4149 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4150 now allow empty (zero character) pass phrases.
4154 * Certificate time validation (X509_cmp_time) enforces stricter
4155 compliance with RFC 5280. Fractional seconds and timezone offsets
4156 are no longer allowed.
4160 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4162 * Constructed ASN.1 types with a recursive definition could exceed the stack
4164 Constructed ASN.1 types with a recursive definition (such as can be found
4165 in PKCS7) could eventually exceed the stack given malicious input with
4166 excessive recursion. This could result in a Denial Of Service attack. There
4167 are no such structures used within SSL/TLS that come from untrusted sources
4168 so this is considered safe.
4170 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4176 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4178 * Read/write after SSL object in error state
4180 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4181 mechanism. The intent was that if a fatal error occurred during a handshake
4182 then OpenSSL would move into the error state and would immediately fail if
4183 you attempted to continue the handshake. This works as designed for the
4184 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4185 SSL_connect()), however due to a bug it does not work correctly if
4186 SSL_read() or SSL_write() is called directly. In that scenario, if the
4187 handshake fails then a fatal error will be returned in the initial function
4188 call. If SSL_read()/SSL_write() is subsequently called by the application
4189 for the same SSL object then it will succeed and the data is passed without
4190 being decrypted/encrypted directly from the SSL/TLS record layer.
4192 In order to exploit this issue an application bug would have to be present
4193 that resulted in a call to SSL_read()/SSL_write() being issued after having
4194 already received a fatal error.
4196 This issue was reported to OpenSSL by David Benjamin (Google).
4201 * rsaz_1024_mul_avx2 overflow bug on x86_64
4203 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4204 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4205 Analysis suggests that attacks against RSA and DSA as a result of this
4206 defect would be very difficult to perform and are not believed likely.
4207 Attacks against DH1024 are considered just feasible, because most of the
4208 work necessary to deduce information about a private key may be performed
4209 offline. The amount of resources required for such an attack would be
4210 significant. However, for an attack on TLS to be meaningful, the server
4211 would have to share the DH1024 private key among multiple clients, which is
4212 no longer an option since CVE-2016-0701.
4214 This only affects processors that support the AVX2 but not ADX extensions
4215 like Intel Haswell (4th generation).
4217 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4218 was originally found via the OSS-Fuzz project.
4223 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4225 * bn_sqrx8x_internal carry bug on x86_64
4227 There is a carry propagating bug in the x86_64 Montgomery squaring
4228 procedure. No EC algorithms are affected. Analysis suggests that attacks
4229 against RSA and DSA as a result of this defect would be very difficult to
4230 perform and are not believed likely. Attacks against DH are considered just
4231 feasible (although very difficult) because most of the work necessary to
4232 deduce information about a private key may be performed offline. The amount
4233 of resources required for such an attack would be very significant and
4234 likely only accessible to a limited number of attackers. An attacker would
4235 additionally need online access to an unpatched system using the target
4236 private key in a scenario with persistent DH parameters and a private
4237 key that is shared between multiple clients.
4239 This only affects processors that support the BMI1, BMI2 and ADX extensions
4240 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4242 This issue was reported to OpenSSL by the OSS-Fuzz project.
4247 * Malformed X.509 IPAddressFamily could cause OOB read
4249 If an X.509 certificate has a malformed IPAddressFamily extension,
4250 OpenSSL could do a one-byte buffer overread. The most likely result
4251 would be an erroneous display of the certificate in text format.
4253 This issue was reported to OpenSSL by the OSS-Fuzz project.
4258 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4260 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4261 platform rather than 'mingw'.
4265 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4267 * Truncated packet could crash via OOB read
4269 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4270 cipher is being used, then a truncated packet can cause that host to
4271 perform an out-of-bounds read, usually resulting in a crash.
4273 This issue was reported to OpenSSL by Robert Święcki of Google.
4278 * BN_mod_exp may produce incorrect results on x86_64
4280 There is a carry propagating bug in the x86_64 Montgomery squaring
4281 procedure. No EC algorithms are affected. Analysis suggests that attacks
4282 against RSA and DSA as a result of this defect would be very difficult to
4283 perform and are not believed likely. Attacks against DH are considered just
4284 feasible (although very difficult) because most of the work necessary to
4285 deduce information about a private key may be performed offline. The amount
4286 of resources required for such an attack would be very significant and
4287 likely only accessible to a limited number of attackers. An attacker would
4288 additionally need online access to an unpatched system using the target
4289 private key in a scenario with persistent DH parameters and a private
4290 key that is shared between multiple clients. For example this can occur by
4291 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4292 similar to CVE-2015-3193 but must be treated as a separate problem.
4294 This issue was reported to OpenSSL by the OSS-Fuzz project.
4299 * Montgomery multiplication may produce incorrect results
4301 There is a carry propagating bug in the Broadwell-specific Montgomery
4302 multiplication procedure that handles input lengths divisible by, but
4303 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4304 and DH private keys are impossible. This is because the subroutine in
4305 question is not used in operations with the private key itself and an input
4306 of the attacker's direct choice. Otherwise the bug can manifest itself as
4307 transient authentication and key negotiation failures or reproducible
4308 erroneous outcome of public-key operations with specially crafted input.
4309 Among EC algorithms only Brainpool P-512 curves are affected and one
4310 presumably can attack ECDH key negotiation. Impact was not analyzed in
4311 detail, because pre-requisites for attack are considered unlikely. Namely
4312 multiple clients have to choose the curve in question and the server has to
4313 share the private key among them, neither of which is default behaviour.
4314 Even then only clients that chose the curve will be affected.
4316 This issue was publicly reported as transient failures and was not
4317 initially recognized as a security issue. Thanks to Richard Morgan for
4318 providing reproducible case.
4323 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4324 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4325 prevent issues where no progress is being made and the peer continually
4326 sends unrecognised record types, using up resources processing them.
4330 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4332 * Missing CRL sanity check
4334 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4335 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4336 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4338 This issue only affects the OpenSSL 1.0.2i
4343 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4345 * OCSP Status Request extension unbounded memory growth
4347 A malicious client can send an excessively large OCSP Status Request
4348 extension. If that client continually requests renegotiation, sending a
4349 large OCSP Status Request extension each time, then there will be unbounded
4350 memory growth on the server. This will eventually lead to a Denial Of
4351 Service attack through memory exhaustion. Servers with a default
4352 configuration are vulnerable even if they do not support OCSP. Builds using
4353 the "no-ocsp" build time option are not affected.
4355 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4360 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4363 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4369 * OOB write in MDC2_Update()
4371 An overflow can occur in MDC2_Update() either if called directly or
4372 through the EVP_DigestUpdate() function using MDC2. If an attacker
4373 is able to supply very large amounts of input data after a previous
4374 call to EVP_EncryptUpdate() with a partial block then a length check
4375 can overflow resulting in a heap corruption.
4377 The amount of data needed is comparable to SIZE_MAX which is impractical
4380 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4385 * Malformed SHA512 ticket DoS
4387 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4388 DoS attack where a malformed ticket will result in an OOB read which will
4391 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4392 a custom server callback and ticket lookup mechanism.
4394 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4399 * OOB write in BN_bn2dec()
4401 The function BN_bn2dec() does not check the return value of BN_div_word().
4402 This can cause an OOB write if an application uses this function with an
4403 overly large BIGNUM. This could be a problem if an overly large certificate
4404 or CRL is printed out from an untrusted source. TLS is not affected because
4405 record limits will reject an oversized certificate before it is parsed.
4407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4412 * OOB read in TS_OBJ_print_bio()
4414 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4415 the total length the OID text representation would use and not the amount
4416 of data written. This will result in OOB reads when large OIDs are
4419 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4424 * Pointer arithmetic undefined behaviour
4426 Avoid some undefined pointer arithmetic
4428 A common idiom in the codebase is to check limits in the following manner:
4431 Where "p" points to some malloc'd data of SIZE bytes and
4434 "len" here could be from some externally supplied data (e.g. from a TLS
4437 The rules of C pointer arithmetic are such that "p + len" is only well
4438 defined where len <= SIZE. Therefore the above idiom is actually
4439 undefined behaviour.
4441 For example this could cause problems if some malloc implementation
4442 provides an address for "p" such that "p + len" actually overflows for
4443 values of len that are too big and therefore p + len < limit.
4445 This issue was reported to OpenSSL by Guido Vranken
4450 * Constant time flag not preserved in DSA signing
4452 Operations in the DSA signing algorithm should run in constant time in
4453 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4454 implementation means that a non-constant time codepath is followed for
4455 certain operations. This has been demonstrated through a cache-timing
4456 attack to be sufficient for an attacker to recover the private DSA key.
4458 This issue was reported by César Pereida (Aalto University), Billy Brumley
4459 (Tampere University of Technology), and Yuval Yarom (The University of
4460 Adelaide and NICTA).
4465 * DTLS buffered message DoS
4467 In a DTLS connection where handshake messages are delivered out-of-order
4468 those messages that OpenSSL is not yet ready to process will be buffered
4469 for later use. Under certain circumstances, a flaw in the logic means that
4470 those messages do not get removed from the buffer even though the handshake
4471 has been completed. An attacker could force up to approx. 15 messages to
4472 remain in the buffer when they are no longer required. These messages will
4473 be cleared when the DTLS connection is closed. The default maximum size for
4474 a message is 100k. Therefore the attacker could force an additional 1500k
4475 to be consumed per connection. By opening many simulataneous connections an
4476 attacker could cause a DoS attack through memory exhaustion.
4478 This issue was reported to OpenSSL by Quan Luo.
4483 * DTLS replay protection DoS
4485 A flaw in the DTLS replay attack protection mechanism means that records
4486 that arrive for future epochs update the replay protection "window" before
4487 the MAC for the record has been validated. This could be exploited by an
4488 attacker by sending a record for the next epoch (which does not have to
4489 decrypt or have a valid MAC), with a very large sequence number. This means
4490 that all subsequent legitimate packets are dropped causing a denial of
4491 service for a specific DTLS connection.
4493 This issue was reported to OpenSSL by the OCAP audit team.
4498 * Certificate message OOB reads
4500 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4501 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4502 theoretical DoS risk but this has not been observed in practice on common
4505 The messages affected are client certificate, client certificate request
4506 and server certificate. As a result the attack can only be performed
4507 against a client or a server which enables client authentication.
4509 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4514 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4516 * Prevent padding oracle in AES-NI CBC MAC check
4518 A MITM attacker can use a padding oracle attack to decrypt traffic
4519 when the connection uses an AES CBC cipher and the server support
4522 This issue was introduced as part of the fix for Lucky 13 padding
4523 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4524 constant time by making sure that always the same bytes are read and
4525 compared against either the MAC or padding bytes. But it no longer
4526 checked that there was enough data to have both the MAC and padding
4529 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4534 * Fix EVP_EncodeUpdate overflow
4536 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4537 Base64 encoding of binary data. If an attacker is able to supply very large
4538 amounts of input data then a length check can overflow resulting in a heap
4541 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4542 the `PEM_write_bio*` family of functions. These are mainly used within the
4543 OpenSSL command line applications, so any application which processes data
4544 from an untrusted source and outputs it as a PEM file should be considered
4545 vulnerable to this issue. User applications that call these APIs directly
4546 with large amounts of untrusted data may also be vulnerable.
4548 This issue was reported by Guido Vranken.
4553 * Fix EVP_EncryptUpdate overflow
4555 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4556 is able to supply very large amounts of input data after a previous call to
4557 EVP_EncryptUpdate() with a partial block then a length check can overflow
4558 resulting in a heap corruption. Following an analysis of all OpenSSL
4559 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4560 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4561 the first called function after an EVP_EncryptInit(), and therefore that
4562 specific call must be safe. The second form is where the length passed to
4563 EVP_EncryptUpdate() can be seen from the code to be some small value and
4564 therefore there is no possibility of an overflow. Since all instances are
4565 one of these two forms, it is believed that there can be no overflows in
4566 internal code due to this problem. It should be noted that
4567 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4568 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4569 of these calls have also been analysed too and it is believed there are no
4570 instances in internal usage where an overflow could occur.
4572 This issue was reported by Guido Vranken.
4577 * Prevent ASN.1 BIO excessive memory allocation
4579 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4580 a short invalid encoding can cause allocation of large amounts of memory
4581 potentially consuming excessive resources or exhausting memory.
4583 Any application parsing untrusted data through d2i BIO functions is
4584 affected. The memory based functions such as d2i_X509() are *not* affected.
4585 Since the memory based functions are used by the TLS library, TLS
4586 applications are not affected.
4588 This issue was reported by Brian Carpenter.
4595 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4596 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4597 in arbitrary stack data being returned in the buffer.
4599 This issue was reported by Guido Vranken.
4604 * Modify behavior of ALPN to invoke callback after SNI/servername
4605 callback, such that updates to the SSL_CTX affect ALPN.
4609 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4614 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4615 methods are enabled and ssl2 is disabled the methods return NULL.
4619 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4621 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4622 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4623 provide any "EXPORT" or "LOW" strength ciphers.
4627 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4628 is by default disabled at build-time. Builds that are not configured with
4629 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4630 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4631 will need to explicitly call either of:
4633 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4635 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4637 as appropriate. Even if either of those is used, or the application
4638 explicitly uses the version-specific SSLv2_method() or its client and
4639 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4640 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4641 ciphers, and SSLv2 56-bit DES are no longer available.
4646 * Fix a double-free in DSA code
4648 A double free bug was discovered when OpenSSL parses malformed DSA private
4649 keys and could lead to a DoS attack or memory corruption for applications
4650 that receive DSA private keys from untrusted sources. This scenario is
4653 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4659 * Disable SRP fake user seed to address a server memory leak.
4661 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4663 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4664 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4665 was changed to ignore the "fake user" SRP seed, even if the seed
4668 Users should use SRP_VBASE_get1_by_user instead. Note that in
4669 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4670 also that even though configuring the SRP seed attempts to hide
4671 invalid usernames by continuing the handshake with fake
4672 credentials, this behaviour is not constant time and no strong
4673 guarantees are made that the handshake is indistinguishable from
4674 that of a valid user.
4679 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4681 In the BN_hex2bn function the number of hex digits is calculated using an
4682 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4683 large values of `i` this can result in `bn_expand` not allocating any
4684 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4685 field as NULL leading to a subsequent NULL ptr deref. For very large values
4686 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4687 In this case memory is allocated to the internal BIGNUM data field, but it
4688 is insufficiently sized leading to heap corruption. A similar issue exists
4689 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4690 is ever called by user applications with very large untrusted hex/dec data.
4691 This is anticipated to be a rare occurrence.
4693 All OpenSSL internal usage of these functions use data that is not expected
4694 to be untrusted, e.g. config file data or application command line
4695 arguments. If user developed applications generate config file data based
4696 on untrusted data then it is possible that this could also lead to security
4697 consequences. This is also anticipated to be rare.
4699 This issue was reported to OpenSSL by Guido Vranken.
4704 * Fix memory issues in `BIO_*printf` functions
4706 The internal `fmtstr` function used in processing a "%s" format string in
4707 the `BIO_*printf` functions could overflow while calculating the length of a
4708 string and cause an OOB read when printing very long strings.
4710 Additionally the internal `doapr_outch` function can attempt to write to an
4711 OOB memory location (at an offset from the NULL pointer) in the event of a
4712 memory allocation failure. In 1.0.2 and below this could be caused where
4713 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4714 could be in processing a very long "%s" format string. Memory leaks can
4717 The first issue may mask the second issue dependent on compiler behaviour.
4718 These problems could enable attacks where large amounts of untrusted data
4719 is passed to the `BIO_*printf` functions. If applications use these functions
4720 in this way then they could be vulnerable. OpenSSL itself uses these
4721 functions when printing out human-readable dumps of ASN.1 data. Therefore
4722 applications that print this data could be vulnerable if the data is from
4723 untrusted sources. OpenSSL command line applications could also be
4724 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4725 as command line arguments.
4727 Libssl is not considered directly vulnerable. Additionally certificates etc
4728 received via remote connections via libssl are also unlikely to be able to
4729 trigger these issues because of message size limits enforced within libssl.
4731 This issue was reported to OpenSSL Guido Vranken.
4736 * Side channel attack on modular exponentiation
4738 A side-channel attack was found which makes use of cache-bank conflicts on
4739 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4740 of RSA keys. The ability to exploit this issue is limited as it relies on
4741 an attacker who has control of code in a thread running on the same
4742 hyper-threaded core as the victim thread which is performing decryptions.
4744 This issue was reported to OpenSSL by Yuval Yarom, The University of
4745 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4746 Nadia Heninger, University of Pennsylvania with more information at
4747 <http://cachebleed.info>.
4752 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4753 if no keysize is specified with default_bits. This fixes an
4754 omission in an earlier change that changed all RSA/DSA key generation
4755 apps to use 2048 bits by default.
4759 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4761 * DH small subgroups
4763 Historically OpenSSL only ever generated DH parameters based on "safe"
4764 primes. More recently (in version 1.0.2) support was provided for
4765 generating X9.42 style parameter files such as those required for RFC 5114
4766 support. The primes used in such files may not be "safe". Where an
4767 application is using DH configured with parameters based on primes that are
4768 not "safe" then an attacker could use this fact to find a peer's private
4769 DH exponent. This attack requires that the attacker complete multiple
4770 handshakes in which the peer uses the same private DH exponent. For example
4771 this could be used to discover a TLS server's private DH exponent if it's
4772 reusing the private DH exponent or it's using a static DH ciphersuite.
4774 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4775 TLS. It is not on by default. If the option is not set then the server
4776 reuses the same private DH exponent for the life of the server process and
4777 would be vulnerable to this attack. It is believed that many popular
4778 applications do set this option and would therefore not be at risk.
4780 The fix for this issue adds an additional check where a "q" parameter is
4781 available (as is the case in X9.42 based parameters). This detects the
4782 only known attack, and is the only possible defense for static DH
4783 ciphersuites. This could have some performance impact.
4785 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4786 default and cannot be disabled. This could have some performance impact.
4788 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4793 * SSLv2 doesn't block disabled ciphers
4795 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4796 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4797 been disabled, provided that the SSLv2 protocol was not also disabled via
4800 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4801 and Sebastian Schinzel.
4806 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4808 * BN_mod_exp may produce incorrect results on x86_64
4810 There is a carry propagating bug in the x86_64 Montgomery squaring
4811 procedure. No EC algorithms are affected. Analysis suggests that attacks
4812 against RSA and DSA as a result of this defect would be very difficult to
4813 perform and are not believed likely. Attacks against DH are considered just
4814 feasible (although very difficult) because most of the work necessary to
4815 deduce information about a private key may be performed offline. The amount
4816 of resources required for such an attack would be very significant and
4817 likely only accessible to a limited number of attackers. An attacker would
4818 additionally need online access to an unpatched system using the target
4819 private key in a scenario with persistent DH parameters and a private
4820 key that is shared between multiple clients. For example this can occur by
4821 default in OpenSSL DHE based SSL/TLS ciphersuites.
4823 This issue was reported to OpenSSL by Hanno Böck.
4828 * Certificate verify crash with missing PSS parameter
4830 The signature verification routines will crash with a NULL pointer
4831 dereference if presented with an ASN.1 signature using the RSA PSS
4832 algorithm and absent mask generation function parameter. Since these
4833 routines are used to verify certificate signature algorithms this can be
4834 used to crash any certificate verification operation and exploited in a
4835 DoS attack. Any application which performs certificate verification is
4836 vulnerable including OpenSSL clients and servers which enable client
4839 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4844 * X509_ATTRIBUTE memory leak
4846 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4847 memory. This structure is used by the PKCS#7 and CMS routines so any
4848 application which reads PKCS#7 or CMS data from untrusted sources is
4849 affected. SSL/TLS is not affected.
4851 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4857 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4858 This changes the decoding behaviour for some invalid messages,
4859 though the change is mostly in the more lenient direction, and
4860 legacy behaviour is preserved as much as possible.
4864 * In DSA_generate_parameters_ex, if the provided seed is too short,
4867 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4869 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4871 * Alternate chains certificate forgery
4873 During certificate verification, OpenSSL will attempt to find an
4874 alternative certificate chain if the first attempt to build such a chain
4875 fails. An error in the implementation of this logic can mean that an
4876 attacker could cause certain checks on untrusted certificates to be
4877 bypassed, such as the CA flag, enabling them to use a valid leaf
4878 certificate to act as a CA and "issue" an invalid certificate.
4880 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4885 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4887 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4888 incompatibility in the handling of HMAC. The previous ABI has now been
4893 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4895 * Malformed ECParameters causes infinite loop
4897 When processing an ECParameters structure OpenSSL enters an infinite loop
4898 if the curve specified is over a specially malformed binary polynomial
4901 This can be used to perform denial of service against any
4902 system which processes public keys, certificate requests or
4903 certificates. This includes TLS clients and TLS servers with
4904 client authentication enabled.
4906 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4911 * Exploitable out-of-bounds read in X509_cmp_time
4913 X509_cmp_time does not properly check the length of the ASN1_TIME
4914 string and can read a few bytes out of bounds. In addition,
4915 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4918 An attacker can use this to craft malformed certificates and CRLs of
4919 various sizes and potentially cause a segmentation fault, resulting in
4920 a DoS on applications that verify certificates or CRLs. TLS clients
4921 that verify CRLs are affected. TLS clients and servers with client
4922 authentication enabled may be affected if they use custom verification
4925 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4926 independently by Hanno Böck.
4931 * PKCS7 crash with missing EnvelopedContent
4933 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4934 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4935 with missing content and trigger a NULL pointer dereference on parsing.
4937 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4938 structures from untrusted sources are affected. OpenSSL clients and
4939 servers are not affected.
4941 This issue was reported to OpenSSL by Michal Zalewski (Google).
4946 * CMS verify infinite loop with unknown hash function
4948 When verifying a signedData message the CMS code can enter an infinite loop
4949 if presented with an unknown hash function OID. This can be used to perform
4950 denial of service against any system which verifies signedData messages using
4952 This issue was reported to OpenSSL by Johannes Bauer.
4957 * Race condition handling NewSessionTicket
4959 If a NewSessionTicket is received by a multi-threaded client when attempting to
4960 reuse a previous ticket then a race condition can occur potentially leading to
4961 a double free of the ticket data.
4966 * Only support 256-bit or stronger elliptic curves with the
4967 'ecdh_auto' setting (server) or by default (client). Of supported
4968 curves, prefer P-256 (both).
4972 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4974 * ClientHello sigalgs DoS fix
4976 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4977 invalid signature algorithms extension a NULL pointer dereference will
4978 occur. This can be exploited in a DoS attack against the server.
4980 This issue was was reported to OpenSSL by David Ramos of Stanford
4984 *Stephen Henson and Matt Caswell*
4986 * Multiblock corrupted pointer fix
4988 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4989 feature only applies on 64 bit x86 architecture platforms that support AES
4990 NI instructions. A defect in the implementation of "multiblock" can cause
4991 OpenSSL's internal write buffer to become incorrectly set to NULL when
4992 using non-blocking IO. Typically, when the user application is using a
4993 socket BIO for writing, this will only result in a failed connection.
4994 However if some other BIO is used then it is likely that a segmentation
4995 fault will be triggered, thus enabling a potential DoS attack.
4997 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5002 * Segmentation fault in DTLSv1_listen fix
5004 The DTLSv1_listen function is intended to be stateless and processes the
5005 initial ClientHello from many peers. It is common for user code to loop
5006 over the call to DTLSv1_listen until a valid ClientHello is received with
5007 an associated cookie. A defect in the implementation of DTLSv1_listen means
5008 that state is preserved in the SSL object from one invocation to the next
5009 that can lead to a segmentation fault. Errors processing the initial
5010 ClientHello can trigger this scenario. An example of such an error could be
5011 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5014 This issue was reported to OpenSSL by Per Allansson.
5019 * Segmentation fault in ASN1_TYPE_cmp fix
5021 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5022 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5023 certificate signature algorithm consistency this can be used to crash any
5024 certificate verification operation and exploited in a DoS attack. Any
5025 application which performs certificate verification is vulnerable including
5026 OpenSSL clients and servers which enable client authentication.
5031 * Segmentation fault for invalid PSS parameters fix
5033 The signature verification routines will crash with a NULL pointer
5034 dereference if presented with an ASN.1 signature using the RSA PSS
5035 algorithm and invalid parameters. Since these routines are used to verify
5036 certificate signature algorithms this can be used to crash any
5037 certificate verification operation and exploited in a DoS attack. Any
5038 application which performs certificate verification is vulnerable including
5039 OpenSSL clients and servers which enable client authentication.
5041 This issue was was reported to OpenSSL by Brian Carpenter.
5046 * ASN.1 structure reuse memory corruption fix
5048 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5049 memory corruption via an invalid write. Such reuse is and has been
5050 strongly discouraged and is believed to be rare.
5052 Applications that parse structures containing CHOICE or ANY DEFINED BY
5053 components may be affected. Certificate parsing (d2i_X509 and related
5054 functions) are however not affected. OpenSSL clients and servers are
5060 * PKCS7 NULL pointer dereferences fix
5062 The PKCS#7 parsing code does not handle missing outer ContentInfo
5063 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5064 missing content and trigger a NULL pointer dereference on parsing.
5066 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5067 otherwise parse PKCS#7 structures from untrusted sources are
5068 affected. OpenSSL clients and servers are not affected.
5070 This issue was reported to OpenSSL by Michal Zalewski (Google).
5075 * DoS via reachable assert in SSLv2 servers fix
5077 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5078 servers that both support SSLv2 and enable export cipher suites by sending
5079 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5081 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5082 (OpenSSL development team).
5087 * Empty CKE with client auth and DHE fix
5089 If client auth is used then a server can seg fault in the event of a DHE
5090 ciphersuite being selected and a zero length ClientKeyExchange message
5091 being sent by the client. This could be exploited in a DoS attack.
5096 * Handshake with unseeded PRNG fix
5098 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5099 with an unseeded PRNG. The conditions are:
5100 - The client is on a platform where the PRNG has not been seeded
5101 automatically, and the user has not seeded manually
5102 - A protocol specific client method version has been used (i.e. not
5103 SSL_client_methodv23)
5104 - A ciphersuite is used that does not require additional random data from
5105 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5107 If the handshake succeeds then the client random that has been used will
5108 have been generated from a PRNG with insufficient entropy and therefore the
5109 output may be predictable.
5111 For example using the following command with an unseeded openssl will
5112 succeed on an unpatched platform:
5114 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5119 * Use After Free following d2i_ECPrivatekey error fix
5121 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5122 could cause a use after free condition. This, in turn, could cause a double
5123 free in several private key parsing functions (such as d2i_PrivateKey
5124 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5125 for applications that receive EC private keys from untrusted
5126 sources. This scenario is considered rare.
5128 This issue was discovered by the BoringSSL project and fixed in their
5134 * X509_to_X509_REQ NULL pointer deref fix
5136 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5137 the certificate key is invalid. This function is rarely used in practice.
5139 This issue was discovered by Brian Carpenter.
5144 * Removed the export ciphers from the DEFAULT ciphers
5148 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5150 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5151 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5152 So far those who have to target multiple platforms would compromise
5153 and argue that binary targeting say ARMv5 would still execute on
5154 ARMv8. "Universal" build resolves this compromise by providing
5155 near-optimal performance even on newer platforms.
5159 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5160 (other platforms pending).
5162 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5164 * Add support for the SignedCertificateTimestampList certificate and
5165 OCSP response extensions from RFC6962.
5169 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5170 for corner cases. (Certain input points at infinity could lead to
5171 bogus results, with non-infinity inputs mapped to infinity too.)
5175 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5176 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5177 common cases are optimized and there still is room for further
5178 improvements. Vector Permutation AES for Altivec is also added.
5182 * Add support for little-endian ppc64 Linux target.
5184 *Marcelo Cerri (IBM)*
5186 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5187 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5188 are optimized and there still is room for further improvements.
5189 Both 32- and 64-bit modes are supported.
5191 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5193 * Improved ARMv7 NEON support.
5197 * Support for SPARC Architecture 2011 crypto extensions, first
5198 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5199 SHA256/512, MD5, GHASH and modular exponentiation.
5201 *Andy Polyakov, David Miller*
5203 * Accelerated modular exponentiation for Intel processors, a.k.a.
5206 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5208 * Support for new and upcoming Intel processors, including AVX2,
5209 BMI and SHA ISA extensions. This includes additional "stitched"
5210 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5213 This work was sponsored by Intel Corp.
5217 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5218 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5219 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5223 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5224 this fixes a limitation in previous versions of OpenSSL.
5228 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5229 MGF1 digest and OAEP label.
5233 * Add EVP support for key wrapping algorithms, to avoid problems with
5234 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5235 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5236 algorithms and include tests cases.
5240 * Add functions to allocate and set the fields of an ECDSA_METHOD
5243 *Douglas E. Engert, Steve Henson*
5245 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5246 difference in days and seconds between two tm or ASN1_TIME structures.
5250 * Add -rev test option to s_server to just reverse order of characters
5251 received by client and send back to server. Also prints an abbreviated
5252 summary of the connection parameters.
5256 * New option -brief for s_client and s_server to print out a brief summary
5257 of connection parameters.
5261 * Add callbacks for arbitrary TLS extensions.
5263 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5265 * New option -crl_download in several openssl utilities to download CRLs
5266 from CRLDP extension in certificates.
5270 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5274 * New function X509_CRL_diff to generate a delta CRL from the difference
5275 of two full CRLs. Add support to "crl" utility.
5279 * New functions to set lookup_crls function and to retrieve
5280 X509_STORE from X509_STORE_CTX.
5284 * Print out deprecated issuer and subject unique ID fields in
5289 * Extend OCSP I/O functions so they can be used for simple general purpose
5290 HTTP as well as OCSP. New wrapper function which can be used to download
5291 CRLs using the OCSP API.
5295 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5299 * `SSL_CONF*` functions. These provide a common framework for application
5300 configuration using configuration files or command lines.
5304 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5305 message callback and prints the results. Needs compile time option
5306 "enable-ssl-trace". New options to s_client and s_server to enable
5311 * New ctrl and macro to retrieve supported points extensions.
5312 Print out extension in s_server and s_client.
5316 * New functions to retrieve certificate signature and signature
5321 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5326 * New Suite B modes for TLS code. These use and enforce the requirements
5327 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5328 only use Suite B curves. The Suite B modes can be set by using the
5329 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5333 * New chain verification flags for Suite B levels of security. Check
5334 algorithms are acceptable when flags are set in X509_verify_cert.
5338 * Make tls1_check_chain return a set of flags indicating checks passed
5339 by a certificate chain. Add additional tests to handle client
5340 certificates: checks for matching certificate type and issuer name
5345 * If an attempt is made to use a signature algorithm not in the peer
5346 preference list abort the handshake. If client has no suitable
5347 signature algorithms in response to a certificate request do not
5348 use the certificate.
5352 * If server EC tmp key is not in client preference list abort handshake.
5356 * Add support for certificate stores in CERT structure. This makes it
5357 possible to have different stores per SSL structure or one store in
5358 the parent SSL_CTX. Include distinct stores for certificate chain
5359 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5360 to build and store a certificate chain in CERT structure: returning
5361 an error if the chain cannot be built: this will allow applications
5362 to test if a chain is correctly configured.
5364 Note: if the CERT based stores are not set then the parent SSL_CTX
5365 store is used to retain compatibility with existing behaviour.
5369 * New function ssl_set_client_disabled to set a ciphersuite disabled
5370 mask based on the current session, check mask when sending client
5371 hello and checking the requested ciphersuite.
5375 * New ctrls to retrieve and set certificate types in a certificate
5376 request message. Print out received values in s_client. If certificate
5377 types is not set with custom values set sensible values based on
5378 supported signature algorithms.
5382 * Support for distinct client and server supported signature algorithms.
5386 * Add certificate callback. If set this is called whenever a certificate
5387 is required by client or server. An application can decide which
5388 certificate chain to present based on arbitrary criteria: for example
5389 supported signature algorithms. Add very simple example to s_server.
5390 This fixes many of the problems and restrictions of the existing client
5391 certificate callback: for example you can now clear an existing
5392 certificate and specify the whole chain.
5396 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5397 the certificate can be used for (if anything). Set valid_flags field
5398 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5399 to have similar checks in it.
5401 Add new "cert_flags" field to CERT structure and include a "strict mode".
5402 This enforces some TLS certificate requirements (such as only permitting
5403 certificate signature algorithms contained in the supported algorithms
5404 extension) which some implementations ignore: this option should be used
5405 with caution as it could cause interoperability issues.
5409 * Update and tidy signature algorithm extension processing. Work out
5410 shared signature algorithms based on preferences and peer algorithms
5411 and print them out in s_client and s_server. Abort handshake if no
5412 shared signature algorithms.
5416 * Add new functions to allow customised supported signature algorithms
5417 for SSL and SSL_CTX structures. Add options to s_client and s_server
5422 * New function SSL_certs_clear() to delete all references to certificates
5423 from an SSL structure. Before this once a certificate had been added
5424 it couldn't be removed.
5428 * Integrate hostname, email address and IP address checking with certificate
5429 verification. New verify options supporting checking in openssl utility.
5433 * Fixes and wildcard matching support to hostname and email checking
5434 functions. Add manual page.
5436 *Florian Weimer (Red Hat Product Security Team)*
5438 * New functions to check a hostname email or IP address against a
5439 certificate. Add options x509 utility to print results of checks against
5444 * Fix OCSP checking.
5446 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5448 * Initial experimental support for explicitly trusted non-root CAs.
5449 OpenSSL still tries to build a complete chain to a root but if an
5450 intermediate CA has a trust setting included that is used. The first
5451 setting is used: whether to trust (e.g., -addtrust option to the x509
5456 * Add -trusted_first option which attempts to find certificates in the
5457 trusted store even if an untrusted chain is also supplied.
5461 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5462 platform support for Linux and Android.
5466 * Support for linux-x32, ILP32 environment in x86_64 framework.
5470 * Experimental multi-implementation support for FIPS capable OpenSSL.
5471 When in FIPS mode the approved implementations are used as normal,
5472 when not in FIPS mode the internal unapproved versions are used instead.
5473 This means that the FIPS capable OpenSSL isn't forced to use the
5474 (often lower performance) FIPS implementations outside FIPS mode.
5478 * Transparently support X9.42 DH parameters when calling
5479 PEM_read_bio_DHparameters. This means existing applications can handle
5480 the new parameter format automatically.
5484 * Initial experimental support for X9.42 DH parameter format: mainly
5485 to support use of 'q' parameter for RFC5114 parameters.
5489 * Add DH parameters from RFC5114 including test data to dhtest.
5493 * Support for automatic EC temporary key parameter selection. If enabled
5494 the most preferred EC parameters are automatically used instead of
5495 hardcoded fixed parameters. Now a server just has to call:
5496 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5497 support ECDH and use the most appropriate parameters.
5501 * Enhance and tidy EC curve and point format TLS extension code. Use
5502 static structures instead of allocation if default values are used.
5503 New ctrls to set curves we wish to support and to retrieve shared curves.
5504 Print out shared curves in s_server. New options to s_server and s_client
5505 to set list of supported curves.
5509 * New ctrls to retrieve supported signature algorithms and
5510 supported curve values as an array of NIDs. Extend openssl utility
5511 to print out received values.
5515 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5516 between NIDs and the more common NIST names such as "P-256". Enhance
5517 ecparam utility and ECC method to recognise the NIST names for curves.
5521 * Enhance SSL/TLS certificate chain handling to support different
5522 chains for each certificate instead of one chain in the parent SSL_CTX.
5526 * Support for fixed DH ciphersuite client authentication: where both
5527 server and client use DH certificates with common parameters.
5531 * Support for fixed DH ciphersuites: those requiring DH server
5536 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5538 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5539 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5540 X509_CINF_get_signature were reverted post internal team review.
5545 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5547 * OCSP Status Request extension unbounded memory growth
5549 A malicious client can send an excessively large OCSP Status Request
5550 extension. If that client continually requests renegotiation, sending a
5551 large OCSP Status Request extension each time, then there will be unbounded
5552 memory growth on the server. This will eventually lead to a Denial Of
5553 Service attack through memory exhaustion. Servers with a default
5554 configuration are vulnerable even if they do not support OCSP. Builds using
5555 the "no-ocsp" build time option are not affected.
5557 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5562 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5565 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5571 * OOB write in MDC2_Update()
5573 An overflow can occur in MDC2_Update() either if called directly or
5574 through the EVP_DigestUpdate() function using MDC2. If an attacker
5575 is able to supply very large amounts of input data after a previous
5576 call to EVP_EncryptUpdate() with a partial block then a length check
5577 can overflow resulting in a heap corruption.
5579 The amount of data needed is comparable to SIZE_MAX which is impractical
5582 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5587 * Malformed SHA512 ticket DoS
5589 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5590 DoS attack where a malformed ticket will result in an OOB read which will
5593 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5594 a custom server callback and ticket lookup mechanism.
5596 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5601 * OOB write in BN_bn2dec()
5603 The function BN_bn2dec() does not check the return value of BN_div_word().
5604 This can cause an OOB write if an application uses this function with an
5605 overly large BIGNUM. This could be a problem if an overly large certificate
5606 or CRL is printed out from an untrusted source. TLS is not affected because
5607 record limits will reject an oversized certificate before it is parsed.
5609 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5614 * OOB read in TS_OBJ_print_bio()
5616 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5617 the total length the OID text representation would use and not the amount
5618 of data written. This will result in OOB reads when large OIDs are
5621 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5626 * Pointer arithmetic undefined behaviour
5628 Avoid some undefined pointer arithmetic
5630 A common idiom in the codebase is to check limits in the following manner:
5633 Where "p" points to some malloc'd data of SIZE bytes and
5636 "len" here could be from some externally supplied data (e.g. from a TLS
5639 The rules of C pointer arithmetic are such that "p + len" is only well
5640 defined where len <= SIZE. Therefore the above idiom is actually
5641 undefined behaviour.
5643 For example this could cause problems if some malloc implementation
5644 provides an address for "p" such that "p + len" actually overflows for
5645 values of len that are too big and therefore p + len < limit.
5647 This issue was reported to OpenSSL by Guido Vranken
5652 * Constant time flag not preserved in DSA signing
5654 Operations in the DSA signing algorithm should run in constant time in
5655 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5656 implementation means that a non-constant time codepath is followed for
5657 certain operations. This has been demonstrated through a cache-timing
5658 attack to be sufficient for an attacker to recover the private DSA key.
5660 This issue was reported by César Pereida (Aalto University), Billy Brumley
5661 (Tampere University of Technology), and Yuval Yarom (The University of
5662 Adelaide and NICTA).
5667 * DTLS buffered message DoS
5669 In a DTLS connection where handshake messages are delivered out-of-order
5670 those messages that OpenSSL is not yet ready to process will be buffered
5671 for later use. Under certain circumstances, a flaw in the logic means that
5672 those messages do not get removed from the buffer even though the handshake
5673 has been completed. An attacker could force up to approx. 15 messages to
5674 remain in the buffer when they are no longer required. These messages will
5675 be cleared when the DTLS connection is closed. The default maximum size for
5676 a message is 100k. Therefore the attacker could force an additional 1500k
5677 to be consumed per connection. By opening many simulataneous connections an
5678 attacker could cause a DoS attack through memory exhaustion.
5680 This issue was reported to OpenSSL by Quan Luo.
5685 * DTLS replay protection DoS
5687 A flaw in the DTLS replay attack protection mechanism means that records
5688 that arrive for future epochs update the replay protection "window" before
5689 the MAC for the record has been validated. This could be exploited by an
5690 attacker by sending a record for the next epoch (which does not have to
5691 decrypt or have a valid MAC), with a very large sequence number. This means
5692 that all subsequent legitimate packets are dropped causing a denial of
5693 service for a specific DTLS connection.
5695 This issue was reported to OpenSSL by the OCAP audit team.
5700 * Certificate message OOB reads
5702 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5703 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5704 theoretical DoS risk but this has not been observed in practice on common
5707 The messages affected are client certificate, client certificate request
5708 and server certificate. As a result the attack can only be performed
5709 against a client or a server which enables client authentication.
5711 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5716 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5718 * Prevent padding oracle in AES-NI CBC MAC check
5720 A MITM attacker can use a padding oracle attack to decrypt traffic
5721 when the connection uses an AES CBC cipher and the server support
5724 This issue was introduced as part of the fix for Lucky 13 padding
5725 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5726 constant time by making sure that always the same bytes are read and
5727 compared against either the MAC or padding bytes. But it no longer
5728 checked that there was enough data to have both the MAC and padding
5731 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5736 * Fix EVP_EncodeUpdate overflow
5738 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5739 Base64 encoding of binary data. If an attacker is able to supply very large
5740 amounts of input data then a length check can overflow resulting in a heap
5743 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5744 the `PEM_write_bio*` family of functions. These are mainly used within the
5745 OpenSSL command line applications, so any application which processes data
5746 from an untrusted source and outputs it as a PEM file should be considered
5747 vulnerable to this issue. User applications that call these APIs directly
5748 with large amounts of untrusted data may also be vulnerable.
5750 This issue was reported by Guido Vranken.
5755 * Fix EVP_EncryptUpdate overflow
5757 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5758 is able to supply very large amounts of input data after a previous call to
5759 EVP_EncryptUpdate() with a partial block then a length check can overflow
5760 resulting in a heap corruption. Following an analysis of all OpenSSL
5761 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5762 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5763 the first called function after an EVP_EncryptInit(), and therefore that
5764 specific call must be safe. The second form is where the length passed to
5765 EVP_EncryptUpdate() can be seen from the code to be some small value and
5766 therefore there is no possibility of an overflow. Since all instances are
5767 one of these two forms, it is believed that there can be no overflows in
5768 internal code due to this problem. It should be noted that
5769 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5770 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5771 of these calls have also been analysed too and it is believed there are no
5772 instances in internal usage where an overflow could occur.
5774 This issue was reported by Guido Vranken.
5779 * Prevent ASN.1 BIO excessive memory allocation
5781 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5782 a short invalid encoding can casuse allocation of large amounts of memory
5783 potentially consuming excessive resources or exhausting memory.
5785 Any application parsing untrusted data through d2i BIO functions is
5786 affected. The memory based functions such as d2i_X509() are *not* affected.
5787 Since the memory based functions are used by the TLS library, TLS
5788 applications are not affected.
5790 This issue was reported by Brian Carpenter.
5797 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5798 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5799 in arbitrary stack data being returned in the buffer.
5801 This issue was reported by Guido Vranken.
5806 * Modify behavior of ALPN to invoke callback after SNI/servername
5807 callback, such that updates to the SSL_CTX affect ALPN.
5811 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5816 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5817 methods are enabled and ssl2 is disabled the methods return NULL.
5821 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5823 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5824 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5825 provide any "EXPORT" or "LOW" strength ciphers.
5829 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5830 is by default disabled at build-time. Builds that are not configured with
5831 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5832 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5833 will need to explicitly call either of:
5835 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5837 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5839 as appropriate. Even if either of those is used, or the application
5840 explicitly uses the version-specific SSLv2_method() or its client and
5841 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5842 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5843 ciphers, and SSLv2 56-bit DES are no longer available.
5848 * Fix a double-free in DSA code
5850 A double free bug was discovered when OpenSSL parses malformed DSA private
5851 keys and could lead to a DoS attack or memory corruption for applications
5852 that receive DSA private keys from untrusted sources. This scenario is
5855 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5861 * Disable SRP fake user seed to address a server memory leak.
5863 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5865 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5866 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5867 was changed to ignore the "fake user" SRP seed, even if the seed
5870 Users should use SRP_VBASE_get1_by_user instead. Note that in
5871 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5872 also that even though configuring the SRP seed attempts to hide
5873 invalid usernames by continuing the handshake with fake
5874 credentials, this behaviour is not constant time and no strong
5875 guarantees are made that the handshake is indistinguishable from
5876 that of a valid user.
5881 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5883 In the BN_hex2bn function the number of hex digits is calculated using an
5884 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5885 large values of `i` this can result in `bn_expand` not allocating any
5886 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5887 field as NULL leading to a subsequent NULL ptr deref. For very large values
5888 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5889 In this case memory is allocated to the internal BIGNUM data field, but it
5890 is insufficiently sized leading to heap corruption. A similar issue exists
5891 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5892 is ever called by user applications with very large untrusted hex/dec data.
5893 This is anticipated to be a rare occurrence.
5895 All OpenSSL internal usage of these functions use data that is not expected
5896 to be untrusted, e.g. config file data or application command line
5897 arguments. If user developed applications generate config file data based
5898 on untrusted data then it is possible that this could also lead to security
5899 consequences. This is also anticipated to be rare.
5901 This issue was reported to OpenSSL by Guido Vranken.
5906 * Fix memory issues in `BIO_*printf` functions
5908 The internal `fmtstr` function used in processing a "%s" format string in
5909 the `BIO_*printf` functions could overflow while calculating the length of a
5910 string and cause an OOB read when printing very long strings.
5912 Additionally the internal `doapr_outch` function can attempt to write to an
5913 OOB memory location (at an offset from the NULL pointer) in the event of a
5914 memory allocation failure. In 1.0.2 and below this could be caused where
5915 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5916 could be in processing a very long "%s" format string. Memory leaks can
5919 The first issue may mask the second issue dependent on compiler behaviour.
5920 These problems could enable attacks where large amounts of untrusted data
5921 is passed to the `BIO_*printf` functions. If applications use these functions
5922 in this way then they could be vulnerable. OpenSSL itself uses these
5923 functions when printing out human-readable dumps of ASN.1 data. Therefore
5924 applications that print this data could be vulnerable if the data is from
5925 untrusted sources. OpenSSL command line applications could also be
5926 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5927 as command line arguments.
5929 Libssl is not considered directly vulnerable. Additionally certificates etc
5930 received via remote connections via libssl are also unlikely to be able to
5931 trigger these issues because of message size limits enforced within libssl.
5933 This issue was reported to OpenSSL Guido Vranken.
5938 * Side channel attack on modular exponentiation
5940 A side-channel attack was found which makes use of cache-bank conflicts on
5941 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5942 of RSA keys. The ability to exploit this issue is limited as it relies on
5943 an attacker who has control of code in a thread running on the same
5944 hyper-threaded core as the victim thread which is performing decryptions.
5946 This issue was reported to OpenSSL by Yuval Yarom, The University of
5947 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5948 Nadia Heninger, University of Pennsylvania with more information at
5949 <http://cachebleed.info>.
5954 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5955 if no keysize is specified with default_bits. This fixes an
5956 omission in an earlier change that changed all RSA/DSA key generation
5957 apps to use 2048 bits by default.
5961 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5963 * Protection for DH small subgroup attacks
5965 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5966 switched on by default and cannot be disabled. This could have some
5971 * SSLv2 doesn't block disabled ciphers
5973 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5974 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5975 been disabled, provided that the SSLv2 protocol was not also disabled via
5978 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5979 and Sebastian Schinzel.
5984 * Reject DH handshakes with parameters shorter than 1024 bits.
5988 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5990 * Certificate verify crash with missing PSS parameter
5992 The signature verification routines will crash with a NULL pointer
5993 dereference if presented with an ASN.1 signature using the RSA PSS
5994 algorithm and absent mask generation function parameter. Since these
5995 routines are used to verify certificate signature algorithms this can be
5996 used to crash any certificate verification operation and exploited in a
5997 DoS attack. Any application which performs certificate verification is
5998 vulnerable including OpenSSL clients and servers which enable client
6001 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6006 * X509_ATTRIBUTE memory leak
6008 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6009 memory. This structure is used by the PKCS#7 and CMS routines so any
6010 application which reads PKCS#7 or CMS data from untrusted sources is
6011 affected. SSL/TLS is not affected.
6013 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6019 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6020 This changes the decoding behaviour for some invalid messages,
6021 though the change is mostly in the more lenient direction, and
6022 legacy behaviour is preserved as much as possible.
6026 * In DSA_generate_parameters_ex, if the provided seed is too short,
6027 use a random seed, as already documented.
6029 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6031 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6033 * Alternate chains certificate forgery
6035 During certificate verfification, OpenSSL will attempt to find an
6036 alternative certificate chain if the first attempt to build such a chain
6037 fails. An error in the implementation of this logic can mean that an
6038 attacker could cause certain checks on untrusted certificates to be
6039 bypassed, such as the CA flag, enabling them to use a valid leaf
6040 certificate to act as a CA and "issue" an invalid certificate.
6042 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6048 * Race condition handling PSK identify hint
6050 If PSK identity hints are received by a multi-threaded client then
6051 the values are wrongly updated in the parent SSL_CTX structure. This can
6052 result in a race condition potentially leading to a double free of the
6058 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6060 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6061 incompatibility in the handling of HMAC. The previous ABI has now been
6064 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6066 * Malformed ECParameters causes infinite loop
6068 When processing an ECParameters structure OpenSSL enters an infinite loop
6069 if the curve specified is over a specially malformed binary polynomial
6072 This can be used to perform denial of service against any
6073 system which processes public keys, certificate requests or
6074 certificates. This includes TLS clients and TLS servers with
6075 client authentication enabled.
6077 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6082 * Exploitable out-of-bounds read in X509_cmp_time
6084 X509_cmp_time does not properly check the length of the ASN1_TIME
6085 string and can read a few bytes out of bounds. In addition,
6086 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6089 An attacker can use this to craft malformed certificates and CRLs of
6090 various sizes and potentially cause a segmentation fault, resulting in
6091 a DoS on applications that verify certificates or CRLs. TLS clients
6092 that verify CRLs are affected. TLS clients and servers with client
6093 authentication enabled may be affected if they use custom verification
6096 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6097 independently by Hanno Böck.
6102 * PKCS7 crash with missing EnvelopedContent
6104 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6105 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6106 with missing content and trigger a NULL pointer dereference on parsing.
6108 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6109 structures from untrusted sources are affected. OpenSSL clients and
6110 servers are not affected.
6112 This issue was reported to OpenSSL by Michal Zalewski (Google).
6117 * CMS verify infinite loop with unknown hash function
6119 When verifying a signedData message the CMS code can enter an infinite loop
6120 if presented with an unknown hash function OID. This can be used to perform
6121 denial of service against any system which verifies signedData messages using
6123 This issue was reported to OpenSSL by Johannes Bauer.
6128 * Race condition handling NewSessionTicket
6130 If a NewSessionTicket is received by a multi-threaded client when attempting to
6131 reuse a previous ticket then a race condition can occur potentially leading to
6132 a double free of the ticket data.
6137 * Reject DH handshakes with parameters shorter than 768 bits.
6139 *Kurt Roeckx and Emilia Kasper*
6141 * dhparam: generate 2048-bit parameters by default.
6143 *Kurt Roeckx and Emilia Kasper*
6145 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6147 * Segmentation fault in ASN1_TYPE_cmp fix
6149 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6150 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6151 certificate signature algorithm consistency this can be used to crash any
6152 certificate verification operation and exploited in a DoS attack. Any
6153 application which performs certificate verification is vulnerable including
6154 OpenSSL clients and servers which enable client authentication.
6159 * ASN.1 structure reuse memory corruption fix
6161 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6162 memory corruption via an invalid write. Such reuse is and has been
6163 strongly discouraged and is believed to be rare.
6165 Applications that parse structures containing CHOICE or ANY DEFINED BY
6166 components may be affected. Certificate parsing (d2i_X509 and related
6167 functions) are however not affected. OpenSSL clients and servers are
6173 * PKCS7 NULL pointer dereferences fix
6175 The PKCS#7 parsing code does not handle missing outer ContentInfo
6176 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6177 missing content and trigger a NULL pointer dereference on parsing.
6179 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6180 otherwise parse PKCS#7 structures from untrusted sources are
6181 affected. OpenSSL clients and servers are not affected.
6183 This issue was reported to OpenSSL by Michal Zalewski (Google).
6188 * DoS via reachable assert in SSLv2 servers fix
6190 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6191 servers that both support SSLv2 and enable export cipher suites by sending
6192 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6194 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6195 (OpenSSL development team).
6200 * Use After Free following d2i_ECPrivatekey error fix
6202 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6203 could cause a use after free condition. This, in turn, could cause a double
6204 free in several private key parsing functions (such as d2i_PrivateKey
6205 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6206 for applications that receive EC private keys from untrusted
6207 sources. This scenario is considered rare.
6209 This issue was discovered by the BoringSSL project and fixed in their
6215 * X509_to_X509_REQ NULL pointer deref fix
6217 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6218 the certificate key is invalid. This function is rarely used in practice.
6220 This issue was discovered by Brian Carpenter.
6225 * Removed the export ciphers from the DEFAULT ciphers
6229 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6231 * Build fixes for the Windows and OpenVMS platforms
6233 *Matt Caswell and Richard Levitte*
6235 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6237 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6238 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6239 dereference. This could lead to a Denial Of Service attack. Thanks to
6240 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6245 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6246 dtls1_buffer_record function under certain conditions. In particular this
6247 could occur if an attacker sent repeated DTLS records with the same
6248 sequence number but for the next epoch. The memory leak could be exploited
6249 by an attacker in a Denial of Service attack through memory exhaustion.
6250 Thanks to Chris Mueller for reporting this issue.
6255 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6256 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6257 method would be set to NULL which could later result in a NULL pointer
6258 dereference. Thanks to Frank Schmirler for reporting this issue.
6263 * Abort handshake if server key exchange message is omitted for ephemeral
6266 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6267 reporting this issue.
6272 * Remove non-export ephemeral RSA code on client and server. This code
6273 violated the TLS standard by allowing the use of temporary RSA keys in
6274 non-export ciphersuites and could be used by a server to effectively
6275 downgrade the RSA key length used to a value smaller than the server
6276 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6277 INRIA or reporting this issue.
6282 * Fixed issue where DH client certificates are accepted without verification.
6283 An OpenSSL server will accept a DH certificate for client authentication
6284 without the certificate verify message. This effectively allows a client to
6285 authenticate without the use of a private key. This only affects servers
6286 which trust a client certificate authority which issues certificates
6287 containing DH keys: these are extremely rare and hardly ever encountered.
6288 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6294 * Ensure that the session ID context of an SSL is updated when its
6295 SSL_CTX is updated via SSL_set_SSL_CTX.
6297 The session ID context is typically set from the parent SSL_CTX,
6298 and can vary with the CTX.
6302 * Fix various certificate fingerprint issues.
6304 By using non-DER or invalid encodings outside the signed portion of a
6305 certificate the fingerprint can be changed without breaking the signature.
6306 Although no details of the signed portion of the certificate can be changed
6307 this can cause problems with some applications: e.g. those using the
6308 certificate fingerprint for blacklists.
6310 1. Reject signatures with non zero unused bits.
6312 If the BIT STRING containing the signature has non zero unused bits reject
6313 the signature. All current signature algorithms require zero unused bits.
6315 2. Check certificate algorithm consistency.
6317 Check the AlgorithmIdentifier inside TBS matches the one in the
6318 certificate signature. NB: this will result in signature failure
6319 errors for some broken certificates.
6321 Thanks to Konrad Kraszewski from Google for reporting this issue.
6323 3. Check DSA/ECDSA signatures use DER.
6325 Re-encode DSA/ECDSA signatures and compare with the original received
6326 signature. Return an error if there is a mismatch.
6328 This will reject various cases including garbage after signature
6329 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6330 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6331 (negative or with leading zeroes).
6333 Further analysis was conducted and fixes were developed by Stephen Henson
6334 of the OpenSSL core team.
6340 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6341 results on some platforms, including x86_64. This bug occurs at random
6342 with a very low probability, and is not known to be exploitable in any
6343 way, though its exact impact is difficult to determine. Thanks to Pieter
6344 Wuille (Blockstream) who reported this issue and also suggested an initial
6345 fix. Further analysis was conducted by the OpenSSL development team and
6346 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6347 the OpenSSL core team.
6352 * Do not resume sessions on the server if the negotiated protocol
6353 version does not match the session's version. Resuming with a different
6354 version, while not strictly forbidden by the RFC, is of questionable
6355 sanity and breaks all known clients.
6357 *David Benjamin, Emilia Käsper*
6359 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6360 early CCS messages during renegotiation. (Note that because
6361 renegotiation is encrypted, this early CCS was not exploitable.)
6365 * Tighten client-side session ticket handling during renegotiation:
6366 ensure that the client only accepts a session ticket if the server sends
6367 the extension anew in the ServerHello. Previously, a TLS client would
6368 reuse the old extension state and thus accept a session ticket if one was
6369 announced in the initial ServerHello.
6371 Similarly, ensure that the client requires a session ticket if one
6372 was advertised in the ServerHello. Previously, a TLS client would
6373 ignore a missing NewSessionTicket message.
6377 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6381 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6382 sends a carefully crafted handshake message, to cause OpenSSL to fail
6383 to free up to 64k of memory causing a memory leak. This could be
6384 exploited in a Denial Of Service attack. This issue affects OpenSSL
6385 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6386 whether SRTP is used or configured. Implementations of OpenSSL that
6387 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6389 The fix was developed by the OpenSSL team.
6394 * Session Ticket Memory Leak.
6396 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6397 integrity of that ticket is first verified. In the event of a session
6398 ticket integrity check failing, OpenSSL will fail to free memory
6399 causing a memory leak. By sending a large number of invalid session
6400 tickets an attacker could exploit this issue in a Denial Of Service
6406 * Build option no-ssl3 is incomplete.
6408 When OpenSSL is configured with "no-ssl3" as a build option, servers
6409 could accept and complete a SSL 3.0 handshake, and clients could be
6410 configured to send them.
6413 *Akamai and the OpenSSL team*
6415 * Add support for TLS_FALLBACK_SCSV.
6416 Client applications doing fallback retries should call
6417 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6420 *Adam Langley, Bodo Moeller*
6422 * Add additional DigestInfo checks.
6424 Re-encode DigestInto in DER and check against the original when
6425 verifying RSA signature: this will reject any improperly encoded
6426 DigestInfo structures.
6428 Note: this is a precautionary measure and no attacks are currently known.
6432 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6434 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6435 SRP code can be overrun an internal buffer. Add sanity check that
6436 g, A, B < N to SRP code.
6438 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6439 Group for discovering this issue.
6444 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6445 TLS 1.0 instead of higher protocol versions when the ClientHello message
6446 is badly fragmented. This allows a man-in-the-middle attacker to force a
6447 downgrade to TLS 1.0 even if both the server and the client support a
6448 higher protocol version, by modifying the client's TLS records.
6450 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6451 researching this issue.
6456 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6457 to a denial of service attack. A malicious server can crash the client
6458 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6459 ciphersuite and sending carefully crafted handshake messages.
6461 Thanks to Felix Gröbert (Google) for discovering and researching this
6467 * By sending carefully crafted DTLS packets an attacker could cause openssl
6468 to leak memory. This can be exploited through a Denial of Service attack.
6469 Thanks to Adam Langley for discovering and researching this issue.
6474 * An attacker can force openssl to consume large amounts of memory whilst
6475 processing DTLS handshake messages. This can be exploited through a
6476 Denial of Service attack.
6477 Thanks to Adam Langley for discovering and researching this issue.
6482 * An attacker can force an error condition which causes openssl to crash
6483 whilst processing DTLS packets due to memory being freed twice. This
6484 can be exploited through a Denial of Service attack.
6485 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6491 * If a multithreaded client connects to a malicious server using a resumed
6492 session and the server sends an ec point format extension it could write
6493 up to 255 bytes to freed memory.
6495 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6501 * A malicious server can crash an OpenSSL client with a null pointer
6502 dereference (read) by specifying an SRP ciphersuite even though it was not
6503 properly negotiated with the client. This can be exploited through a
6504 Denial of Service attack.
6506 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6507 discovering and researching this issue.
6512 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6513 X509_name_oneline, X509_name_print_ex et al. to leak some information
6514 from the stack. Applications may be affected if they echo pretty printing
6515 output to the attacker.
6517 Thanks to Ivan Fratric (Google) for discovering this issue.
6520 *Emilia Käsper, and Steve Henson*
6522 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6523 for corner cases. (Certain input points at infinity could lead to
6524 bogus results, with non-infinity inputs mapped to infinity too.)
6528 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6530 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6531 handshake can force the use of weak keying material in OpenSSL
6532 SSL/TLS clients and servers.
6534 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6535 researching this issue. ([CVE-2014-0224])
6537 *KIKUCHI Masashi, Steve Henson*
6539 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6540 OpenSSL DTLS client the code can be made to recurse eventually crashing
6543 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6546 *Imre Rad, Steve Henson*
6548 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6549 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6550 client or server. This is potentially exploitable to run arbitrary
6551 code on a vulnerable client or server.
6553 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6555 *Jüri Aedla, Steve Henson*
6557 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6558 are subject to a denial of service attack.
6560 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6561 this issue. ([CVE-2014-3470])
6563 *Felix Gröbert, Ivan Fratric, Steve Henson*
6565 * Harmonize version and its documentation. -f flag is used to display
6568 *mancha <mancha1@zoho.com>*
6570 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6571 in i2d_ECPrivateKey.
6573 *mancha <mancha1@zoho.com>*
6575 * Fix some double frees. These are not thought to be exploitable.
6577 *mancha <mancha1@zoho.com>*
6579 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6581 * A missing bounds check in the handling of the TLS heartbeat extension
6582 can be used to reveal up to 64k of memory to a connected client or
6585 Thanks for Neel Mehta of Google Security for discovering this bug and to
6586 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6587 preparing the fix ([CVE-2014-0160])
6589 *Adam Langley, Bodo Moeller*
6591 * Fix for the attack described in the paper "Recovering OpenSSL
6592 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6593 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6594 <http://eprint.iacr.org/2014/140>
6596 Thanks to Yuval Yarom and Naomi Benger for discovering this
6597 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6599 *Yuval Yarom and Naomi Benger*
6601 * TLS pad extension: draft-agl-tls-padding-03
6603 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6604 TLS client Hello record length value would otherwise be > 255 and
6605 less that 512 pad with a dummy extension containing zeroes so it
6606 is at least 512 bytes long.
6608 *Adam Langley, Steve Henson*
6610 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6612 * Fix for TLS record tampering bug. A carefully crafted invalid
6613 handshake could crash OpenSSL with a NULL pointer exception.
6614 Thanks to Anton Johansson for reporting this issues.
6617 * Keep original DTLS digest and encryption contexts in retransmission
6618 structures so we can use the previous session parameters if they need
6619 to be resent. ([CVE-2013-6450])
6623 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6624 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6625 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6626 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6627 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6628 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6630 *Rob Stradling, Adam Langley*
6632 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6634 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6635 supporting platforms or when small records were transferred.
6637 *Andy Polyakov, Steve Henson*
6639 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6641 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6643 This addresses the flaw in CBC record processing discovered by
6644 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6645 at: <http://www.isg.rhul.ac.uk/tls/>
6647 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6648 Security Group at Royal Holloway, University of London
6649 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6650 Emilia Käsper for the initial patch.
6653 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6655 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6656 ciphersuites which can be exploited in a denial of service attack.
6657 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6658 and detecting this bug and to Wolfgang Ettlinger
6659 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6664 * Return an error when checking OCSP signatures when key is NULL.
6665 This fixes a DoS attack. ([CVE-2013-0166])
6669 * Make openssl verify return errors.
6671 *Chris Palmer <palmer@google.com> and Ben Laurie*
6673 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6674 the right response is stapled. Also change SSL_get_certificate()
6675 so it returns the certificate actually sent.
6676 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6678 *Rob Stradling <rob.stradling@comodo.com>*
6680 * Fix possible deadlock when decoding public keys.
6684 * Don't use TLS 1.0 record version number in initial client hello
6689 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6691 * Sanity check record length before skipping explicit IV in TLS
6692 1.2, 1.1 and DTLS to fix DoS attack.
6694 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6695 fuzzing as a service testing platform.
6700 * Initialise tkeylen properly when encrypting CMS messages.
6701 Thanks to Solar Designer of Openwall for reporting this issue.
6705 * In FIPS mode don't try to use composite ciphers as they are not
6710 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6712 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6713 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6714 mean any application compiled against OpenSSL 1.0.0 headers setting
6715 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6716 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6717 0x10000000L Any application which was previously compiled against
6718 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6719 will need to be recompiled as a result. Letting be results in
6720 inability to disable specifically TLS 1.1 and in client context,
6721 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6725 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6726 disable just protocol X, but all protocols above X *if* there are
6727 protocols *below* X still enabled. In more practical terms it means
6728 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6729 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6730 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6735 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6737 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6738 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6739 in CRYPTO_realloc_clean.
6741 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6742 issue and to Adam Langley <agl@chromium.org> for fixing it.
6745 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6747 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6751 * Workarounds for some broken servers that "hang" if a client hello
6752 record length exceeds 255 bytes.
6754 1. Do not use record version number > TLS 1.0 in initial client
6755 hello: some (but not all) hanging servers will now work.
6756 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6757 the number of ciphers sent in the client hello. This should be
6758 set to an even number, such as 50, for example by passing:
6759 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6760 Most broken servers should now work.
6761 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6762 TLS 1.2 client support entirely.
6766 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6770 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6772 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6773 STRING form instead of a DigestInfo.
6777 * The format used for MDC2 RSA signatures is inconsistent between EVP
6778 and the RSA_sign/RSA_verify functions. This was made more apparent when
6779 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6780 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6781 the correct format in RSA_verify so both forms transparently work.
6785 * Some servers which support TLS 1.0 can choke if we initially indicate
6786 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6787 encrypted premaster secret. As a workaround use the maximum permitted
6788 client version in client hello, this should keep such servers happy
6789 and still work with previous versions of OpenSSL.
6793 * Add support for TLS/DTLS heartbeats.
6795 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6797 * Add support for SCTP.
6799 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6801 * Improved PRNG seeding for VOS.
6803 *Paul Green <Paul.Green@stratus.com>*
6805 * Extensive assembler packs updates, most notably:
6807 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6808 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6809 - x86_64: bit-sliced AES implementation;
6810 - ARM: NEON support, contemporary platforms optimizations;
6811 - s390x: z196 support;
6812 - `*`: GHASH and GF(2^m) multiplication implementations;
6816 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6817 (removal of unnecessary code)
6819 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6821 * Add TLS key material exporter from RFC 5705.
6825 * Add DTLS-SRTP negotiation from RFC 5764.
6829 * Add Next Protocol Negotiation,
6830 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6831 disabled with a no-npn flag to config or Configure. Code donated
6834 *Adam Langley <agl@google.com> and Ben Laurie*
6836 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6837 NIST-P256, NIST-P521, with constant-time single point multiplication on
6838 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6839 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6840 Code made available under Apache License version 2.0.
6842 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6843 line to include this in your build of OpenSSL, and run "make depend" (or
6844 "make update"). This enables the following EC_METHODs:
6846 EC_GFp_nistp224_method()
6847 EC_GFp_nistp256_method()
6848 EC_GFp_nistp521_method()
6850 EC_GROUP_new_by_curve_name() will automatically use these (while
6851 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6854 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6856 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6857 all platforms. Move ssize_t definition from e_os.h to the public
6858 header file e_os2.h as it now appears in public header file cms.h
6862 * New -sigopt option to the ca, req and x509 utilities. Additional
6863 signature parameters can be passed using this option and in
6868 * Add RSA PSS signing function. This will generate and set the
6869 appropriate AlgorithmIdentifiers for PSS based on those in the
6870 corresponding EVP_MD_CTX structure. No application support yet.
6874 * Support for companion algorithm specific ASN1 signing routines.
6875 New function ASN1_item_sign_ctx() signs a pre-initialised
6876 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6877 the appropriate parameters.
6881 * Add new algorithm specific ASN1 verification initialisation function
6882 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6883 handling will be the same no matter what EVP_PKEY_METHOD is used.
6884 Add a PSS handler to support verification of PSS signatures: checked
6885 against a number of sample certificates.
6889 * Add signature printing for PSS. Add PSS OIDs.
6891 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6893 * Add algorithm specific signature printing. An individual ASN1 method
6894 can now print out signatures instead of the standard hex dump.
6896 More complex signatures (e.g. PSS) can print out more meaningful
6897 information. Include DSA version that prints out the signature
6902 * Password based recipient info support for CMS library: implementing
6907 * Split password based encryption into PBES2 and PBKDF2 functions. This
6908 neatly separates the code into cipher and PBE sections and is required
6909 for some algorithms that split PBES2 into separate pieces (such as
6910 password based CMS).
6914 * Session-handling fixes:
6915 - Fix handling of connections that are resuming with a session ID,
6916 but also support Session Tickets.
6917 - Fix a bug that suppressed issuing of a new ticket if the client
6918 presented a ticket with an expired session.
6919 - Try to set the ticket lifetime hint to something reasonable.
6920 - Make tickets shorter by excluding irrelevant information.
6921 - On the client side, don't ignore renewed tickets.
6923 *Adam Langley, Bodo Moeller (Google)*
6925 * Fix PSK session representation.
6929 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6931 This work was sponsored by Intel.
6935 * Add GCM support to TLS library. Some custom code is needed to split
6936 the IV between the fixed (from PRF) and explicit (from TLS record)
6937 portions. This adds all GCM ciphersuites supported by RFC5288 and
6938 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6939 add a special AESGCM string for GCM only.
6943 * Expand range of ctrls for AES GCM. Permit setting invocation
6944 field on decrypt and retrieval of invocation field only on encrypt.
6948 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6949 As required by RFC5289 these ciphersuites cannot be used if for
6950 versions of TLS earlier than 1.2.
6954 * For FIPS capable OpenSSL interpret a NULL default public key method
6955 as unset and return the appropriate default but do *not* set the default.
6956 This means we can return the appropriate method in applications that
6957 switch between FIPS and non-FIPS modes.
6961 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6962 ENGINE is used then we cannot handle that in the FIPS module so we
6963 keep original code iff non-FIPS operations are allowed.
6967 * Add -attime option to openssl utilities.
6969 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6971 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6975 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6976 FIPS EC methods unconditionally for now.
6980 * New build option no-ec2m to disable characteristic 2 code.
6984 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6985 all cases can be covered as some introduce binary incompatibilities.
6989 * Redirect RSA operations to FIPS module including keygen,
6990 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6994 * Add similar low level API blocking to ciphers.
6998 * Low level digest APIs are not approved in FIPS mode: any attempt
6999 to use these will cause a fatal error. Applications that *really* want
7000 to use them can use the `private_*` version instead.
7004 * Redirect cipher operations to FIPS module for FIPS builds.
7008 * Redirect digest operations to FIPS module for FIPS builds.
7012 * Update build system to add "fips" flag which will link in fipscanister.o
7013 for static and shared library builds embedding a signature if needed.
7017 * Output TLS supported curves in preference order instead of numerical
7018 order. This is currently hardcoded for the highest order curves first.
7019 This should be configurable so applications can judge speed vs strength.
7023 * Add TLS v1.2 server support for client authentication.
7027 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7032 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7033 FIPS modules versions.
7037 * Add TLS v1.2 client side support for client authentication. Keep cache
7038 of handshake records longer as we don't know the hash algorithm to use
7039 until after the certificate request message is received.
7043 * Initial TLS v1.2 client support. Add a default signature algorithms
7044 extension including all the algorithms we support. Parse new signature
7045 format in client key exchange. Relax some ECC signing restrictions for
7046 TLS v1.2 as indicated in RFC5246.
7050 * Add server support for TLS v1.2 signature algorithms extension. Switch
7051 to new signature format when needed using client digest preference.
7052 All server ciphersuites should now work correctly in TLS v1.2. No client
7053 support yet and no support for client certificates.
7057 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7058 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7059 ciphersuites. At present only RSA key exchange ciphersuites work with
7060 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7061 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7062 and version checking.
7066 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7067 with this defined it will not be affected by any changes to ssl internal
7068 structures. Add several utility functions to allow openssl application
7069 to work with OPENSSL_NO_SSL_INTERN defined.
7073 * A long standing patch to add support for SRP from EdelWeb (Peter
7074 Sylvester and Christophe Renou) was integrated.
7075 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7076 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7079 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7083 * Permit abbreviated handshakes when renegotiating using the function
7084 SSL_renegotiate_abbreviated().
7086 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7088 * Add call to ENGINE_register_all_complete() to
7089 ENGINE_load_builtin_engines(), so some implementations get used
7090 automatically instead of needing explicit application support.
7094 * Add support for TLS key exporter as described in RFC5705.
7096 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7098 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7099 a few changes are required:
7101 Add SSL_OP_NO_TLSv1_1 flag.
7102 Add TLSv1_1 methods.
7103 Update version checking logic to handle version 1.1.
7104 Add explicit IV handling (ported from DTLS code).
7105 Add command line options to s_client/s_server.
7112 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7114 * X509_ATTRIBUTE memory leak
7116 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7117 memory. This structure is used by the PKCS#7 and CMS routines so any
7118 application which reads PKCS#7 or CMS data from untrusted sources is
7119 affected. SSL/TLS is not affected.
7121 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7127 * Race condition handling PSK identify hint
7129 If PSK identity hints are received by a multi-threaded client then
7130 the values are wrongly updated in the parent SSL_CTX structure. This can
7131 result in a race condition potentially leading to a double free of the
7137 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7139 * Malformed ECParameters causes infinite loop
7141 When processing an ECParameters structure OpenSSL enters an infinite loop
7142 if the curve specified is over a specially malformed binary polynomial
7145 This can be used to perform denial of service against any
7146 system which processes public keys, certificate requests or
7147 certificates. This includes TLS clients and TLS servers with
7148 client authentication enabled.
7150 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7155 * Exploitable out-of-bounds read in X509_cmp_time
7157 X509_cmp_time does not properly check the length of the ASN1_TIME
7158 string and can read a few bytes out of bounds. In addition,
7159 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7162 An attacker can use this to craft malformed certificates and CRLs of
7163 various sizes and potentially cause a segmentation fault, resulting in
7164 a DoS on applications that verify certificates or CRLs. TLS clients
7165 that verify CRLs are affected. TLS clients and servers with client
7166 authentication enabled may be affected if they use custom verification
7169 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7170 independently by Hanno Böck.
7175 * PKCS7 crash with missing EnvelopedContent
7177 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7178 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7179 with missing content and trigger a NULL pointer dereference on parsing.
7181 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7182 structures from untrusted sources are affected. OpenSSL clients and
7183 servers are not affected.
7185 This issue was reported to OpenSSL by Michal Zalewski (Google).
7190 * CMS verify infinite loop with unknown hash function
7192 When verifying a signedData message the CMS code can enter an infinite loop
7193 if presented with an unknown hash function OID. This can be used to perform
7194 denial of service against any system which verifies signedData messages using
7196 This issue was reported to OpenSSL by Johannes Bauer.
7201 * Race condition handling NewSessionTicket
7203 If a NewSessionTicket is received by a multi-threaded client when attempting to
7204 reuse a previous ticket then a race condition can occur potentially leading to
7205 a double free of the ticket data.
7210 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7212 * Segmentation fault in ASN1_TYPE_cmp fix
7214 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7215 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7216 certificate signature algorithm consistency this can be used to crash any
7217 certificate verification operation and exploited in a DoS attack. Any
7218 application which performs certificate verification is vulnerable including
7219 OpenSSL clients and servers which enable client authentication.
7224 * ASN.1 structure reuse memory corruption fix
7226 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7227 memory corruption via an invalid write. Such reuse is and has been
7228 strongly discouraged and is believed to be rare.
7230 Applications that parse structures containing CHOICE or ANY DEFINED BY
7231 components may be affected. Certificate parsing (d2i_X509 and related
7232 functions) are however not affected. OpenSSL clients and servers are
7238 * PKCS7 NULL pointer dereferences fix
7240 The PKCS#7 parsing code does not handle missing outer ContentInfo
7241 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7242 missing content and trigger a NULL pointer dereference on parsing.
7244 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7245 otherwise parse PKCS#7 structures from untrusted sources are
7246 affected. OpenSSL clients and servers are not affected.
7248 This issue was reported to OpenSSL by Michal Zalewski (Google).
7253 * DoS via reachable assert in SSLv2 servers fix
7255 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7256 servers that both support SSLv2 and enable export cipher suites by sending
7257 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7259 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7260 (OpenSSL development team).
7265 * Use After Free following d2i_ECPrivatekey error fix
7267 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7268 could cause a use after free condition. This, in turn, could cause a double
7269 free in several private key parsing functions (such as d2i_PrivateKey
7270 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7271 for applications that receive EC private keys from untrusted
7272 sources. This scenario is considered rare.
7274 This issue was discovered by the BoringSSL project and fixed in their
7280 * X509_to_X509_REQ NULL pointer deref fix
7282 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7283 the certificate key is invalid. This function is rarely used in practice.
7285 This issue was discovered by Brian Carpenter.
7290 * Removed the export ciphers from the DEFAULT ciphers
7294 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7296 * Build fixes for the Windows and OpenVMS platforms
7298 *Matt Caswell and Richard Levitte*
7300 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7302 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7303 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7304 dereference. This could lead to a Denial Of Service attack. Thanks to
7305 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7310 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7311 dtls1_buffer_record function under certain conditions. In particular this
7312 could occur if an attacker sent repeated DTLS records with the same
7313 sequence number but for the next epoch. The memory leak could be exploited
7314 by an attacker in a Denial of Service attack through memory exhaustion.
7315 Thanks to Chris Mueller for reporting this issue.
7320 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7321 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7322 method would be set to NULL which could later result in a NULL pointer
7323 dereference. Thanks to Frank Schmirler for reporting this issue.
7328 * Abort handshake if server key exchange message is omitted for ephemeral
7331 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7332 reporting this issue.
7337 * Remove non-export ephemeral RSA code on client and server. This code
7338 violated the TLS standard by allowing the use of temporary RSA keys in
7339 non-export ciphersuites and could be used by a server to effectively
7340 downgrade the RSA key length used to a value smaller than the server
7341 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7342 INRIA or reporting this issue.
7347 * Fixed issue where DH client certificates are accepted without verification.
7348 An OpenSSL server will accept a DH certificate for client authentication
7349 without the certificate verify message. This effectively allows a client to
7350 authenticate without the use of a private key. This only affects servers
7351 which trust a client certificate authority which issues certificates
7352 containing DH keys: these are extremely rare and hardly ever encountered.
7353 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7359 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7360 results on some platforms, including x86_64. This bug occurs at random
7361 with a very low probability, and is not known to be exploitable in any
7362 way, though its exact impact is difficult to determine. Thanks to Pieter
7363 Wuille (Blockstream) who reported this issue and also suggested an initial
7364 fix. Further analysis was conducted by the OpenSSL development team and
7365 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7366 the OpenSSL core team.
7371 * Fix various certificate fingerprint issues.
7373 By using non-DER or invalid encodings outside the signed portion of a
7374 certificate the fingerprint can be changed without breaking the signature.
7375 Although no details of the signed portion of the certificate can be changed
7376 this can cause problems with some applications: e.g. those using the
7377 certificate fingerprint for blacklists.
7379 1. Reject signatures with non zero unused bits.
7381 If the BIT STRING containing the signature has non zero unused bits reject
7382 the signature. All current signature algorithms require zero unused bits.
7384 2. Check certificate algorithm consistency.
7386 Check the AlgorithmIdentifier inside TBS matches the one in the
7387 certificate signature. NB: this will result in signature failure
7388 errors for some broken certificates.
7390 Thanks to Konrad Kraszewski from Google for reporting this issue.
7392 3. Check DSA/ECDSA signatures use DER.
7394 Reencode DSA/ECDSA signatures and compare with the original received
7395 signature. Return an error if there is a mismatch.
7397 This will reject various cases including garbage after signature
7398 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7399 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7400 (negative or with leading zeroes).
7402 Further analysis was conducted and fixes were developed by Stephen Henson
7403 of the OpenSSL core team.
7409 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7411 * Session Ticket Memory Leak.
7413 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7414 integrity of that ticket is first verified. In the event of a session
7415 ticket integrity check failing, OpenSSL will fail to free memory
7416 causing a memory leak. By sending a large number of invalid session
7417 tickets an attacker could exploit this issue in a Denial Of Service
7423 * Build option no-ssl3 is incomplete.
7425 When OpenSSL is configured with "no-ssl3" as a build option, servers
7426 could accept and complete a SSL 3.0 handshake, and clients could be
7427 configured to send them.
7430 *Akamai and the OpenSSL team*
7432 * Add support for TLS_FALLBACK_SCSV.
7433 Client applications doing fallback retries should call
7434 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7437 *Adam Langley, Bodo Moeller*
7439 * Add additional DigestInfo checks.
7441 Reencode DigestInto in DER and check against the original when
7442 verifying RSA signature: this will reject any improperly encoded
7443 DigestInfo structures.
7445 Note: this is a precautionary measure and no attacks are currently known.
7449 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7451 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7452 to a denial of service attack. A malicious server can crash the client
7453 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7454 ciphersuite and sending carefully crafted handshake messages.
7456 Thanks to Felix Gröbert (Google) for discovering and researching this
7462 * By sending carefully crafted DTLS packets an attacker could cause openssl
7463 to leak memory. This can be exploited through a Denial of Service attack.
7464 Thanks to Adam Langley for discovering and researching this issue.
7469 * An attacker can force openssl to consume large amounts of memory whilst
7470 processing DTLS handshake messages. This can be exploited through a
7471 Denial of Service attack.
7472 Thanks to Adam Langley for discovering and researching this issue.
7477 * An attacker can force an error condition which causes openssl to crash
7478 whilst processing DTLS packets due to memory being freed twice. This
7479 can be exploited through a Denial of Service attack.
7480 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7486 * If a multithreaded client connects to a malicious server using a resumed
7487 session and the server sends an ec point format extension it could write
7488 up to 255 bytes to freed memory.
7490 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7496 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7497 X509_name_oneline, X509_name_print_ex et al. to leak some information
7498 from the stack. Applications may be affected if they echo pretty printing
7499 output to the attacker.
7501 Thanks to Ivan Fratric (Google) for discovering this issue.
7504 *Emilia Käsper, and Steve Henson*
7506 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7507 for corner cases. (Certain input points at infinity could lead to
7508 bogus results, with non-infinity inputs mapped to infinity too.)
7512 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7514 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7515 handshake can force the use of weak keying material in OpenSSL
7516 SSL/TLS clients and servers.
7518 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7519 researching this issue. ([CVE-2014-0224])
7521 *KIKUCHI Masashi, Steve Henson*
7523 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7524 OpenSSL DTLS client the code can be made to recurse eventually crashing
7527 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7530 *Imre Rad, Steve Henson*
7532 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7533 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7534 client or server. This is potentially exploitable to run arbitrary
7535 code on a vulnerable client or server.
7537 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7539 *Jüri Aedla, Steve Henson*
7541 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7542 are subject to a denial of service attack.
7544 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7545 this issue. ([CVE-2014-3470])
7547 *Felix Gröbert, Ivan Fratric, Steve Henson*
7549 * Harmonize version and its documentation. -f flag is used to display
7552 *mancha <mancha1@zoho.com>*
7554 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7555 in i2d_ECPrivateKey.
7557 *mancha <mancha1@zoho.com>*
7559 * Fix some double frees. These are not thought to be exploitable.
7561 *mancha <mancha1@zoho.com>*
7563 * Fix for the attack described in the paper "Recovering OpenSSL
7564 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7565 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7566 <http://eprint.iacr.org/2014/140>
7568 Thanks to Yuval Yarom and Naomi Benger for discovering this
7569 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7571 *Yuval Yarom and Naomi Benger*
7573 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7575 * Keep original DTLS digest and encryption contexts in retransmission
7576 structures so we can use the previous session parameters if they need
7577 to be resent. ([CVE-2013-6450])
7581 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7582 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7583 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7584 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7585 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7586 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7588 *Rob Stradling, Adam Langley*
7590 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7592 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7594 This addresses the flaw in CBC record processing discovered by
7595 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7596 at: <http://www.isg.rhul.ac.uk/tls/>
7598 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7599 Security Group at Royal Holloway, University of London
7600 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7601 Emilia Käsper for the initial patch.
7604 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7606 * Return an error when checking OCSP signatures when key is NULL.
7607 This fixes a DoS attack. ([CVE-2013-0166])
7611 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7612 the right response is stapled. Also change SSL_get_certificate()
7613 so it returns the certificate actually sent.
7614 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7615 (This is a backport)
7617 *Rob Stradling <rob.stradling@comodo.com>*
7619 * Fix possible deadlock when decoding public keys.
7623 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7625 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7628 * Sanity check record length before skipping explicit IV in DTLS
7631 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7632 fuzzing as a service testing platform.
7637 * Initialise tkeylen properly when encrypting CMS messages.
7638 Thanks to Solar Designer of Openwall for reporting this issue.
7642 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7644 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7645 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7646 in CRYPTO_realloc_clean.
7648 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7649 issue and to Adam Langley <agl@chromium.org> for fixing it.
7652 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7654 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7656 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7657 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7658 content decryption and always return the same error. Note: this attack
7659 needs on average 2^20 messages so it only affects automated senders. The
7660 old behaviour can be re-enabled in the CMS code by setting the
7661 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7662 an MMA defence is not necessary.
7663 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7664 this issue. ([CVE-2012-0884])
7668 * Fix CVE-2011-4619: make sure we really are receiving a
7669 client hello before rejecting multiple SGC restarts. Thanks to
7670 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7674 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7676 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7677 Thanks to Antonio Martin, Enterprise Secure Access Research and
7678 Development, Cisco Systems, Inc. for discovering this bug and
7679 preparing a fix. ([CVE-2012-0050])
7683 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7685 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7686 of the Vaudenay padding oracle attack on CBC mode encryption
7687 which enables an efficient plaintext recovery attack against
7688 the OpenSSL implementation of DTLS. Their attack exploits timing
7689 differences arising during decryption processing. A research
7690 paper describing this attack can be found at:
7691 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7692 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7693 Security Group at Royal Holloway, University of London
7694 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7695 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7696 for preparing the fix. ([CVE-2011-4108])
7698 *Robin Seggelmann, Michael Tuexen*
7700 * Clear bytes used for block padding of SSL 3.0 records.
7703 *Adam Langley (Google)*
7705 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7706 Kadianakis <desnacked@gmail.com> for discovering this issue and
7707 Adam Langley for preparing the fix. ([CVE-2011-4619])
7709 *Adam Langley (Google)*
7711 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7713 *Andrey Kulikov <amdeich@gmail.com>*
7715 * Prevent malformed RFC3779 data triggering an assertion failure.
7716 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7717 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7719 *Rob Austein <sra@hactrn.net>*
7721 * Improved PRNG seeding for VOS.
7723 *Paul Green <Paul.Green@stratus.com>*
7725 * Fix ssl_ciph.c set-up race.
7727 *Adam Langley (Google)*
7729 * Fix spurious failures in ecdsatest.c.
7731 *Emilia Käsper (Google)*
7733 * Fix the BIO_f_buffer() implementation (which was mixing different
7734 interpretations of the `..._len` fields).
7736 *Adam Langley (Google)*
7738 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7739 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7740 threads won't reuse the same blinding coefficients.
7742 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7743 lock to call BN_BLINDING_invert_ex, and avoids one use of
7744 BN_BLINDING_update for each BN_BLINDING structure (previously,
7745 the last update always remained unused).
7747 *Emilia Käsper (Google)*
7749 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7751 *Bob Buckholz (Google)*
7753 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7755 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7756 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7758 *Kaspar Brand <ossl@velox.ch>*
7760 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7761 for multi-threaded use of ECDH. ([CVE-2011-3210])
7763 *Adam Langley (Google)*
7765 * Fix x509_name_ex_d2i memory leak on bad inputs.
7769 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7770 signature public key algorithm by using OID xref utilities instead.
7771 Before this you could only use some ECC ciphersuites with SHA1 only.
7775 * Add protection against ECDSA timing attacks as mentioned in the paper
7776 by Billy Bob Brumley and Nicola Tuveri, see:
7777 <http://eprint.iacr.org/2011/232.pdf>
7779 *Billy Bob Brumley and Nicola Tuveri*
7781 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7783 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7785 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7787 * Fix bug in string printing code: if *any* escaping is enabled we must
7788 escape the escape character (backslash) or the resulting string is
7793 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7795 * Disable code workaround for ancient and obsolete Netscape browsers
7796 and servers: an attacker can use it in a ciphersuite downgrade attack.
7797 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7801 * Fixed J-PAKE implementation error, originally discovered by
7802 Sebastien Martini, further info and confirmation from Stefan
7803 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7807 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7809 * Fix extension code to avoid race conditions which can result in a buffer
7810 overrun vulnerability: resumed sessions must not be modified as they can
7811 be shared by multiple threads. CVE-2010-3864
7815 * Fix WIN32 build system to correctly link an ENGINE directory into
7820 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7822 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7825 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7827 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7829 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7830 context. The operation can be customised via the ctrl mechanism in
7831 case ENGINEs want to include additional functionality.
7835 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7839 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7840 output hashes compatible with older versions of OpenSSL.
7842 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7844 * Fix compression algorithm handling: if resuming a session use the
7845 compression algorithm of the resumed session instead of determining
7846 it from client hello again. Don't allow server to change algorithm.
7850 * Add load_crls() function to apps tidying load_certs() too. Add option
7851 to verify utility to allow additional CRLs to be included.
7855 * Update OCSP request code to permit adding custom headers to the request:
7856 some responders need this.
7860 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7863 *Julia Lawall <julia@diku.dk>*
7865 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7866 needlessly dereferenced structures, used obsolete functions and
7867 didn't handle all updated verify codes correctly.
7871 * Disable MD2 in the default configuration.
7875 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7876 indicate the initial BIO being pushed or popped. This makes it possible
7877 to determine whether the BIO is the one explicitly called or as a result
7878 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7879 it handles reference counts correctly and doesn't zero out the I/O bio
7880 when it is not being explicitly popped. WARNING: applications which
7881 included workarounds for the old buggy behaviour will need to be modified
7882 or they could free up already freed BIOs.
7886 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7887 renaming to all platforms (within the 0.9.8 branch, this was
7888 done conditionally on Netware platforms to avoid a name clash).
7890 *Guenter <lists@gknw.net>*
7892 * Add ECDHE and PSK support to DTLS.
7894 *Michael Tuexen <tuexen@fh-muenster.de>*
7896 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7901 * Add "missing" function EVP_MD_flags() (without this the only way to
7902 retrieve a digest flags is by accessing the structure directly. Update
7903 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7904 or cipher is registered as in the "from" argument. Print out all
7905 registered digests in the dgst usage message instead of manually
7906 attempting to work them out.
7910 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7911 this allows the use of compression and extensions. Change default cipher
7912 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7913 by default unless an application cipher string requests it.
7917 * Alter match criteria in PKCS12_parse(). It used to try to use local
7918 key ids to find matching certificates and keys but some PKCS#12 files
7919 don't follow the (somewhat unwritten) rules and this strategy fails.
7920 Now just gather all certificates together and the first private key
7921 then look for the first certificate that matches the key.
7925 * Support use of registered digest and cipher names for dgst and cipher
7926 commands instead of having to add each one as a special case. So now
7933 openssl dgst -sha256 foo
7935 and this works for ENGINE based algorithms too.
7939 * Update Gost ENGINE to support parameter files.
7941 *Victor B. Wagner <vitus@cryptocom.ru>*
7943 * Support GeneralizedTime in ca utility.
7945 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7947 * Enhance the hash format used for certificate directory links. The new
7948 form uses the canonical encoding (meaning equivalent names will work
7949 even if they aren't identical) and uses SHA1 instead of MD5. This form
7950 is incompatible with the older format and as a result c_rehash should
7951 be used to rebuild symbolic links.
7955 * Make PKCS#8 the default write format for private keys, replacing the
7956 traditional format. This form is standardised, more secure and doesn't
7957 include an implicit MD5 dependency.
7961 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7962 committed to OpenSSL should pass this lot as a minimum.
7966 * Add session ticket override functionality for use by EAP-FAST.
7968 *Jouni Malinen <j@w1.fi>*
7970 * Modify HMAC functions to return a value. Since these can be implemented
7971 in an ENGINE errors can occur.
7975 * Type-checked OBJ_bsearch_ex.
7979 * Type-checked OBJ_bsearch. Also some constification necessitated
7980 by type-checking. Still to come: TXT_DB, bsearch(?),
7981 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7986 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7987 seconds to a tm structure directly, instead of going through OS
7988 specific date routines. This avoids any issues with OS routines such
7989 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7990 and X509_time_adj_ex() to cover the extended range. The existing
7991 X509_time_adj() is still usable and will no longer have any date issues.
7995 * Delta CRL support. New use deltas option which will attempt to locate
7996 and search any appropriate delta CRLs available.
7998 This work was sponsored by Google.
8002 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8003 code and add additional score elements. Validate alternate CRL paths
8004 as part of the CRL checking and indicate a new error "CRL path validation
8005 error" in this case. Applications wanting additional details can use
8006 the verify callback and check the new "parent" field. If this is not
8007 NULL CRL path validation is taking place. Existing applications won't
8008 see this because it requires extended CRL support which is off by
8011 This work was sponsored by Google.
8015 * Support for freshest CRL extension.
8017 This work was sponsored by Google.
8021 * Initial indirect CRL support. Currently only supported in the CRLs
8022 passed directly and not via lookup. Process certificate issuer
8023 CRL entry extension and lookup CRL entries by bother issuer name
8024 and serial number. Check and process CRL issuer entry in IDP extension.
8026 This work was sponsored by Google.
8030 * Add support for distinct certificate and CRL paths. The CRL issuer
8031 certificate is validated separately in this case. Only enabled if
8032 an extended CRL support flag is set: this flag will enable additional
8033 CRL functionality in future.
8035 This work was sponsored by Google.
8039 * Add support for policy mappings extension.
8041 This work was sponsored by Google.
8045 * Fixes to pathlength constraint, self issued certificate handling,
8046 policy processing to align with RFC3280 and PKITS tests.
8048 This work was sponsored by Google.
8052 * Support for name constraints certificate extension. DN, email, DNS
8053 and URI types are currently supported.
8055 This work was sponsored by Google.
8059 * To cater for systems that provide a pointer-based thread ID rather
8060 than numeric, deprecate the current numeric thread ID mechanism and
8061 replace it with a structure and associated callback type. This
8062 mechanism allows a numeric "hash" to be extracted from a thread ID in
8063 either case, and on platforms where pointers are larger than 'long',
8064 mixing is done to help ensure the numeric 'hash' is usable even if it
8065 can't be guaranteed unique. The default mechanism is to use "&errno"
8066 as a pointer-based thread ID to distinguish between threads.
8068 Applications that want to provide their own thread IDs should now use
8069 CRYPTO_THREADID_set_callback() to register a callback that will call
8070 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8072 Note that ERR_remove_state() is now deprecated, because it is tied
8073 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8074 to free the current thread's error state should be replaced by
8075 ERR_remove_thread_state(NULL).
8077 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8078 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8079 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8080 application was previously providing a numeric thread callback that
8081 was inappropriate for distinguishing threads, then uniqueness might
8082 have been obtained with &errno that happened immediately in the
8083 intermediate development versions of OpenSSL; this is no longer the
8084 case, the numeric thread callback will now override the automatic use
8087 *Geoff Thorpe, with help from Bodo Moeller*
8089 * Initial support for different CRL issuing certificates. This covers a
8090 simple case where the self issued certificates in the chain exist and
8091 the real CRL issuer is higher in the existing chain.
8093 This work was sponsored by Google.
8097 * Removed effectively defunct crypto/store from the build.
8101 * Revamp of STACK to provide stronger type-checking. Still to come:
8102 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8103 ASN1_STRING, CONF_VALUE.
8107 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8108 RAM on SSL connections. This option can save about 34k per idle SSL.
8112 * Revamp of LHASH to provide stronger type-checking. Still to come:
8113 STACK, TXT_DB, bsearch, qsort.
8117 * Initial support for Cryptographic Message Syntax (aka CMS) based
8118 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8119 support for data, signedData, compressedData, digestedData and
8120 encryptedData, envelopedData types included. Scripts to check against
8121 RFC4134 examples draft and interop and consistency checks of many
8122 content types and variants.
8126 * Add options to enc utility to support use of zlib compression BIO.
8130 * Extend mk1mf to support importing of options and assembly language
8131 files from Configure script, currently only included in VC-WIN32.
8132 The assembly language rules can now optionally generate the source
8133 files from the associated perl scripts.
8137 * Implement remaining functionality needed to support GOST ciphersuites.
8138 Interop testing has been performed using CryptoPro implementations.
8140 *Victor B. Wagner <vitus@cryptocom.ru>*
8142 * s390x assembler pack.
8146 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8151 * Implement Opaque PRF Input TLS extension as specified in
8152 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8153 official specification yet and no extension type assignment by
8154 IANA exists, this extension (for now) will have to be explicitly
8155 enabled when building OpenSSL by providing the extension number
8156 to use. For example, specify an option
8158 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8160 to the "config" or "Configure" script to enable the extension,
8161 assuming extension number 0x9527 (which is a completely arbitrary
8162 and unofficial assignment based on the MD5 hash of the Internet
8163 Draft). Note that by doing so, you potentially lose
8164 interoperability with other TLS implementations since these might
8165 be using the same extension number for other purposes.
8167 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8168 opaque PRF input value to use in the handshake. This will create
8169 an internal copy of the length-'len' string at 'src', and will
8170 return non-zero for success.
8172 To get more control and flexibility, provide a callback function
8175 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8176 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8180 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8183 Callback function 'cb' will be called in handshakes, and is
8184 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8185 Argument 'arg' is for application purposes (the value as given to
8186 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8187 be provided to the callback function). The callback function
8188 has to return non-zero to report success: usually 1 to use opaque
8189 PRF input just if possible, or 2 to enforce use of the opaque PRF
8190 input. In the latter case, the library will abort the handshake
8191 if opaque PRF input is not successfully negotiated.
8193 Arguments 'peerinput' and 'len' given to the callback function
8194 will always be NULL and 0 in the case of a client. A server will
8195 see the client's opaque PRF input through these variables if
8196 available (NULL and 0 otherwise). Note that if the server
8197 provides an opaque PRF input, the length must be the same as the
8198 length of the client's opaque PRF input.
8200 Note that the callback function will only be called when creating
8201 a new session (session resumption can resume whatever was
8202 previously negotiated), and will not be called in SSL 2.0
8203 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8204 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8205 for applications that need to enforce opaque PRF input.
8209 * Update ssl code to support digests other than SHA1+MD5 for handshake
8212 *Victor B. Wagner <vitus@cryptocom.ru>*
8214 * Add RFC4507 support to OpenSSL. This includes the corrections in
8215 RFC4507bis. The encrypted ticket format is an encrypted encoded
8216 SSL_SESSION structure, that way new session features are automatically
8219 If a client application caches session in an SSL_SESSION structure
8220 support is transparent because tickets are now stored in the encoded
8223 The SSL_CTX structure automatically generates keys for ticket
8224 protection in servers so again support should be possible
8225 with no application modification.
8227 If a client or server wishes to disable RFC4507 support then the option
8228 SSL_OP_NO_TICKET can be set.
8230 Add a TLS extension debugging callback to allow the contents of any client
8231 or server extensions to be examined.
8233 This work was sponsored by Google.
8237 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8238 OpenSSL should now compile cleanly on gcc 4.2
8240 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8242 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8243 support including streaming MAC support: this is required for GOST
8244 ciphersuite support.
8246 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8248 * Add option -stream to use PKCS#7 streaming in smime utility. New
8249 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8250 to output in BER and PEM format.
8254 * Experimental support for use of HMAC via EVP_PKEY interface. This
8255 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8256 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8257 ENGINE support for HMAC keys which are unextractable. New -mac and
8258 -macopt options to dgst utility.
8262 * New option -sigopt to dgst utility. Update dgst to use
8263 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8264 alternative signing parameters such as X9.31 or PSS in the dgst
8269 * Change ssl_cipher_apply_rule(), the internal function that does
8270 the work each time a ciphersuite string requests enabling
8271 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8272 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8273 the order of disabled ciphersuites such that those ciphersuites
8274 that most recently went from enabled to disabled not only stay
8275 in order with respect to each other, but also have higher priority
8276 than other disabled ciphersuites the next time ciphersuites are
8279 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8280 the same ciphersuites as with "HIGH" alone, but in a specific
8281 order where the PSK ciphersuites come first (since they are the
8282 most recently disabled ciphersuites when "HIGH" is parsed).
8284 Also, change ssl_create_cipher_list() (using this new
8285 functionality) such that between otherwise identical
8286 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8291 * Change ssl_create_cipher_list() so that it automatically
8292 arranges the ciphersuites in reasonable order before starting
8293 to process the rule string. Thus, the definition for "DEFAULT"
8294 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8295 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8296 This makes it much easier to arrive at a reasonable default order
8297 in applications for which anonymous ciphers are OK (meaning
8298 that you can't actually use DEFAULT).
8300 *Bodo Moeller; suggested by Victor Duchovni*
8302 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8303 processing) into multiple integers instead of setting
8304 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8305 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8306 (These masks as well as the individual bit definitions are hidden
8307 away into the non-exported interface ssl/ssl_locl.h, so this
8308 change to the definition of the SSL_CIPHER structure shouldn't
8309 affect applications.) This give us more bits for each of these
8310 categories, so there is no longer a need to coagulate AES128 and
8311 AES256 into a single algorithm bit, and to coagulate Camellia128
8312 and Camellia256 into a single algorithm bit, which has led to all
8315 Thus, among other things, the kludge introduced in 0.9.7m and
8316 0.9.8e for masking out AES256 independently of AES128 or masking
8317 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8319 With the change, we also introduce new ciphersuite aliases that
8320 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8325 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8326 Use the leftmost N bytes of the signature input if the input is
8327 larger than the prime q (with N being the size in bytes of q).
8331 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8332 it yet and it is largely untested.
8336 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8340 * Initial incomplete changes to avoid need for function casts in OpenSSL
8341 some compilers (gcc 4.2 and later) reject their use. Safestack is
8342 reimplemented. Update ASN1 to avoid use of legacy functions.
8346 * Win32/64 targets are linked with Winsock2.
8350 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8351 to external functions. This can be used to increase CRL handling
8352 efficiency especially when CRLs are very large by (for example) storing
8353 the CRL revoked certificates in a database.
8357 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8358 new CRLs added to a directory can be used. New command line option
8359 -verify_return_error to s_client and s_server. This causes real errors
8360 to be returned by the verify callback instead of carrying on no matter
8361 what. This reflects the way a "real world" verify callback would behave.
8365 * GOST engine, supporting several GOST algorithms and public key formats.
8366 Kindly donated by Cryptocom.
8370 * Partial support for Issuing Distribution Point CRL extension. CRLs
8371 partitioned by DP are handled but no indirect CRL or reason partitioning
8372 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8373 selected via a scoring technique which handles IDP and AKID in CRLs.
8377 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8378 will ultimately be used for all verify operations: this will remove the
8379 X509_STORE dependency on certificate verification and allow alternative
8380 lookup methods. X509_STORE based implementations of these two callbacks.
8384 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8385 Modify get_crl() to find a valid (unexpired) CRL if possible.
8389 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8390 this would be called X509_CRL_cmp() but that name is already used by
8391 a function that just compares CRL issuer names. Cache several CRL
8392 extensions in X509_CRL structure and cache CRLDP in X509.
8396 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8397 this maps equivalent X509_NAME structures into a consistent structure.
8398 Name comparison can then be performed rapidly using memcmp().
8402 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8407 * Allow digests to supply their own micalg string for S/MIME type using
8408 the ctrl EVP_MD_CTRL_MICALG.
8412 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8413 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8414 ctrl. It can then customise the structure before and/or after signing
8419 * New function OBJ_add_sigid() to allow application defined signature OIDs
8420 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8421 to free up any added signature OIDs.
8425 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8426 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8427 digest and cipher tables. New options added to openssl utility:
8428 list-message-digest-algorithms and list-cipher-algorithms.
8432 * Change the array representation of binary polynomials: the list
8433 of degrees of non-zero coefficients is now terminated with -1.
8434 Previously it was terminated with 0, which was also part of the
8435 value; thus, the array representation was not applicable to
8436 polynomials where t^0 has coefficient zero. This change makes
8437 the array representation useful in a more general context.
8441 * Various modifications and fixes to SSL/TLS cipher string
8442 handling. For ECC, the code now distinguishes between fixed ECDH
8443 with RSA certificates on the one hand and with ECDSA certificates
8444 on the other hand, since these are separate ciphersuites. The
8445 unused code for Fortezza ciphersuites has been removed.
8447 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8448 (not "ECDHE"). For consistency with the code for DH
8449 certificates, use of ECDH certificates is now considered ECDH
8450 authentication, not RSA or ECDSA authentication (the latter is
8451 merely the CA's signing algorithm and not actively used in the
8454 The temporary ciphersuite alias "ECCdraft" is no longer
8455 available, and ECC ciphersuites are no longer excluded from "ALL"
8456 and "DEFAULT". The following aliases now exist for RFC 4492
8457 ciphersuites, most of these by analogy with the DH case:
8459 kECDHr - ECDH cert, signed with RSA
8460 kECDHe - ECDH cert, signed with ECDSA
8461 kECDH - ECDH cert (signed with either RSA or ECDSA)
8462 kEECDH - ephemeral ECDH
8463 ECDH - ECDH cert or ephemeral ECDH
8469 AECDH - anonymous ECDH
8470 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8474 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8475 Use correct micalg parameters depending on digest(s) in signed message.
8479 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8480 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8484 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8485 an engine to register a method. Add ENGINE lookups for methods and
8486 functional reference processing.
8490 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8491 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8496 * New -resign option to smime utility. This adds one or more signers
8497 to an existing PKCS#7 signedData structure. Also -md option to use an
8498 alternative message digest algorithm for signing.
8502 * Tidy up PKCS#7 routines and add new functions to make it easier to
8503 create PKCS7 structures containing multiple signers. Update smime
8504 application to support multiple signers.
8508 * New -macalg option to pkcs12 utility to allow setting of an alternative
8513 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8514 Reorganize PBE internals to lookup from a static table using NIDs,
8515 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8516 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8517 PRF which will be automatically used with PBES2.
8521 * Replace the algorithm specific calls to generate keys in "req" with the
8526 * Update PKCS#7 enveloped data routines to use new API. This is now
8527 supported by any public key method supporting the encrypt operation. A
8528 ctrl is added to allow the public key algorithm to examine or modify
8529 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8534 * Add a ctrl to asn1 method to allow a public key algorithm to express
8535 a default digest type to use. In most cases this will be SHA1 but some
8536 algorithms (such as GOST) need to specify an alternative digest. The
8537 return value indicates how strong the preference is 1 means optional and
8538 2 is mandatory (that is it is the only supported type). Modify
8539 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8540 use the default md. Update openssl utilities to use the default digest
8541 type for signing if it is not explicitly indicated.
8545 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8546 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8547 signing method from the key type. This effectively removes the link
8548 between digests and public key types.
8552 * Add an OID cross reference table and utility functions. Its purpose is to
8553 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8554 rsaEncryption. This will allow some of the algorithm specific hackery
8555 needed to use the correct OID to be removed.
8559 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8560 structures for PKCS7_sign(). They are now set up by the relevant public
8565 * Add provisional EC pkey method with support for ECDSA and ECDH.
8569 * Add support for key derivation (agreement) in the API, DH method and
8574 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8575 public and private key formats. As a side effect these add additional
8576 command line functionality not previously available: DSA signatures can be
8577 generated and verified using pkeyutl and DH key support and generation in
8584 *Oliver Tappe <zooey@hirschkaefer.de>*
8586 * New make target "install_html_docs" installs HTML renditions of the
8589 *Oliver Tappe <zooey@hirschkaefer.de>*
8591 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8592 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8593 support key and parameter generation and add initial key generation
8594 functionality for RSA.
8598 * Add functions for main EVP_PKEY_method operations. The undocumented
8599 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8600 `EVP_PKEY_{encrypt,decrypt}_old`.
8604 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8605 key API, doesn't do much yet.
8609 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8610 public key algorithms. New option to openssl utility:
8611 "list-public-key-algorithms" to print out info.
8615 * Implement the Supported Elliptic Curves Extension for
8616 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8620 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8621 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8625 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8626 utilities such as rsa, dsa, dsaparam etc except they process any key
8631 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8632 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8633 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8638 * Initial support for pluggable public key ASN1.
8639 De-spaghettify the public key ASN1 handling. Move public and private
8640 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8641 algorithm specific handling to a single module within the relevant
8642 algorithm directory. Add functions to allow (near) opaque processing
8643 of public and private key structures.
8647 * Implement the Supported Point Formats Extension for
8648 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8652 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8653 for the psk identity [hint] and the psk callback functions to the
8654 SSL_SESSION, SSL and SSL_CTX structure.
8657 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8661 SSL_CTX_use_psk_identity_hint
8662 SSL_get_psk_identity_hint
8663 SSL_get_psk_identity
8664 SSL_use_psk_identity_hint
8666 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8668 * Add RFC 3161 compliant time stamp request creation, response generation
8669 and response verification functionality.
8671 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8673 * Add initial support for TLS extensions, specifically for the server_name
8674 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8675 have new members for a host name. The SSL data structure has an
8676 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8677 stored in that context to allow for session resumption, even after the
8678 SSL has been switched to a new SSL_CTX in reaction to a client's
8679 server_name extension.
8681 New functions (subject to change):
8683 SSL_get_servername()
8684 SSL_get_servername_type()
8687 New CTRL codes and macros (subject to change):
8689 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8690 - SSL_CTX_set_tlsext_servername_callback()
8691 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8692 - SSL_CTX_set_tlsext_servername_arg()
8693 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8695 openssl s_client has a new '-servername ...' option.
8697 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8698 '-key2 ...', '-servername_fatal' (subject to change). This allows
8699 testing the HostName extension for a specific single host name ('-cert'
8700 and '-key' remain fallbacks for handshakes without HostName
8701 negotiation). If the unrecognized_name alert has to be sent, this by
8702 default is a warning; it becomes fatal with the '-servername_fatal'
8705 *Peter Sylvester, Remy Allais, Christophe Renou*
8707 * Whirlpool hash implementation is added.
8711 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8712 bn(64,32). Because of instruction set limitations it doesn't have
8713 any negative impact on performance. This was done mostly in order
8714 to make it possible to share assembler modules, such as bn_mul_mont
8715 implementations, between 32- and 64-bit builds without hassle.
8719 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8720 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8725 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8726 dedicated Montgomery multiplication procedure, is introduced.
8727 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8728 "64-bit" performance on certain 32-bit targets.
8732 * New option SSL_OP_NO_COMP to disable use of compression selectively
8733 in SSL structures. New SSL ctrl to set maximum send fragment size.
8734 Save memory by setting the I/O buffer sizes dynamically instead of
8735 using the maximum available value.
8739 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8740 in addition to the text details.
8744 * Very, very preliminary EXPERIMENTAL support for printing of general
8745 ASN1 structures. This currently produces rather ugly output and doesn't
8746 handle several customised structures at all.
8750 * Integrated support for PVK file format and some related formats such
8751 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8752 these in the 'rsa' and 'dsa' utilities.
8756 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8760 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8761 place for the (very old) "NETSCAPE" format certificates which are now
8762 handled using new ASN1 code equivalents.
8766 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8767 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8768 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8772 * Modify CRL distribution points extension code to print out previously
8773 unsupported fields. Enhance extension setting code to allow setting of
8778 * Add print and set support for Issuing Distribution Point CRL extension.
8782 * Change 'Configure' script to enable Camellia by default.
8789 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8791 * When rejecting SSL/TLS records due to an incorrect version number, never
8792 update s->server with a new major version number. As of
8793 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8794 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8795 the previous behavior could result in a read attempt at NULL when
8796 receiving specific incorrect SSL/TLS records once record payload
8797 protection is active. ([CVE-2010-0740])
8799 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8801 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8802 could be crashed if the relevant tables were not present (e.g. chrooted).
8804 *Tomas Hoger <thoger@redhat.com>*
8806 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8808 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8810 *Martin Olsson, Neel Mehta*
8812 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8813 accommodate for stack sorting, always a write lock!).
8817 * On some versions of WIN32 Heap32Next is very slow. This can cause
8818 excessive delays in the RAND_poll(): over a minute. As a workaround
8819 include a time check in the inner Heap32Next loop too.
8823 * The code that handled flushing of data in SSL/TLS originally used the
8824 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8825 the problem outlined in PR#1949. The fix suggested there however can
8826 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8827 of Apache). So instead simplify the code to flush unconditionally.
8828 This should be fine since flushing with no data to flush is a no op.
8832 * Handle TLS versions 2.0 and later properly and correctly use the
8833 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8834 off ancient servers have a habit of sticking around for a while...
8838 * Modify compression code so it frees up structures without using the
8839 ex_data callbacks. This works around a problem where some applications
8840 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8841 restarting) then use compression (e.g. SSL with compression) later.
8842 This results in significant per-connection memory leaks and
8843 has caused some security issues including CVE-2008-1678 and
8848 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8849 change when encrypting or decrypting.
8853 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8854 connect and renegotiate with servers which do not support RI.
8855 Until RI is more widely deployed this option is enabled by default.
8859 * Add "missing" ssl ctrls to clear options and mode.
8863 * If client attempts to renegotiate and doesn't support RI respond with
8864 a no_renegotiation alert as required by RFC5746. Some renegotiating
8865 TLS clients will continue a connection gracefully when they receive
8866 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8867 waiting for a server hello which it will never receive. Now we treat a
8868 received no_renegotiation alert as a fatal error. This is because
8869 applications requesting a renegotiation might well expect it to succeed
8870 and would have no code in place to handle the server denying it so the
8871 only safe thing to do is to terminate the connection.
8875 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8876 peer supports secure renegotiation and 0 otherwise. Print out peer
8877 renegotiation support in s_client/s_server.
8881 * Replace the highly broken and deprecated SPKAC certification method with
8882 the updated NID creation version. This should correctly handle UTF8.
8886 * Implement RFC5746. Re-enable renegotiation but require the extension
8887 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8888 turns out to be a bad idea. It has been replaced by
8889 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8890 SSL_CTX_set_options(). This is really not recommended unless you
8891 know what you are doing.
8893 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8895 * Fixes to stateless session resumption handling. Use initial_ctx when
8896 issuing and attempting to decrypt tickets in case it has changed during
8897 servername handling. Use a non-zero length session ID when attempting
8898 stateless session resumption: this makes it possible to determine if
8899 a resumption has occurred immediately after receiving server hello
8900 (several places in OpenSSL subtly assume this) instead of later in
8905 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8906 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8907 fixes for a few places where the return code is not checked
8910 *Julia Lawall <julia@diku.dk>*
8912 * Add --strict-warnings option to Configure script to include devteam
8913 warnings in other configurations.
8917 * Add support for --libdir option and LIBDIR variable in makefiles. This
8918 makes it possible to install openssl libraries in locations which
8919 have names other than "lib", for example "/usr/lib64" which some
8922 *Steve Henson, based on patch from Jeremy Utley*
8924 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8925 X690 8.9.12 and can produce some misleading textual output of OIDs.
8927 *Steve Henson, reported by Dan Kaminsky*
8929 * Delete MD2 from algorithm tables. This follows the recommendation in
8930 several standards that it is not used in new applications due to
8931 several cryptographic weaknesses. For binary compatibility reasons
8932 the MD2 API is still compiled in by default.
8936 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8941 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8942 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8945 *Guenter <lists@gknw.net>*
8947 * Fix the server certificate chain building code to use X509_verify_cert(),
8948 it used to have an ad-hoc builder which was unable to cope with anything
8949 other than a simple chain.
8951 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8953 * Don't check self signed certificate signatures in X509_verify_cert()
8954 by default (a flag can override this): it just wastes time without
8955 adding any security. As a useful side effect self signed root CAs
8956 with non-FIPS digests are now usable in FIPS mode.
8960 * In dtls1_process_out_of_seq_message() the check if the current message
8961 is already buffered was missing. For every new message was memory
8962 allocated, allowing an attacker to perform an denial of service attack
8963 with sending out of seq handshake messages until there is no memory
8964 left. Additionally every future message was buffered, even if the
8965 sequence number made no sense and would be part of another handshake.
8966 So only messages with sequence numbers less than 10 in advance will be
8967 buffered. ([CVE-2009-1378])
8969 *Robin Seggelmann, discovered by Daniel Mentz*
8971 * Records are buffered if they arrive with a future epoch to be
8972 processed after finishing the corresponding handshake. There is
8973 currently no limitation to this buffer allowing an attacker to perform
8974 a DOS attack with sending records with future epochs until there is no
8975 memory left. This patch adds the pqueue_size() function to determine
8976 the size of a buffer and limits the record buffer to 100 entries.
8979 *Robin Seggelmann, discovered by Daniel Mentz*
8981 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8982 parent structure is freed. ([CVE-2009-1379])
8986 * Handle non-blocking I/O properly in SSL_shutdown() call.
8988 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8990 * Add `2.5.4.*` OIDs
8992 *Ilya O. <vrghost@gmail.com>*
8994 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8996 * Disable renegotiation completely - this fixes a severe security
8997 problem ([CVE-2009-3555]) at the cost of breaking all
8998 renegotiation. Renegotiation can be re-enabled by setting
8999 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9000 run-time. This is really not recommended unless you know what
9005 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9007 * Don't set val to NULL when freeing up structures, it is freed up by
9008 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9009 zeroing past the valid field. ([CVE-2009-0789])
9011 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9013 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9014 checked correctly. This would allow some invalid signed attributes to
9015 appear to verify correctly. ([CVE-2009-0591])
9017 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9019 * Reject UniversalString and BMPString types with invalid lengths. This
9020 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9021 a legal length. ([CVE-2009-0590])
9025 * Set S/MIME signing as the default purpose rather than setting it
9026 unconditionally. This allows applications to override it at the store
9031 * Permit restricted recursion of ASN1 strings. This is needed in practice
9032 to handle some structures.
9036 * Improve efficiency of mem_gets: don't search whole buffer each time
9039 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9041 * New -hex option for openssl rand.
9045 * Print out UTF8String and NumericString when parsing ASN1.
9049 * Support NumericString type for name components.
9053 * Allow CC in the environment to override the automatically chosen
9054 compiler. Note that nothing is done to ensure flags work with the
9059 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9061 * Properly check EVP_VerifyFinal() and similar return values
9064 *Ben Laurie, Bodo Moeller, Google Security Team*
9066 * Enable TLS extensions by default.
9070 * Allow the CHIL engine to be loaded, whether the application is
9071 multithreaded or not. (This does not release the developer from the
9072 obligation to set up the dynamic locking callbacks.)
9074 *Sander Temme <sander@temme.net>*
9076 * Use correct exit code if there is an error in dgst command.
9078 *Steve Henson; problem pointed out by Roland Dirlewanger*
9080 * Tweak Configure so that you need to say "experimental-jpake" to enable
9081 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9085 * Add experimental JPAKE support, including demo authentication in
9086 s_client and s_server.
9090 * Set the comparison function in v3_addr_canonize().
9092 *Rob Austein <sra@hactrn.net>*
9094 * Add support for XMPP STARTTLS in s_client.
9096 *Philip Paeps <philip@freebsd.org>*
9098 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9099 to ensure that even with this option, only ciphersuites in the
9100 server's preference list will be accepted. (Note that the option
9101 applies only when resuming a session, so the earlier behavior was
9102 just about the algorithm choice for symmetric cryptography.)
9106 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9108 * Fix NULL pointer dereference if a DTLS server received
9109 ChangeCipherSpec as first record ([CVE-2009-1386]).
9113 * Fix a state transition in s3_srvr.c and d1_srvr.c
9114 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9118 * The fix in 0.9.8c that supposedly got rid of unsafe
9119 double-checked locking was incomplete for RSA blinding,
9120 addressing just one layer of what turns out to have been
9121 doubly unsafe triple-checked locking.
9123 So now fix this for real by retiring the MONT_HELPER macro
9124 in crypto/rsa/rsa_eay.c.
9126 *Bodo Moeller; problem pointed out by Marius Schilder*
9128 * Various precautionary measures:
9130 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9132 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9133 (NB: This would require knowledge of the secret session ticket key
9134 to exploit, in which case you'd be SOL either way.)
9136 - Change bn_nist.c so that it will properly handle input BIGNUMs
9137 outside the expected range.
9139 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9142 *Neel Mehta, Bodo Moeller*
9144 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9145 the load fails. Useful for distros.
9147 *Ben Laurie and the FreeBSD team*
9149 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9153 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9157 * Expand ENGINE to support engine supplied SSL client certificate functions.
9159 This work was sponsored by Logica.
9163 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9164 keystores. Support for SSL/TLS client authentication too.
9165 Not compiled unless enable-capieng specified to Configure.
9167 This work was sponsored by Logica.
9171 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9172 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9173 attribute creation routines such as certificate requests and PKCS#12
9178 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9180 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9181 handshake which could lead to a client crash as found using the
9182 Codenomicon TLS test suite ([CVE-2008-1672])
9184 *Steve Henson, Mark Cox*
9186 * Fix double free in TLS server name extensions which could lead to
9187 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9191 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9193 Clear the error queue to ensure that error entries left from
9194 older function calls do not interfere with the correct operation.
9196 *Lutz Jaenicke, Erik de Castro Lopo*
9198 * Remove root CA certificates of commercial CAs:
9200 The OpenSSL project does not recommend any specific CA and does not
9201 have any policy with respect to including or excluding any CA.
9202 Therefore it does not make any sense to ship an arbitrary selection
9203 of root CA certificates with the OpenSSL software.
9207 * RSA OAEP patches to fix two separate invalid memory reads.
9208 The first one involves inputs when 'lzero' is greater than
9209 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9210 before the beginning of from). The second one involves inputs where
9211 the 'db' section contains nothing but zeroes (there is a one-byte
9212 invalid read after the end of 'db').
9214 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9216 * Partial backport from 0.9.9-dev:
9218 Introduce bn_mul_mont (dedicated Montgomery multiplication
9219 procedure) as a candidate for BIGNUM assembler implementation.
9220 While 0.9.9-dev uses assembler for various architectures, only
9221 x86_64 is available by default here in the 0.9.8 branch, and
9222 32-bit x86 is available through a compile-time setting.
9224 To try the 32-bit x86 assembler implementation, use Configure
9225 option "enable-montasm" (which exists only for this backport).
9227 As "enable-montasm" for 32-bit x86 disclaims code stability
9228 anyway, in this constellation we activate additional code
9229 backported from 0.9.9-dev for further performance improvements,
9230 namely BN_from_montgomery_word. (To enable this otherwise,
9231 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9233 *Andy Polyakov (backport partially by Bodo Moeller)*
9235 * Add TLS session ticket callback. This allows an application to set
9236 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9237 values. This is useful for key rollover for example where several key
9238 sets may exist with different names.
9242 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9243 This was broken until now in 0.9.8 releases, such that the only way
9244 a registered ENGINE could be used (assuming it initialises
9245 successfully on the host) was to explicitly set it as the default
9246 for the relevant algorithms. This is in contradiction with 0.9.7
9247 behaviour and the documentation. With this fix, when an ENGINE is
9248 registered into a given algorithm's table of implementations, the
9249 'uptodate' flag is reset so that auto-discovery will be used next
9250 time a new context for that algorithm attempts to select an
9253 *Ian Lister (tweaked by Geoff Thorpe)*
9255 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9256 implementation in the following ways:
9258 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9261 Lack of BER streaming support means one pass streaming processing is
9262 only supported if data is detached: setting the streaming flag is
9263 ignored for embedded content.
9265 CMS support is disabled by default and must be explicitly enabled
9266 with the enable-cms configuration option.
9270 * Update the GMP engine glue to do direct copies between BIGNUM and
9271 mpz_t when openssl and GMP use the same limb size. Otherwise the
9272 existing "conversion via a text string export" trick is still used.
9274 *Paul Sheer <paulsheer@gmail.com>*
9276 * Zlib compression BIO. This is a filter BIO which compressed and
9277 uncompresses any data passed through it.
9281 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9282 RFC3394 compatible AES key wrapping.
9286 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9287 sets string data without copying. X509_ALGOR_set0() and
9288 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9289 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9290 from an X509_ATTRIBUTE structure optionally checking it occurs only
9291 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9296 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9297 to get the expected BN_FLG_CONSTTIME behavior.
9299 *Bodo Moeller (Google)*
9303 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9304 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9305 - added some more tests to do_tests.pl
9306 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9307 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9308 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9309 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9310 - various changes to netware.pl to enable gcc-cross builds on Win32
9312 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9313 - various changes to fix missing prototype warnings
9314 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9315 - added AES, WHIRLPOOL and CPUID assembler code to build files
9316 - added missing AES assembler make rules to mk1mf.pl
9317 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9319 *Guenter Knauf <eflash@gmx.net>*
9321 * Implement certificate status request TLS extension defined in RFC3546.
9322 A client can set the appropriate parameters and receive the encoded
9323 OCSP response via a callback. A server can query the supplied parameters
9324 and set the encoded OCSP response in the callback. Add simplified examples
9325 to s_client and s_server.
9329 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9332 + Binary incompatibility of ssl_ctx_st structure
9333 + DTLS interoperation with non-compliant servers
9334 + Don't call get_session_cb() without proposed session
9335 + Fix ia64 assembler code
9337 *Andy Polyakov, Steve Henson*
9339 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9341 * DTLS Handshake overhaul. There were longstanding issues with
9342 OpenSSL DTLS implementation, which were making it impossible for
9343 RFC 4347 compliant client to communicate with OpenSSL server.
9344 Unfortunately just fixing these incompatibilities would "cut off"
9345 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9346 server keeps tolerating non RFC compliant syntax. The opposite is
9347 not true, 0.9.8f client can not communicate with earlier server.
9348 This update even addresses CVE-2007-4995.
9352 * Changes to avoid need for function casts in OpenSSL: some compilers
9353 (gcc 4.2 and later) reject their use.
9354 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9357 * Add RFC4507 support to OpenSSL. This includes the corrections in
9358 RFC4507bis. The encrypted ticket format is an encrypted encoded
9359 SSL_SESSION structure, that way new session features are automatically
9362 If a client application caches session in an SSL_SESSION structure
9363 support is transparent because tickets are now stored in the encoded
9366 The SSL_CTX structure automatically generates keys for ticket
9367 protection in servers so again support should be possible
9368 with no application modification.
9370 If a client or server wishes to disable RFC4507 support then the option
9371 SSL_OP_NO_TICKET can be set.
9373 Add a TLS extension debugging callback to allow the contents of any client
9374 or server extensions to be examined.
9376 This work was sponsored by Google.
9380 * Add initial support for TLS extensions, specifically for the server_name
9381 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9382 have new members for a host name. The SSL data structure has an
9383 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9384 stored in that context to allow for session resumption, even after the
9385 SSL has been switched to a new SSL_CTX in reaction to a client's
9386 server_name extension.
9388 New functions (subject to change):
9390 SSL_get_servername()
9391 SSL_get_servername_type()
9394 New CTRL codes and macros (subject to change):
9396 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9397 - SSL_CTX_set_tlsext_servername_callback()
9398 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9399 - SSL_CTX_set_tlsext_servername_arg()
9400 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9402 openssl s_client has a new '-servername ...' option.
9404 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9405 '-key2 ...', '-servername_fatal' (subject to change). This allows
9406 testing the HostName extension for a specific single host name ('-cert'
9407 and '-key' remain fallbacks for handshakes without HostName
9408 negotiation). If the unrecognized_name alert has to be sent, this by
9409 default is a warning; it becomes fatal with the '-servername_fatal'
9412 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9414 * Add AES and SSE2 assembly language support to VC++ build.
9418 * Mitigate attack on final subtraction in Montgomery reduction.
9422 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9423 (which previously caused an internal error).
9427 * Squeeze another 10% out of IGE mode when in != out.
9431 * AES IGE mode speedup.
9433 *Dean Gaudet (Google)*
9435 * Add the Korean symmetric 128-bit cipher SEED (see
9436 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9437 add SEED ciphersuites from RFC 4162:
9439 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9440 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9441 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9442 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9444 To minimize changes between patchlevels in the OpenSSL 0.9.8
9445 series, SEED remains excluded from compilation unless OpenSSL
9446 is configured with 'enable-seed'.
9448 *KISA, Bodo Moeller*
9450 * Mitigate branch prediction attacks, which can be practical if a
9451 single processor is shared, allowing a spy process to extract
9452 information. For detailed background information, see
9453 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9454 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9455 and Necessary Software Countermeasures"). The core of the change
9456 are new versions BN_div_no_branch() and
9457 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9458 respectively, which are slower, but avoid the security-relevant
9459 conditional branches. These are automatically called by BN_div()
9460 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9461 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9462 remove a conditional branch.
9464 BN_FLG_CONSTTIME is the new name for the previous
9465 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9466 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9467 in the exponent causes BN_mod_exp_mont() to use the alternative
9468 implementation in BN_mod_exp_mont_consttime().) The old name
9469 remains as a deprecated alias.
9471 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9472 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9473 constant-time implementations for more than just exponentiation.
9474 Here too the old name is kept as a deprecated alias.
9476 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9477 the BN_BLINDING structure gets an independent copy of the
9478 modulus. This means that the previous `BIGNUM *m` argument to
9479 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9480 essentially becomes `const BIGNUM *m`, although we can't actually
9481 change this in the header file before 0.9.9. It allows
9482 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9483 enable BN_FLG_CONSTTIME.
9485 *Matthew D Wood (Intel Corp)*
9487 * In the SSL/TLS server implementation, be strict about session ID
9488 context matching (which matters if an application uses a single
9489 external cache for different purposes). Previously,
9490 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9491 set. This did ensure strict client verification, but meant that,
9492 with applications using a single external cache for quite
9493 different requirements, clients could circumvent ciphersuite
9494 restrictions for a given session ID context by starting a session
9495 in a different context.
9499 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9500 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9501 authentication-only ciphersuites.
9505 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9506 not complete and could lead to a possible single byte overflow
9507 ([CVE-2007-5135]) [Ben Laurie]
9509 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9511 * Since AES128 and AES256 (and similarly Camellia128 and
9512 Camellia256) share a single mask bit in the logic of
9513 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9514 kludge to work properly if AES128 is available and AES256 isn't
9515 (or if Camellia128 is available and Camellia256 isn't).
9519 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9520 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9521 When a point or a seed is encoded in a BIT STRING, we need to
9522 prevent the removal of trailing zero bits to get the proper DER
9523 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9524 of a NamedBitList, for which trailing 0 bits need to be removed.)
9528 * Have SSL/TLS server implementation tolerate "mismatched" record
9529 protocol version while receiving ClientHello even if the
9530 ClientHello is fragmented. (The server can't insist on the
9531 particular protocol version it has chosen before the ServerHello
9532 message has informed the client about his choice.)
9536 * Add RFC 3779 support.
9538 *Rob Austein for ARIN, Ben Laurie*
9540 * Load error codes if they are not already present instead of using a
9541 static variable. This allows them to be cleanly unloaded and reloaded.
9542 Improve header file function name parsing.
9546 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9547 or CAPABILITY handshake as required by RFCs.
9551 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9553 * Introduce limits to prevent malicious keys being able to
9554 cause a denial of service. ([CVE-2006-2940])
9556 *Steve Henson, Bodo Moeller*
9558 * Fix ASN.1 parsing of certain invalid structures that can result
9559 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9561 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9562 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9564 * Fix SSL client code which could crash if connecting to a
9565 malicious SSLv2 server. ([CVE-2006-4343])
9567 *Tavis Ormandy and Will Drewry, Google Security Team*
9569 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9570 match only those. Before that, "AES256-SHA" would be interpreted
9571 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9572 the same strength classification in 0.9.7h) as we currently only
9573 have a single AES bit in the ciphersuite description bitmap.
9574 That change, however, also applied to ciphersuite strings such as
9575 "RC4-MD5" that intentionally matched multiple ciphersuites --
9576 namely, SSL 2.0 ciphersuites in addition to the more common ones
9577 from SSL 3.0/TLS 1.0.
9579 So we change the selection algorithm again: Naming an explicit
9580 ciphersuite selects this one ciphersuite, and any other similar
9581 ciphersuite (same bitmap) from *other* protocol versions.
9582 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9583 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9585 Since SSL 2.0 does not have any ciphersuites for which the
9586 128/256 bit distinction would be relevant, this works for now.
9587 The proper fix will be to use different bits for AES128 and
9588 AES256, which would have avoided the problems from the beginning;
9589 however, bits are scarce, so we can only do this in a new release
9590 (not just a patchlevel) when we can change the SSL_CIPHER
9591 definition to split the single 'unsigned long mask' bitmap into
9592 multiple values to extend the available space.
9596 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9598 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9599 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9601 * Add AES IGE and biIGE modes.
9605 * Change the Unix randomness entropy gathering to use poll() when
9606 possible instead of select(), since the latter has some
9607 undesirable limitations.
9609 *Darryl Miles via Richard Levitte and Bodo Moeller*
9611 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9612 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9613 cannot be implicitly activated as part of, e.g., the "AES" alias.
9614 However, please upgrade to OpenSSL 0.9.9[-dev] for
9615 non-experimental use of the ECC ciphersuites to get TLS extension
9616 support, which is required for curve and point format negotiation
9617 to avoid potential handshake problems.
9621 * Disable rogue ciphersuites:
9623 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9624 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9625 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9627 The latter two were purportedly from
9628 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9631 Also deactivate the remaining ciphersuites from
9632 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9633 unofficial, and the ID has long expired.
9637 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9638 dual-core machines) and other potential thread-safety issues.
9642 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9643 versions), which is now available for royalty-free use
9644 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9645 Also, add Camellia TLS ciphersuites from RFC 4132.
9647 To minimize changes between patchlevels in the OpenSSL 0.9.8
9648 series, Camellia remains excluded from compilation unless OpenSSL
9649 is configured with 'enable-camellia'.
9653 * Disable the padding bug check when compression is in use. The padding
9654 bug check assumes the first packet is of even length, this is not
9655 necessarily true if compression is enabled and can result in false
9656 positives causing handshake failure. The actual bug test is ancient
9657 code so it is hoped that implementations will either have fixed it by
9658 now or any which still have the bug do not support compression.
9662 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9664 * When applying a cipher rule check to see if string match is an explicit
9665 cipher suite and only match that one cipher suite if it is.
9669 * Link in manifests for VC++ if needed.
9671 *Austin Ziegler <halostatue@gmail.com>*
9673 * Update support for ECC-based TLS ciphersuites according to
9674 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9675 TLS extensions, which are supported starting with the 0.9.9
9676 branch, not in the OpenSSL 0.9.8 branch).
9680 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9681 opaque EVP_CIPHER_CTX handling.
9685 * Fixes and enhancements to zlib compression code. We now only use
9686 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9687 to conform with the standards mentioned here:
9688 <http://www.zlib.net/DLL_FAQ.txt>
9689 Static zlib linking now works on Windows and the new --with-zlib-include
9690 --with-zlib-lib options to Configure can be used to supply the location
9691 of the headers and library. Gracefully handle case where zlib library
9696 * Several fixes and enhancements to the OID generation code. The old code
9697 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9698 handle numbers larger than ULONG_MAX, truncated printing and had a
9699 non standard OBJ_obj2txt() behaviour.
9703 * Add support for building of engines under engine/ as shared libraries
9704 under VC++ build system.
9708 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9709 Hopefully, we will not see any false combination of paths any more.
9713 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9715 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9716 (part of SSL_OP_ALL). This option used to disable the
9717 countermeasure against man-in-the-middle protocol-version
9718 rollback in the SSL 2.0 server implementation, which is a bad
9719 idea. ([CVE-2005-2969])
9721 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9722 for Information Security, National Institute of Advanced Industrial
9723 Science and Technology [AIST], Japan)*
9725 * Add two function to clear and return the verify parameter flags.
9729 * Keep cipherlists sorted in the source instead of sorting them at
9730 runtime, thus removing the need for a lock.
9734 * Avoid some small subgroup attacks in Diffie-Hellman.
9736 *Nick Mathewson and Ben Laurie*
9738 * Add functions for well-known primes.
9742 * Extended Windows CE support.
9744 *Satoshi Nakamura and Andy Polyakov*
9746 * Initialize SSL_METHOD structures at compile time instead of during
9747 runtime, thus removing the need for a lock.
9751 * Make PKCS7_decrypt() work even if no certificate is supplied by
9752 attempting to decrypt each encrypted key in turn. Add support to
9757 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9759 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9762 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9766 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9767 key into the same file any more.
9771 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9775 * Add -utf8 command line and config file option to 'ca'.
9777 *Stefan <stf@udoma.org*
9779 * Removed the macro des_crypt(), as it seems to conflict with some
9780 libraries. Use DES_crypt().
9784 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9785 involves renaming the source and generated shared-libs for
9786 both. The engines will accept the corrected or legacy ids
9787 ('ncipher' and '4758_cca' respectively) when binding. NB,
9788 this only applies when building 'shared'.
9790 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9792 * Add attribute functions to EVP_PKEY structure. Modify
9793 PKCS12_create() to recognize a CSP name attribute and
9794 use it. Make -CSP option work again in pkcs12 utility.
9798 * Add new functionality to the bn blinding code:
9799 - automatic re-creation of the BN_BLINDING parameters after
9800 a fixed number of uses (currently 32)
9801 - add new function for parameter creation
9802 - introduce flags to control the update behaviour of the
9803 BN_BLINDING parameters
9804 - hide BN_BLINDING structure
9805 Add a second BN_BLINDING slot to the RSA structure to improve
9806 performance when a single RSA object is shared among several
9811 * Add support for DTLS.
9813 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9815 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9816 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9820 * Remove buggy and incomplete DH cert support from
9821 ssl/ssl_rsa.c and ssl/s3_both.c
9825 * Use SHA-1 instead of MD5 as the default digest algorithm for
9826 the apps/openssl applications.
9830 * Compile clean with "-Wall -Wmissing-prototypes
9831 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9832 DEBUG_SAFESTACK must also be set.
9836 * Change ./Configure so that certain algorithms can be disabled by default.
9837 The new counterpiece to "no-xxx" is "enable-xxx".
9839 The patented RC5 and MDC2 algorithms will now be disabled unless
9840 "enable-rc5" and "enable-mdc2", respectively, are specified.
9842 (IDEA remains enabled despite being patented. This is because IDEA
9843 is frequently required for interoperability, and there is no license
9844 fee for non-commercial use. As before, "no-idea" can be used to
9845 avoid this algorithm.)
9849 * Add processing of proxy certificates (see RFC 3820). This work was
9850 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9851 EGEE (Enabling Grids for E-science in Europe).
9855 * RC4 performance overhaul on modern architectures/implementations, such
9856 as Intel P4, IA-64 and AMD64.
9860 * New utility extract-section.pl. This can be used specify an alternative
9861 section number in a pod file instead of having to treat each file as
9862 a separate case in Makefile. This can be done by adding two lines to the
9865 =for comment openssl_section:XXX
9867 The blank line is mandatory.
9871 * New arguments -certform, -keyform and -pass for s_client and s_server
9872 to allow alternative format key and certificate files and passphrase
9877 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9878 update associated structures and add various utility functions.
9880 Add new policy related verify parameters, include policy checking in
9881 standard verify code. Enhance 'smime' application with extra parameters
9882 to support policy checking and print out.
9886 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9887 Nehemiah processors. These extensions support AES encryption in hardware
9888 as well as RNG (though RNG support is currently disabled).
9890 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9892 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9896 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9898 *Andy Polyakov and a number of other people*
9900 * Improved PowerPC platform support. Most notably BIGNUM assembler
9901 implementation contributed by IBM.
9903 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9905 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9906 exponent rather than 'unsigned long'. There is a corresponding change to
9907 the new 'rsa_keygen' element of the RSA_METHOD structure.
9909 *Jelte Jansen, Geoff Thorpe*
9911 * Functionality for creating the initial serial number file is now
9912 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9914 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9915 number file to 1, which is bound to cause problems. To avoid
9916 the problems while respecting compatibility between different 0.9.7
9917 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9918 CA.pl for serial number initialization. With the new release 0.9.8,
9919 we can fix the problem directly in the 'ca' utility.)
9923 * Reduced header interdependencies by declaring more opaque objects in
9924 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9925 give fewer recursive includes, which could break lazy source code - so
9926 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9927 developers should define this symbol when building and using openssl to
9928 ensure they track the recommended behaviour, interfaces, [etc], but
9929 backwards-compatible behaviour prevails when this isn't defined.
9933 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9937 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9938 This will generate a random key of the appropriate length based on the
9939 cipher context. The EVP_CIPHER can provide its own random key generation
9940 routine to support keys of a specific form. This is used in the des and
9941 3des routines to generate a key of the correct parity. Update S/MIME
9942 code to use new functions and hence generate correct parity DES keys.
9943 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9944 valid (weak or incorrect parity).
9948 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9949 as looking them up. This is useful when the verified structure may contain
9950 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9951 present unless the new PKCS7_NO_CRL flag is asserted.
9955 * Extend ASN1 oid configuration module. It now additionally accepts the
9958 shortName = some long name, 1.2.3.4
9962 * Reimplemented the BN_CTX implementation. There is now no more static
9963 limitation on the number of variables it can handle nor the depth of the
9964 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9965 information can now expand as required, and rather than having a single
9966 static array of bignums, BN_CTX now uses a linked-list of such arrays
9967 allowing it to expand on demand whilst maintaining the usefulness of
9968 BN_CTX's "bundling".
9972 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9973 to allow all RSA operations to function using a single BN_CTX.
9977 * Preliminary support for certificate policy evaluation and checking. This
9978 is initially intended to pass the tests outlined in "Conformance Testing
9979 of Relying Party Client Certificate Path Processing Logic" v1.07.
9983 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9984 remained unused and not that useful. A variety of other little bignum
9985 tweaks and fixes have also been made continuing on from the audit (see
9990 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9991 associated ASN1, EVP and SSL functions and old ASN1 macros.
9995 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9996 and this should never fail. So the return value from the use of
9997 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9998 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10002 * BN_CTX_get() should return zero-valued bignums, providing the same
10003 initialised value as BN_new().
10005 *Geoff Thorpe, suggested by Ulf Möller*
10007 * Support for inhibitAnyPolicy certificate extension.
10011 * An audit of the BIGNUM code is underway, for which debugging code is
10012 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10013 is considered valid when processing BIGNUMs, and causes execution to
10014 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10015 further steps are taken to deliberately pollute unused data in BIGNUM
10016 structures to try and expose faulty code further on. For now, openssl will
10017 (in its default mode of operation) continue to tolerate the inconsistent
10018 forms that it has tolerated in the past, but authors and packagers should
10019 consider trying openssl and their own applications when compiled with
10020 these debugging symbols defined. It will help highlight potential bugs in
10021 their own code, and will improve the test coverage for OpenSSL itself. At
10022 some point, these tighter rules will become openssl's default to improve
10023 maintainability, though the assert()s and other overheads will remain only
10024 in debugging configurations. See bn.h for more details.
10026 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10028 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10029 that can only be obtained through BN_CTX_new() (which implicitly
10030 initialises it). The presence of this function only made it possible
10031 to overwrite an existing structure (and cause memory leaks).
10035 * Because of the callback-based approach for implementing LHASH as a
10036 template type, lh_insert() adds opaque objects to hash-tables and
10037 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10038 to clean up those corresponding objects before destroying the hash table
10039 (and losing the object pointers). So some over-zealous constifications in
10040 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10041 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10042 prototyped to have "const" restrictions on the object pointers they are
10043 given (and so aren't required to cast them away any more).
10047 * The tmdiff.h API was so ugly and minimal that our own timing utility
10048 (speed) prefers to use its own implementation. The two implementations
10049 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10050 its object type properly exposed (MS_TM) instead of casting to/from
10051 `char *`. This may still change yet if someone realises MS_TM and
10053 aren't necessarily the greatest nomenclatures - but this is what was used
10054 internally to the implementation so I've used that for now.
10058 * Ensure that deprecated functions do not get compiled when
10059 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10060 the self-tests were still using deprecated key-generation functions so
10061 these have been updated also.
10065 * Reorganise PKCS#7 code to separate the digest location functionality
10066 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10067 New function PKCS7_set_digest() to set the digest type for PKCS#7
10068 digestedData type. Add additional code to correctly generate the
10069 digestedData type and add support for this type in PKCS7 initialization
10074 * New function PKCS7_set0_type_other() this initializes a PKCS7
10075 structure of type "other".
10079 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10080 sure the loop does correctly stop and breaking ("division by zero")
10081 modulus operations are not performed. The (pre-generated) prime
10082 table crypto/bn/bn_prime.h was already correct, but it could not be
10083 re-generated on some platforms because of the "division by zero"
10084 situation in the script.
10086 *Ralf S. Engelschall*
10088 * Update support for ECC-based TLS ciphersuites according to
10089 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10090 SHA-1 now is only used for "small" curves (where the
10091 representation of a field element takes up to 24 bytes); for
10092 larger curves, the field element resulting from ECDH is directly
10093 used as premaster secret.
10095 *Douglas Stebila (Sun Microsystems Laboratories)*
10097 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10098 curve secp160r1 to the tests.
10100 *Douglas Stebila (Sun Microsystems Laboratories)*
10102 * Add the possibility to load symbols globally with DSO.
10104 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10106 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10107 control of the error stack.
10111 * Add support for STORE in ENGINE.
10115 * Add the STORE type. The intention is to provide a common interface
10116 to certificate and key stores, be they simple file-based stores, or
10117 HSM-type store, or LDAP stores, or...
10118 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10122 * Add a generic structure called OPENSSL_ITEM. This can be used to
10123 pass a list of arguments to any function as well as provide a way
10124 for a function to pass data back to the caller.
10128 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10129 works like BUF_strdup() but can be used to duplicate a portion of
10130 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10135 * Add the function sk_find_ex() which works like sk_find(), but will
10136 return an index to an element even if an exact match couldn't be
10137 found. The index is guaranteed to point at the element where the
10138 searched-for key would be inserted to preserve sorting order.
10142 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10143 takes an extra flags argument for optional functionality. Currently,
10144 the following flags are defined:
10146 OBJ_BSEARCH_VALUE_ON_NOMATCH
10147 This one gets OBJ_bsearch_ex() to return a pointer to the first
10148 element where the comparing function returns a negative or zero
10151 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10152 This one gets OBJ_bsearch_ex() to return a pointer to the first
10153 element where the comparing function returns zero. This is useful
10154 if there are more than one element where the comparing function
10159 * Make it possible to create self-signed certificates with 'openssl ca'
10160 in such a way that the self-signed certificate becomes part of the
10161 CA database and uses the same mechanisms for serial number generation
10162 as all other certificate signing. The new flag '-selfsign' enables
10163 this functionality. Adapt CA.sh and CA.pl.in.
10167 * Add functionality to check the public key of a certificate request
10168 against a given private. This is useful to check that a certificate
10169 request can be signed by that key (self-signing).
10173 * Make it possible to have multiple active certificates with the same
10174 subject in the CA index file. This is done only if the keyword
10175 'unique_subject' is set to 'no' in the main CA section (default
10176 if 'CA_default') of the configuration file. The value is saved
10177 with the database itself in a separate index attribute file,
10178 named like the index file with '.attr' appended to the name.
10182 * Generate multi-valued AVAs using '+' notation in config files for
10187 * Support for nameConstraints certificate extension.
10191 * Support for policyConstraints certificate extension.
10195 * Support for policyMappings certificate extension.
10199 * Make sure the default DSA_METHOD implementation only uses its
10200 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10201 and change its own handlers to be NULL so as to remove unnecessary
10202 indirection. This lets alternative implementations fallback to the
10203 default implementation more easily.
10207 * Support for directoryName in GeneralName related extensions
10212 * Make it possible to link applications using Makefile.shared.
10213 Make that possible even when linking against static libraries!
10217 * Support for single pass processing for S/MIME signing. This now
10218 means that S/MIME signing can be done from a pipe, in addition
10219 cleartext signing (multipart/signed type) is effectively streaming
10220 and the signed data does not need to be all held in memory.
10222 This is done with a new flag PKCS7_STREAM. When this flag is set
10223 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10224 is done after the data is output (and digests calculated) in
10225 SMIME_write_PKCS7().
10229 * Add full support for -rpath/-R, both in shared libraries and
10230 applications, at least on the platforms where it's known how
10235 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10236 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10237 will now compute a table of multiples of the generator that
10238 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10239 faster (notably in the case of a single point multiplication,
10240 scalar * generator).
10242 *Nils Larsch, Bodo Moeller*
10244 * IPv6 support for certificate extensions. The various extensions
10245 which use the IP:a.b.c.d can now take IPv6 addresses using the
10246 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10251 * Added an ENGINE that implements RSA by performing private key
10252 exponentiations with the GMP library. The conversions to and from
10253 GMP's mpz_t format aren't optimised nor are any montgomery forms
10254 cached, and on x86 it appears OpenSSL's own performance has caught up.
10255 However there are likely to be other architectures where GMP could
10256 provide a boost. This ENGINE is not built in by default, but it can be
10257 specified at Configure time and should be accompanied by the necessary
10258 linker additions, eg;
10259 ./config -DOPENSSL_USE_GMP -lgmp
10263 * "openssl engine" will not display ENGINE/DSO load failure errors when
10264 testing availability of engines with "-t" - the old behaviour is
10265 produced by increasing the feature's verbosity with "-tt".
10269 * ECDSA routines: under certain error conditions uninitialized BN objects
10270 could be freed. Solution: make sure initialization is performed early
10271 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10276 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10277 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10278 software implementations. For DSA and DH, parameter generation can
10279 also be overridden by providing the appropriate method callbacks.
10283 * Change the "progress" mechanism used in key-generation and
10284 primality testing to functions that take a new BN_GENCB pointer in
10285 place of callback/argument pairs. The new API functions have `_ex`
10286 postfixes and the older functions are reimplemented as wrappers for
10287 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10288 declarations of the old functions to help (graceful) attempts to
10289 migrate to the new functions. Also, the new key-generation API
10290 functions operate on a caller-supplied key-structure and return
10291 success/failure rather than returning a key or NULL - this is to
10292 help make "keygen" another member function of RSA_METHOD etc.
10294 Example for using the new callback interface:
10296 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10297 void *my_arg = ...;
10300 BN_GENCB_set(&my_cb, my_callback, my_arg);
10302 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10303 /* For the meaning of a, b in calls to my_callback(), see the
10304 * documentation of the function that calls the callback.
10305 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10306 * my_callback should return 1 if it wants BN_is_prime_ex()
10307 * to continue, or 0 to stop.
10312 * Change the ZLIB compression method to be stateful, and make it
10313 available to TLS with the number defined in
10314 draft-ietf-tls-compression-04.txt.
10318 * Add the ASN.1 structures and functions for CertificatePair, which
10319 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10321 CertificatePair ::= SEQUENCE {
10322 forward [0] Certificate OPTIONAL,
10323 reverse [1] Certificate OPTIONAL,
10324 -- at least one of the pair shall be present -- }
10326 Also implement the PEM functions to read and write certificate
10327 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10329 This needed to be defined, mostly for the sake of the LDAP
10330 attribute crossCertificatePair, but may prove useful elsewhere as
10335 * Make it possible to inhibit symlinking of shared libraries in
10336 Makefile.shared, for Cygwin's sake.
10340 * Extend the BIGNUM API by creating a function
10341 void BN_set_negative(BIGNUM *a, int neg);
10342 and a macro that behave like
10343 int BN_is_negative(const BIGNUM *a);
10345 to avoid the need to access 'a->neg' directly in applications.
10349 * Implement fast modular reduction for pseudo-Mersenne primes
10350 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10351 EC_GROUP_new_curve_GFp() will now automatically use this
10354 *Nils Larsch <nla@trustcenter.de>*
10356 * Add new lock type (CRYPTO_LOCK_BN).
10360 * Change the ENGINE framework to automatically load engines
10361 dynamically from specific directories unless they could be
10362 found to already be built in or loaded. Move all the
10363 current engines except for the cryptodev one to a new
10364 directory engines/.
10365 The engines in engines/ are built as shared libraries if
10366 the "shared" options was given to ./Configure or ./config.
10367 Otherwise, they are inserted in libcrypto.a.
10368 /usr/local/ssl/engines is the default directory for dynamic
10369 engines, but that can be overridden at configure time through
10370 the usual use of --prefix and/or --openssldir, and at run
10371 time with the environment variable OPENSSL_ENGINES.
10373 *Geoff Thorpe and Richard Levitte*
10375 * Add Makefile.shared, a helper makefile to build shared
10376 libraries. Adapt Makefile.org.
10380 * Add version info to Win32 DLLs.
10382 *Peter 'Luna' Runestig" <peter@runestig.com>*
10384 * Add new 'medium level' PKCS#12 API. Certificates and keys
10385 can be added using this API to created arbitrary PKCS#12
10386 files while avoiding the low level API.
10388 New options to PKCS12_create(), key or cert can be NULL and
10389 will then be omitted from the output file. The encryption
10390 algorithm NIDs can be set to -1 for no encryption, the mac
10391 iteration count can be set to 0 to omit the mac.
10393 Enhance pkcs12 utility by making the -nokeys and -nocerts
10394 options work when creating a PKCS#12 file. New option -nomac
10395 to omit the mac, NONE can be set for an encryption algorithm.
10396 New code is modified to use the enhanced PKCS12_create()
10397 instead of the low level API.
10401 * Extend ASN1 encoder to support indefinite length constructed
10402 encoding. This can output sequences tags and octet strings in
10403 this form. Modify pk7_asn1.c to support indefinite length
10404 encoding. This is experimental and needs additional code to
10405 be useful, such as an ASN1 bio and some enhanced streaming
10408 Extend template encode functionality so that tagging is passed
10409 down to the template encoder.
10413 * Let 'openssl req' fail if an argument to '-newkey' is not
10414 recognized instead of using RSA as a default.
10418 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10419 As these are not official, they are not included in "ALL";
10420 the "ECCdraft" ciphersuite group alias can be used to select them.
10422 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10424 * Add ECDH engine support.
10426 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10428 * Add ECDH in new directory crypto/ecdh/.
10430 *Douglas Stebila (Sun Microsystems Laboratories)*
10432 * Let BN_rand_range() abort with an error after 100 iterations
10433 without success (which indicates a broken PRNG).
10437 * Change BN_mod_sqrt() so that it verifies that the input value
10438 is really the square of the return value. (Previously,
10439 BN_mod_sqrt would show GIGO behaviour.)
10443 * Add named elliptic curves over binary fields from X9.62, SECG,
10444 and WAP/WTLS; add OIDs that were still missing.
10446 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10448 * Extend the EC library for elliptic curves over binary fields
10449 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10452 EC_GF2m_simple_method
10456 EC_GROUP_new_curve_GF2m
10457 EC_GROUP_set_curve_GF2m
10458 EC_GROUP_get_curve_GF2m
10459 EC_POINT_set_affine_coordinates_GF2m
10460 EC_POINT_get_affine_coordinates_GF2m
10461 EC_POINT_set_compressed_coordinates_GF2m
10463 Point compression for binary fields is disabled by default for
10464 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10467 As binary polynomials are represented as BIGNUMs, various members
10468 of the EC_GROUP and EC_POINT data structures can be shared
10469 between the implementations for prime fields and binary fields;
10470 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10471 are essentially identical to their `..._GFp` counterparts.
10472 (For simplicity, the `..._GFp` prefix has been dropped from
10473 various internal method names.)
10475 An internal 'field_div' method (similar to 'field_mul' and
10476 'field_sqr') has been added; this is used only for binary fields.
10478 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10480 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10481 through methods ('mul', 'precompute_mult').
10483 The generic implementations (now internally called 'ec_wNAF_mul'
10484 and 'ec_wNAF_precomputed_mult') remain the default if these
10485 methods are undefined.
10487 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10489 * New function EC_GROUP_get_degree, which is defined through
10490 EC_METHOD. For curves over prime fields, this returns the bit
10491 length of the modulus.
10493 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10495 * New functions EC_GROUP_dup, EC_POINT_dup.
10496 (These simply call ..._new and ..._copy).
10498 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10500 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10501 Polynomials are represented as BIGNUMs (where the sign bit is not
10502 used) in the following functions [macros]:
10505 BN_GF2m_sub [= BN_GF2m_add]
10506 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10507 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10508 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10510 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10511 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10512 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10513 BN_GF2m_cmp [= BN_ucmp]
10515 (Note that only the 'mod' functions are actually for fields GF(2^m).
10516 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10518 For some functions, an the irreducible polynomial defining a
10519 field can be given as an 'unsigned int[]' with strictly
10520 decreasing elements giving the indices of those bits that are set;
10521 i.e., p[] represents the polynomial
10522 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10524 p[0] > p[1] > ... > p[k] = 0.
10525 This applies to the following functions:
10528 BN_GF2m_mod_mul_arr
10529 BN_GF2m_mod_sqr_arr
10530 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10531 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10532 BN_GF2m_mod_exp_arr
10533 BN_GF2m_mod_sqrt_arr
10534 BN_GF2m_mod_solve_quad_arr
10538 Conversion can be performed by the following functions:
10543 bntest.c has additional tests for binary polynomial arithmetic.
10545 Two implementations for BN_GF2m_mod_div() are available.
10546 The default algorithm simply uses BN_GF2m_mod_inv() and
10547 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10548 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10549 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10551 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10553 * Add new error code 'ERR_R_DISABLED' that can be used when some
10554 functionality is disabled at compile-time.
10556 *Douglas Stebila <douglas.stebila@sun.com>*
10558 * Change default behaviour of 'openssl asn1parse' so that more
10559 information is visible when viewing, e.g., a certificate:
10561 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10562 mode the content of non-printable OCTET STRINGs is output in a
10563 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10564 avoid the appearance of a printable string.
10566 *Nils Larsch <nla@trustcenter.de>*
10568 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10570 EC_GROUP_set_asn1_flag()
10571 EC_GROUP_get_asn1_flag()
10572 EC_GROUP_set_point_conversion_form()
10573 EC_GROUP_get_point_conversion_form()
10574 These control ASN1 encoding details:
10575 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10576 has been set to OPENSSL_EC_NAMED_CURVE.
10577 - Points are encoded in uncompressed form by default; options for
10578 asn1_for are as for point2oct, namely
10579 POINT_CONVERSION_COMPRESSED
10580 POINT_CONVERSION_UNCOMPRESSED
10581 POINT_CONVERSION_HYBRID
10583 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10585 EC_GROUP_set_seed()
10586 EC_GROUP_get0_seed()
10587 EC_GROUP_get_seed_len()
10588 This is used only for ASN1 purposes (so far).
10590 *Nils Larsch <nla@trustcenter.de>*
10592 * Add 'field_type' member to EC_METHOD, which holds the NID
10593 of the appropriate field type OID. The new function
10594 EC_METHOD_get_field_type() returns this value.
10596 *Nils Larsch <nla@trustcenter.de>*
10599 EC_POINT_point2bn()
10600 EC_POINT_bn2point()
10601 EC_POINT_point2hex()
10602 EC_POINT_hex2point()
10603 providing useful interfaces to EC_POINT_point2oct() and
10604 EC_POINT_oct2point().
10606 *Nils Larsch <nla@trustcenter.de>*
10608 * Change internals of the EC library so that the functions
10609 EC_GROUP_set_generator()
10610 EC_GROUP_get_generator()
10611 EC_GROUP_get_order()
10612 EC_GROUP_get_cofactor()
10613 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10614 to methods, which would lead to unnecessary code duplication when
10615 adding different types of curves.
10617 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10619 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10620 arithmetic, and such that modified wNAFs are generated
10621 (which avoid length expansion in many cases).
10625 * Add a function EC_GROUP_check_discriminant() (defined via
10626 EC_METHOD) that verifies that the curve discriminant is non-zero.
10628 Add a function EC_GROUP_check() that makes some sanity tests
10629 on a EC_GROUP, its generator and order. This includes
10630 EC_GROUP_check_discriminant().
10632 *Nils Larsch <nla@trustcenter.de>*
10634 * Add ECDSA in new directory crypto/ecdsa/.
10636 Add applications 'openssl ecparam' and 'openssl ecdsa'
10637 (these are based on 'openssl dsaparam' and 'openssl dsa').
10639 ECDSA support is also included in various other files across the
10640 library. Most notably,
10641 - 'openssl req' now has a '-newkey ecdsa:file' option;
10642 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10643 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10644 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10645 them suitable for ECDSA where domain parameters must be
10646 extracted before the specific public key;
10647 - ECDSA engine support has been added.
10649 *Nils Larsch <nla@trustcenter.de>*
10651 * Include some named elliptic curves, and add OIDs from X9.62,
10652 SECG, and WAP/WTLS. Each curve can be obtained from the new
10654 EC_GROUP_new_by_curve_name(),
10655 and the list of available named curves can be obtained with
10656 EC_get_builtin_curves().
10657 Also add a 'curve_name' member to EC_GROUP objects, which can be
10659 EC_GROUP_set_curve_name()
10660 EC_GROUP_get_curve_name()
10662 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10664 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10665 was actually never needed) and in BN_mul(). The removal in BN_mul()
10666 required a small change in bn_mul_part_recursive() and the addition
10667 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10668 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10669 bn_sub_words() and bn_add_words() except they take arrays with
10674 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10676 * Cleanse PEM buffers before freeing them since they may contain
10679 *Benjamin Bennett <ben@psc.edu>*
10681 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10682 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10683 authentication-only ciphersuites.
10687 * Since AES128 and AES256 share a single mask bit in the logic of
10688 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10689 kludge to work properly if AES128 is available and AES256 isn't.
10693 * Expand security boundary to match 1.1.1 module.
10697 * Remove redundant features: hash file source, editing of test vectors
10698 modify fipsld to use external fips_premain.c signature.
10702 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10703 run algorithm test programs.
10707 * Make algorithm test programs more tolerant of whitespace.
10711 * Have SSL/TLS server implementation tolerate "mismatched" record
10712 protocol version while receiving ClientHello even if the
10713 ClientHello is fragmented. (The server can't insist on the
10714 particular protocol version it has chosen before the ServerHello
10715 message has informed the client about his choice.)
10719 * Load error codes if they are not already present instead of using a
10720 static variable. This allows them to be cleanly unloaded and reloaded.
10724 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10726 * Introduce limits to prevent malicious keys being able to
10727 cause a denial of service. ([CVE-2006-2940])
10729 *Steve Henson, Bodo Moeller*
10731 * Fix ASN.1 parsing of certain invalid structures that can result
10732 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10734 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10735 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10737 * Fix SSL client code which could crash if connecting to a
10738 malicious SSLv2 server. ([CVE-2006-4343])
10740 *Tavis Ormandy and Will Drewry, Google Security Team*
10742 * Change ciphersuite string processing so that an explicit
10743 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10744 will no longer include "AES128-SHA"), and any other similar
10745 ciphersuite (same bitmap) from *other* protocol versions (so that
10746 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10747 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10748 changes from 0.9.8b and 0.9.8d.
10752 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10754 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10755 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10757 * Change the Unix randomness entropy gathering to use poll() when
10758 possible instead of select(), since the latter has some
10759 undesirable limitations.
10761 *Darryl Miles via Richard Levitte and Bodo Moeller*
10763 * Disable rogue ciphersuites:
10765 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10766 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10767 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10769 The latter two were purportedly from
10770 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10773 Also deactivate the remaining ciphersuites from
10774 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10775 unofficial, and the ID has long expired.
10779 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10780 dual-core machines) and other potential thread-safety issues.
10784 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10786 * Adapt fipsld and the build system to link against the validated FIPS
10787 module in FIPS mode.
10791 * Fixes for VC++ 2005 build under Windows.
10795 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10796 from a Windows bash shell such as MSYS. It is autodetected from the
10797 "config" script when run from a VC++ environment. Modify standard VC++
10798 build to use fipscanister.o from the GNU make build.
10802 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10804 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10805 The value now differs depending on if you build for FIPS or not.
10806 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10807 safely run with a non-FIPSed libcrypto, as it may crash because of
10808 the difference induced by this change.
10812 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10814 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10815 (part of SSL_OP_ALL). This option used to disable the
10816 countermeasure against man-in-the-middle protocol-version
10817 rollback in the SSL 2.0 server implementation, which is a bad
10818 idea. ([CVE-2005-2969])
10820 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10821 for Information Security, National Institute of Advanced Industrial
10822 Science and Technology [AIST, Japan)]*
10824 * Minimal support for X9.31 signatures and PSS padding modes. This is
10825 mainly for FIPS compliance and not fully integrated at this stage.
10829 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10830 the exponentiation using a fixed-length exponent. (Otherwise,
10831 the information leaked through timing could expose the secret key
10832 after many signatures; cf. Bleichenbacher's attack on DSA with
10837 * Make a new fixed-window mod_exp implementation the default for
10838 RSA, DSA, and DH private-key operations so that the sequence of
10839 squares and multiplies and the memory access pattern are
10840 independent of the particular secret key. This will mitigate
10841 cache-timing and potential related attacks.
10843 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10844 and this is automatically used by BN_mod_exp_mont() if the new flag
10845 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10846 will use this BN flag for private exponents unless the flag
10847 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10848 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10850 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10852 * Change the client implementation for SSLv23_method() and
10853 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10854 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10855 (Previously, the SSL 2.0 backwards compatible Client Hello
10856 message format would be used even with SSL_OP_NO_SSLv2.)
10860 * Add support for smime-type MIME parameter in S/MIME messages which some
10865 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10866 a threadsafe manner. Modify rsa code to use new function and add calls
10867 to dsa and dh code (which had race conditions before).
10871 * Include the fixed error library code in the C error file definitions
10872 instead of fixing them up at runtime. This keeps the error code
10873 structures constant.
10877 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10879 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10882 * Fixes for newer kerberos headers. NB: the casts are needed because
10883 the 'length' field is signed on one version and unsigned on another
10884 with no (?) obvious way to tell the difference, without these VC++
10885 complains. Also the "definition" of FAR (blank) is no longer included
10886 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10887 some needed definitions.
10891 * Undo Cygwin change.
10895 * Added support for proxy certificates according to RFC 3820.
10896 Because they may be a security thread to unaware applications,
10897 they must be explicitly allowed in run-time. See
10898 docs/HOWTO/proxy_certificates.txt for further information.
10902 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10904 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10905 server and client random values. Previously
10906 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10907 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10909 This change has negligible security impact because:
10911 1. Server and client random values still have 24 bytes of pseudo random
10914 2. Server and client random values are sent in the clear in the initial
10917 3. The master secret is derived using the premaster secret (48 bytes in
10918 size for static RSA ciphersuites) as well as client server and random
10921 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10924 *Stephen Henson, reported by UK NISCC*
10926 * Use Windows randomness collection on Cygwin.
10930 * Fix hang in EGD/PRNGD query when communication socket is closed
10931 prematurely by EGD/PRNGD.
10933 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10935 * Prompt for pass phrases when appropriate for PKCS12 input format.
10939 * Back-port of selected performance improvements from development
10940 branch, as well as improved support for PowerPC platforms.
10944 * Add lots of checks for memory allocation failure, error codes to indicate
10945 failure and freeing up memory if a failure occurs.
10947 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10949 * Add new -passin argument to dgst.
10953 * Perform some character comparisons of different types in X509_NAME_cmp:
10954 this is needed for some certificates that re-encode DNs into UTF8Strings
10955 (in violation of RFC3280) and can't or won't issue name rollover
10960 * Make an explicit check during certificate validation to see that
10961 the CA setting in each certificate on the chain is correct. As a
10962 side effect always do the following basic checks on extensions,
10963 not just when there's an associated purpose to the check:
10965 - if there is an unhandled critical extension (unless the user
10966 has chosen to ignore this fault)
10967 - if the path length has been exceeded (if one is set at all)
10968 - that certain extensions fit the associated purpose (if one has
10973 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10975 * Avoid a race condition when CRLs are checked in a multi threaded
10976 environment. This would happen due to the reordering of the revoked
10977 entries during signature checking and serial number lookup. Now the
10978 encoding is cached and the serial number sort performed under a lock.
10979 Add new STACK function sk_is_sorted().
10983 * Add Delta CRL to the extension code.
10987 * Various fixes to s3_pkt.c so alerts are sent properly.
10989 *David Holmes <d.holmes@f5.com>*
10991 * Reduce the chances of duplicate issuer name and serial numbers (in
10992 violation of RFC3280) using the OpenSSL certificate creation utilities.
10993 This is done by creating a random 64 bit value for the initial serial
10994 number when a serial number file is created or when a self signed
10995 certificate is created using 'openssl req -x509'. The initial serial
10996 number file is created using 'openssl x509 -next_serial' in CA.pl
10997 rather than being initialized to 1.
11001 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11003 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11004 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11006 *Joe Orton, Steve Henson*
11008 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11011 *Joe Orton, Steve Henson*
11013 * Make it possible to have multiple active certificates with the same
11014 subject in the CA index file. This is done only if the keyword
11015 'unique_subject' is set to 'no' in the main CA section (default
11016 if 'CA_default') of the configuration file. The value is saved
11017 with the database itself in a separate index attribute file,
11018 named like the index file with '.attr' appended to the name.
11022 * X509 verify fixes. Disable broken certificate workarounds when
11023 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11024 keyUsage extension present. Don't accept CRLs with unhandled critical
11025 extensions: since verify currently doesn't process CRL extensions this
11026 rejects a CRL with *any* critical extensions. Add new verify error codes
11031 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11032 A clarification of RFC2560 will require the use of OCTET STRINGs and
11033 some implementations cannot handle the current raw format. Since OpenSSL
11034 copies and compares OCSP nonces as opaque blobs without any attempt at
11035 parsing them this should not create any compatibility issues.
11039 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11040 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11041 this HMAC (and other) operations are several times slower than OpenSSL
11046 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11048 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11050 * Use the correct content when signing type "other".
11054 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11056 * Fix various bugs revealed by running the NISCC test suite:
11058 Stop out of bounds reads in the ASN1 code when presented with
11059 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11061 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11063 If verify callback ignores invalid public key errors don't try to check
11064 certificate signature with the NULL public key.
11068 * New -ignore_err option in ocsp application to stop the server
11069 exiting on the first error in a request.
11073 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11074 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11079 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11080 extra data after the compression methods not only for TLS 1.0
11081 but also for SSL 3.0 (as required by the specification).
11083 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11085 * Change X509_certificate_type() to mark the key as exported/exportable
11086 when it's 512 *bits* long, not 512 bytes.
11090 * Change AES_cbc_encrypt() so it outputs exact multiple of
11091 blocks during encryption.
11095 * Various fixes to base64 BIO and non blocking I/O. On write
11096 flushes were not handled properly if the BIO retried. On read
11097 data was not being buffered properly and had various logic bugs.
11098 This also affects blocking I/O when the data being decoded is a
11103 * Various S/MIME bugfixes and compatibility changes:
11104 output correct application/pkcs7 MIME type if
11105 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11106 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11107 of files as .eml work). Correctly handle very long lines in MIME
11112 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11114 * Countermeasure against the Klima-Pokorny-Rosa extension of
11115 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11116 a protocol version number mismatch like a decryption error
11117 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11121 * Turn on RSA blinding by default in the default implementation
11122 to avoid a timing attack. Applications that don't want it can call
11123 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11124 They would be ill-advised to do so in most cases.
11126 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11128 * Change RSA blinding code so that it works when the PRNG is not
11129 seeded (in this case, the secret RSA exponent is abused as
11130 an unpredictable seed -- if it is not unpredictable, there
11131 is no point in blinding anyway). Make RSA blinding thread-safe
11132 by remembering the creator's thread ID in rsa->blinding and
11133 having all other threads use local one-time blinding factors
11134 (this requires more computation than sharing rsa->blinding, but
11135 avoids excessive locking; and if an RSA object is not shared
11136 between threads, blinding will still be very fast).
11140 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11141 ENGINE as defaults for all supported algorithms irrespective of
11142 the 'flags' parameter. 'flags' is now honoured, so applications
11143 should make sure they are passing it correctly.
11147 * Target "mingw" now allows native Windows code to be generated in
11148 the Cygwin environment as well as with the MinGW compiler.
11152 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11154 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11155 via timing by performing a MAC computation even if incorrect
11156 block cipher padding has been found. This is a countermeasure
11157 against active attacks where the attacker has to distinguish
11158 between bad padding and a MAC verification error. ([CVE-2003-0078])
11160 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11161 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11162 Martin Vuagnoux (EPFL, Ilion)*
11164 * Make the no-err option work as intended. The intention with no-err
11165 is not to have the whole error stack handling routines removed from
11166 libcrypto, it's only intended to remove all the function name and
11167 reason texts, thereby removing some of the footprint that may not
11168 be interesting if those errors aren't displayed anyway.
11170 NOTE: it's still possible for any application or module to have its
11171 own set of error texts inserted. The routines are there, just not
11172 used by default when no-err is given.
11176 * Add support for FreeBSD on IA64.
11178 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11180 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11181 Kerberos function mit_des_cbc_cksum(). Before this change,
11182 the value returned by DES_cbc_cksum() was like the one from
11183 mit_des_cbc_cksum(), except the bytes were swapped.
11185 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11187 * Allow an application to disable the automatic SSL chain building.
11188 Before this a rather primitive chain build was always performed in
11189 ssl3_output_cert_chain(): an application had no way to send the
11190 correct chain if the automatic operation produced an incorrect result.
11192 Now the chain builder is disabled if either:
11194 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11196 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11198 The reasoning behind this is that an application would not want the
11199 auto chain building to take place if extra chain certificates are
11200 present and it might also want a means of sending no additional
11201 certificates (for example the chain has two certificates and the
11206 * Add the possibility to build without the ENGINE framework.
11208 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11210 * Under Win32 gmtime() can return NULL: check return value in
11211 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11215 * DSA routines: under certain error conditions uninitialized BN objects
11216 could be freed. Solution: make sure initialization is performed early
11217 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11218 Nils Larsch <nla@trustcenter.de> via PR#459)
11222 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11223 checked on reconnect on the client side, therefore session resumption
11224 could still fail with a "ssl session id is different" error. This
11225 behaviour is masked when SSL_OP_ALL is used due to
11226 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11227 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11228 followup to PR #377.
11232 * IA-32 assembler support enhancements: unified ELF targets, support
11233 for SCO/Caldera platforms, fix for Cygwin shared build.
11237 * Add support for FreeBSD on sparc64. As a consequence, support for
11238 FreeBSD on non-x86 processors is separate from x86 processors on
11239 the config script, much like the NetBSD support.
11241 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11243 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11245 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11248 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11249 code (06) was taken as the first octet of the session ID and the last
11250 octet was ignored consequently. As a result SSLv2 client side session
11251 caching could not have worked due to the session ID mismatch between
11253 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11258 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11259 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11264 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11265 seems that in spite of existing for more than a year, many application
11266 author have done nothing to provide the necessary callbacks, which
11267 means that this particular engine will not work properly anywhere.
11268 This is a very unfortunate situation which forces us, in the name
11269 of usability, to give the hw_ncipher.c a static lock, which is part
11271 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11272 appear in 0.9.8 or later. We EXPECT application authors to have
11273 dealt properly with this when 0.9.8 is released (unless we actually
11274 make such changes in the libcrypto locking code that changes will
11275 have to be made anyway).
11279 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11280 octets have been read, EOF or an error occurs. Without this change
11281 some truncated ASN1 structures will not produce an error.
11285 * Disable Heimdal support, since it hasn't been fully implemented.
11286 Still give the possibility to force the use of Heimdal, but with
11287 warnings and a request that patches get sent to openssl-dev.
11291 * Add the VC-CE target, introduce the WINCE sysname, and add
11292 INSTALL.WCE and appropriate conditionals to make it build.
11294 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11296 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11297 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11298 edit numbers of the version.
11300 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11302 * Introduce safe string copy and catenation functions
11303 (BUF_strlcpy() and BUF_strlcat()).
11305 *Ben Laurie (CHATS) and Richard Levitte*
11307 * Avoid using fixed-size buffers for one-line DNs.
11309 *Ben Laurie (CHATS)*
11311 * Add BUF_MEM_grow_clean() to avoid information leakage when
11312 resizing buffers containing secrets, and use where appropriate.
11314 *Ben Laurie (CHATS)*
11316 * Avoid using fixed size buffers for configuration file location.
11318 *Ben Laurie (CHATS)*
11320 * Avoid filename truncation for various CA files.
11322 *Ben Laurie (CHATS)*
11324 * Use sizeof in preference to magic numbers.
11326 *Ben Laurie (CHATS)*
11328 * Avoid filename truncation in cert requests.
11330 *Ben Laurie (CHATS)*
11332 * Add assertions to check for (supposedly impossible) buffer
11335 *Ben Laurie (CHATS)*
11337 * Don't cache truncated DNS entries in the local cache (this could
11338 potentially lead to a spoofing attack).
11340 *Ben Laurie (CHATS)*
11342 * Fix various buffers to be large enough for hex/decimal
11343 representations in a platform independent manner.
11345 *Ben Laurie (CHATS)*
11347 * Add CRYPTO_realloc_clean() to avoid information leakage when
11348 resizing buffers containing secrets, and use where appropriate.
11350 *Ben Laurie (CHATS)*
11352 * Add BIO_indent() to avoid much slightly worrying code to do
11355 *Ben Laurie (CHATS)*
11357 * Convert sprintf()/BIO_puts() to BIO_printf().
11359 *Ben Laurie (CHATS)*
11361 * buffer_gets() could terminate with the buffer only half
11364 *Ben Laurie (CHATS)*
11366 * Add assertions to prevent user-supplied crypto functions from
11367 overflowing internal buffers by having large block sizes, etc.
11369 *Ben Laurie (CHATS)*
11371 * New OPENSSL_assert() macro (similar to assert(), but enabled
11374 *Ben Laurie (CHATS)*
11376 * Eliminate unused copy of key in RC4.
11378 *Ben Laurie (CHATS)*
11380 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11382 *Ben Laurie (CHATS)*
11384 * Fix off-by-one error in EGD path.
11386 *Ben Laurie (CHATS)*
11388 * If RANDFILE path is too long, ignore instead of truncating.
11390 *Ben Laurie (CHATS)*
11392 * Eliminate unused and incorrectly sized X.509 structure
11395 *Ben Laurie (CHATS)*
11397 * Eliminate unused and dangerous function knumber().
11399 *Ben Laurie (CHATS)*
11401 * Eliminate unused and dangerous structure, KSSL_ERR.
11403 *Ben Laurie (CHATS)*
11405 * Protect against overlong session ID context length in an encoded
11406 session object. Since these are local, this does not appear to be
11409 *Ben Laurie (CHATS)*
11411 * Change from security patch (see 0.9.6e below) that did not affect
11412 the 0.9.6 release series:
11414 Remote buffer overflow in SSL3 protocol - an attacker could
11415 supply an oversized master key in Kerberos-enabled versions.
11418 *Ben Laurie (CHATS)*
11420 * Change the SSL kerb5 codes to match RFC 2712.
11424 * Make -nameopt work fully for req and add -reqopt switch.
11426 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11428 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11430 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11432 * Make sure tests can be performed even if the corresponding algorithms
11433 have been removed entirely. This was also the last step to make
11434 OpenSSL compilable with DJGPP under all reasonable conditions.
11436 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11438 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11439 to allow version independent disabling of normally unselected ciphers,
11440 which may be activated as a side-effect of selecting a single cipher.
11442 (E.g., cipher list string "RSA" enables ciphersuites that are left
11443 out of "ALL" because they do not provide symmetric encryption.
11444 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11446 *Lutz Jaenicke, Bodo Moeller*
11448 * Add appropriate support for separate platform-dependent build
11449 directories. The recommended way to make a platform-dependent
11450 build directory is the following (tested on Linux), maybe with
11453 # Place yourself outside of the OpenSSL source tree. In
11454 # this example, the environment variable OPENSSL_SOURCE
11455 # is assumed to contain the absolute OpenSSL source directory.
11456 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11457 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11458 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11459 mkdir -p `dirname $F`
11460 ln -s $OPENSSL_SOURCE/$F $F
11463 To be absolutely sure not to disturb the source tree, a "make clean"
11464 is a good thing. If it isn't successful, don't worry about it,
11465 it probably means the source directory is very clean.
11469 * Make sure any ENGINE control commands make local copies of string
11470 pointers passed to them whenever necessary. Otherwise it is possible
11471 the caller may have overwritten (or deallocated) the original string
11472 data when a later ENGINE operation tries to use the stored values.
11474 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11476 * Improve diagnostics in file reading and command-line digests.
11478 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11480 * Add AES modes CFB and OFB to the object database. Correct an
11481 error in AES-CFB decryption.
11485 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11486 allows existing EVP_CIPHER_CTX structures to be reused after
11487 calling `EVP_*Final()`. This behaviour is used by encryption
11488 BIOs and some applications. This has the side effect that
11489 applications must explicitly clean up cipher contexts with
11490 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11494 * Check the values of dna and dnb in bn_mul_recursive before calling
11495 bn_mul_comba (a non zero value means the a or b arrays do not contain
11496 n2 elements) and fallback to bn_mul_normal if either is not zero.
11500 * Fix escaping of non-ASCII characters when using the -subj option
11501 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11505 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11506 form for "surname", serialNumber has no short form.
11507 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11508 therefore remove "mail" short name for "internet 7".
11509 The OID for unique identifiers in X509 certificates is
11510 x500UniqueIdentifier, not uniqueIdentifier.
11511 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11515 * Add an "init" command to the ENGINE config module and auto initialize
11516 ENGINEs. Without any "init" command the ENGINE will be initialized
11517 after all ctrl commands have been executed on it. If init=1 the
11518 ENGINE is initialized at that point (ctrls before that point are run
11519 on the uninitialized ENGINE and after on the initialized one). If
11520 init=0 then the ENGINE will not be initialized at all.
11524 * Fix the 'app_verify_callback' interface so that the user-defined
11525 argument is actually passed to the callback: In the
11526 SSL_CTX_set_cert_verify_callback() prototype, the callback
11527 declaration has been changed from
11530 int (*cb)(X509_STORE_CTX *,void *);
11531 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11532 i=s->ctx->app_verify_callback(&ctx)
11533 has been changed into
11534 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11536 To update applications using SSL_CTX_set_cert_verify_callback(),
11537 a dummy argument can be added to their callback functions.
11539 *D. K. Smetters <smetters@parc.xerox.com>*
11541 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11543 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11545 * Add and OPENSSL_LOAD_CONF define which will cause
11546 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11547 This allows older applications to transparently support certain
11548 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11549 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11550 load the config file and OPENSSL_add_all_algorithms_conf() which will
11551 always load it have also been added.
11555 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11556 Adjust NIDs and EVP layer.
11558 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11560 * Config modules support in openssl utility.
11562 Most commands now load modules from the config file,
11563 though in a few (such as version) this isn't done
11564 because it couldn't be used for anything.
11566 In the case of ca and req the config file used is
11567 the same as the utility itself: that is the -config
11568 command line option can be used to specify an
11573 * Move default behaviour from OPENSSL_config(). If appname is NULL
11574 use "openssl_conf" if filename is NULL use default openssl config file.
11578 * Add an argument to OPENSSL_config() to allow the use of an alternative
11579 config section name. Add a new flag to tolerate a missing config file
11580 and move code to CONF_modules_load_file().
11584 * Support for crypto accelerator cards from Accelerated Encryption
11585 Processing, www.aep.ie. (Use engine 'aep')
11586 The support was copied from 0.9.6c [engine] and adapted/corrected
11587 to work with the new engine framework.
11589 *AEP Inc. and Richard Levitte*
11591 * Support for SureWare crypto accelerator cards from Baltimore
11592 Technologies. (Use engine 'sureware')
11593 The support was copied from 0.9.6c [engine] and adapted
11594 to work with the new engine framework.
11598 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11599 make the newer ENGINE framework commands for the CHIL engine work.
11601 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11603 * Make it possible to produce shared libraries on ReliantUNIX.
11605 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11607 * Add the configuration target debug-linux-ppro.
11608 Make 'openssl rsa' use the general key loading routines
11609 implemented in apps.c, and make those routines able to
11610 handle the key format FORMAT_NETSCAPE and the variant
11613 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11615 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11617 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11619 * Add -keyform to rsautl, and document -engine.
11621 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11623 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11624 BIO_R_NO_SUCH_FILE error code rather than the generic
11625 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11629 * Add new functions
11630 ERR_peek_last_error
11631 ERR_peek_last_error_line
11632 ERR_peek_last_error_line_data.
11633 These are similar to
11635 ERR_peek_error_line
11636 ERR_peek_error_line_data,
11637 but report on the latest error recorded rather than the first one
11638 still in the error queue.
11640 *Ben Laurie, Bodo Moeller*
11642 * default_algorithms option in ENGINE config module. This allows things
11644 default_algorithms = ALL
11645 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11649 * Preliminary ENGINE config module.
11653 * New experimental application configuration code.
11657 * Change the AES code to follow the same name structure as all other
11658 symmetric ciphers, and behave the same way. Move everything to
11659 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11661 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11663 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11665 *Ben Laurie and Theo de Raadt*
11667 * Add option to output public keys in req command.
11669 *Massimiliano Pala madwolf@openca.org*
11671 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11672 (up to about 10% better than before for P-192 and P-224).
11676 * New functions/macros
11678 SSL_CTX_set_msg_callback(ctx, cb)
11679 SSL_CTX_set_msg_callback_arg(ctx, arg)
11680 SSL_set_msg_callback(ssl, cb)
11681 SSL_set_msg_callback_arg(ssl, arg)
11683 to request calling a callback function
11685 void cb(int write_p, int version, int content_type,
11686 const void *buf, size_t len, SSL *ssl, void *arg)
11688 whenever a protocol message has been completely received
11689 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11690 protocol version according to which the SSL library interprets
11691 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11692 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11693 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11694 specification (change_cipher_spec(20), alert(21), handshake(22)).
11695 'buf' and 'len' point to the actual message, 'ssl' to the
11696 SSL object, and 'arg' is the application-defined value set by
11697 SSL[_CTX]_set_msg_callback_arg().
11699 'openssl s_client' and 'openssl s_server' have new '-msg' options
11700 to enable a callback that displays all protocol messages.
11704 * Change the shared library support so shared libraries are built as
11705 soon as the corresponding static library is finished, and thereby get
11706 openssl and the test programs linked against the shared library.
11707 This still only happens when the keyword "shard" has been given to
11708 the configuration scripts.
11710 NOTE: shared library support is still an experimental thing, and
11711 backward binary compatibility is still not guaranteed.
11713 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11715 * Add support for Subject Information Access extension.
11717 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11719 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11720 additional bytes when new memory had to be allocated, not just
11721 when reusing an existing buffer.
11725 * New command line and configuration option 'utf8' for the req command.
11726 This allows field values to be specified as UTF8 strings.
11730 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11731 runs for the former and machine-readable output for the latter.
11735 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11736 of the e-mail address in the DN (i.e., it will go into a certificate
11737 extension only). The new configuration file option 'email_in_dn = no'
11738 has the same effect.
11740 *Massimiliano Pala madwolf@openca.org*
11742 * Change all functions with names starting with `des_` to be starting
11743 with `DES_` instead. Add wrappers that are compatible with libdes,
11744 but are named `_ossl_old_des_*`. Finally, add macros that map the
11745 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11746 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11747 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11750 Since we provide two compatibility mappings, the user needs to
11751 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11752 compatibility is desired. The default (i.e., when that macro
11753 isn't defined) is OpenSSL 0.9.6c compatibility.
11755 There are also macros that enable and disable the support of old
11756 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11757 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11758 are defined, the default will apply: to support the old des routines.
11760 In either case, one must include openssl/des.h to get the correct
11761 definitions. Do not try to just include openssl/des_old.h, that
11764 NOTE: This is a major break of an old API into a new one. Software
11765 authors are encouraged to switch to the `DES_` style functions. Some
11766 time in the future, des_old.h and the libdes compatibility functions
11767 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11768 default), and then completely removed.
11772 * Test for certificates which contain unsupported critical extensions.
11773 If such a certificate is found during a verify operation it is
11774 rejected by default: this behaviour can be overridden by either
11775 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11776 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11777 X509_supported_extension() has also been added which returns 1 if a
11778 particular extension is supported.
11782 * Modify the behaviour of EVP cipher functions in similar way to digests
11783 to retain compatibility with existing code.
11787 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11788 compatibility with existing code. In particular the 'ctx' parameter does
11789 not have to be to be initialized before the call to EVP_DigestInit() and
11790 it is tidied up after a call to EVP_DigestFinal(). New function
11791 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11792 EVP_MD_CTX_copy() changed to not require the destination to be
11793 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11794 requires the destination to be valid.
11796 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11797 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11801 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11802 so that complete 'Handshake' protocol structures are kept in memory
11803 instead of overwriting 'msg_type' and 'length' with 'body' data.
11807 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11809 *Massimo Santin via Richard Levitte*
11811 * Major restructuring to the underlying ENGINE code. This includes
11812 reduction of linker bloat, separation of pure "ENGINE" manipulation
11813 (initialisation, etc) from functionality dealing with implementations
11814 of specific crypto interfaces. This change also introduces integrated
11815 support for symmetric ciphers and digest implementations - so ENGINEs
11816 can now accelerate these by providing EVP_CIPHER and EVP_MD
11817 implementations of their own. This is detailed in
11818 [crypto/engine/README.md](crypto/engine/README.md)
11819 as it couldn't be adequately described here. However, there are a few
11820 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11821 were changed in the original introduction of ENGINE code have now
11822 reverted back - the hooking from this code to ENGINE is now a good
11823 deal more passive and at run-time, operations deal directly with
11824 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11825 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11826 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11827 they were not being used by the framework as there is no concept of a
11828 BIGNUM_METHOD and they could not be generalised to the new
11829 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11830 ENGINE_cpy() has been removed as it cannot be consistently defined in
11835 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11839 * Change mkdef.pl to sort symbols that get the same entry number,
11840 and make sure the automatically generated functions `ERR_load_*`
11841 become part of libeay.num as well.
11845 * New function SSL_renegotiate_pending(). This returns true once
11846 renegotiation has been requested (either SSL_renegotiate() call
11847 or HelloRequest/ClientHello received from the peer) and becomes
11848 false once a handshake has been completed.
11849 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11850 sends a HelloRequest, but does not ensure that a handshake takes
11851 place. SSL_renegotiate_pending() is useful for checking if the
11852 client has followed the request.)
11856 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11857 By default, clients may request session resumption even during
11858 renegotiation (if session ID contexts permit); with this option,
11859 session resumption is possible only in the first handshake.
11861 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11862 more bits available for options that should not be part of
11863 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11867 * Add some demos for certificate and certificate request creation.
11871 * Make maximum certificate chain size accepted from the peer application
11872 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11873 "Douglas E. Engert" <deengert@anl.gov>.
11877 * Add support for shared libraries for Unixware-7
11878 (Boyd Lynn Gerber <gerberb@zenez.com>).
11882 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11883 be done prior to destruction. Use this to unload error strings from
11884 ENGINEs that load their own error strings. NB: This adds two new API
11885 functions to "get" and "set" this destroy handler in an ENGINE.
11889 * Alter all existing ENGINE implementations (except "openssl" and
11890 "openbsd") to dynamically instantiate their own error strings. This
11891 makes them more flexible to be built both as statically-linked ENGINEs
11892 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11893 Also, add stub code to each that makes building them as self-contained
11894 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11898 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11899 implementations into applications that are completely implemented in
11900 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11901 commands that can be used to configure what shared-library to load and
11902 to control aspects of the way it is handled. Also, made an update to
11903 the [README-Engine.md](README-Engine.md) file
11904 that brings its information up-to-date and
11905 provides some information and instructions on the "dynamic" ENGINE
11906 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11910 * Make it possible to unload ranges of ERR strings with a new
11911 "ERR_unload_strings" function.
11915 * Add a copy() function to EVP_MD.
11919 * Make EVP_MD routines take a context pointer instead of just the
11920 md_data void pointer.
11924 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11925 that the digest can only process a single chunk of data
11926 (typically because it is provided by a piece of
11927 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11928 is only going to provide a single chunk of data, and hence the
11929 framework needn't accumulate the data for oneshot drivers.
11933 * As with "ERR", make it possible to replace the underlying "ex_data"
11934 functions. This change also alters the storage and management of global
11935 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11936 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11937 index counters. The API functions that use this state have been changed
11938 to take a "class_index" rather than pointers to the class's local STACK
11939 and counter, and there is now an API function to dynamically create new
11940 classes. This centralisation allows us to (a) plug a lot of the
11941 thread-safety problems that existed, and (b) makes it possible to clean
11942 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11943 such data would previously have always leaked in application code and
11944 workarounds were in place to make the memory debugging turn a blind eye
11945 to it. Application code that doesn't use this new function will still
11946 leak as before, but their memory debugging output will announce it now
11947 rather than letting it slide.
11949 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11950 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11951 has a return value to indicate success or failure.
11955 * Make it possible to replace the underlying "ERR" functions such that the
11956 global state (2 LHASH tables and 2 locks) is only used by the "default"
11957 implementation. This change also adds two functions to "get" and "set"
11958 the implementation prior to it being automatically set the first time
11959 any other ERR function takes place. Ie. an application can call "get",
11960 pass the return value to a module it has just loaded, and that module
11961 can call its own "set" function using that value. This means the
11962 module's "ERR" operations will use (and modify) the error state in the
11963 application and not in its own statically linked copy of OpenSSL code.
11967 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11968 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11969 the operation, and provides a more encapsulated way for external code
11970 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11971 to use these functions rather than manually incrementing the counts.
11973 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11977 * Add EVP test program.
11981 * Add symmetric cipher support to ENGINE. Expect the API to change!
11985 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11986 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11987 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11988 These allow a CRL to be built without having to access X509_CRL fields
11989 directly. Modify 'ca' application to use new functions.
11993 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11994 bug workarounds. Rollback attack detection is a security feature.
11995 The problem will only arise on OpenSSL servers when TLSv1 is not
11996 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11997 Software authors not wanting to support TLSv1 will have special reasons
11998 for their choice and can explicitly enable this option.
12000 *Bodo Moeller, Lutz Jaenicke*
12002 * Rationalise EVP so it can be extended: don't include a union of
12003 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12004 (similar to those existing for EVP_CIPHER_CTX).
12009 EVP_MD_CTX_init(&md); /* new function call */
12010 EVP_DigestInit(&md, EVP_sha1());
12011 EVP_DigestUpdate(&md, in, len);
12012 EVP_DigestFinal(&md, out, NULL);
12013 EVP_MD_CTX_cleanup(&md); /* new function call */
12017 * Make DES key schedule conform to the usual scheme, as well as
12018 correcting its structure. This means that calls to DES functions
12019 now have to pass a pointer to a des_key_schedule instead of a
12020 plain des_key_schedule (which was actually always a pointer
12023 des_key_schedule ks;
12025 des_set_key_checked(..., &ks);
12026 des_ncbc_encrypt(..., &ks, ...);
12028 (Note that a later change renames 'des_...' into 'DES_...'.)
12032 * Initial reduction of linker bloat: the use of some functions, such as
12033 PEM causes large amounts of unused functions to be linked in due to
12034 poor organisation. For example pem_all.c contains every PEM function
12035 which has a knock on effect of linking in large amounts of (unused)
12036 ASN1 code. Grouping together similar functions and splitting unrelated
12037 functions prevents this.
12041 * Cleanup of EVP macros.
12045 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12046 correct `_ecb suffix`.
12050 * Add initial OCSP responder support to ocsp application. The
12051 revocation information is handled using the text based index
12052 use by the ca application. The responder can either handle
12053 requests generated internally, supplied in files (for example
12054 via a CGI script) or using an internal minimal server.
12058 * Add configuration choices to get zlib compression for TLS.
12062 * Changes to Kerberos SSL for RFC 2712 compliance:
12063 1. Implemented real KerberosWrapper, instead of just using
12064 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12065 2. Implemented optional authenticator field of KerberosWrapper.
12067 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12068 and authenticator structs; see crypto/krb5/.
12070 Generalized Kerberos calls to support multiple Kerberos libraries.
12071 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12072 via Richard Levitte*
12074 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12075 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12076 values for each of the key sizes rather than having just
12077 parameters (and 'speed' generating keys each time).
12081 * Speed up EVP routines.
12084 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12085 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12086 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12087 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12089 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12090 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12091 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12094 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12096 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12100 * Added the OS2-EMX target.
12102 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12104 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12105 to support NCONF routines in extension code. New function CONF_set_nconf()
12106 to allow functions which take an NCONF to also handle the old LHASH
12107 structure: this means that the old CONF compatible routines can be
12108 retained (in particular wrt extensions) without having to duplicate the
12109 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12113 * Enhance the general user interface with mechanisms for inner control
12114 and with possibilities to have yes/no kind of prompts.
12118 * Change all calls to low level digest routines in the library and
12119 applications to use EVP. Add missing calls to HMAC_cleanup() and
12120 don't assume HMAC_CTX can be copied using memcpy().
12122 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12124 * Add the possibility to control engines through control names but with
12125 arbitrary arguments instead of just a string.
12126 Change the key loaders to take a UI_METHOD instead of a callback
12127 function pointer. NOTE: this breaks binary compatibility with earlier
12128 versions of OpenSSL [engine].
12129 Adapt the nCipher code for these new conditions and add a card insertion
12134 * Enhance the general user interface with mechanisms to better support
12135 dialog box interfaces, application-defined prompts, the possibility
12136 to use defaults (for example default passwords from somewhere else)
12137 and interrupts/cancellations.
12141 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12142 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12146 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12147 tidy up some unnecessarily weird code in 'sk_new()').
12149 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12151 * Change the key loading routines for ENGINEs to use the same kind
12152 callback (pem_password_cb) as all other routines that need this
12157 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12158 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12159 than this minimum value is recommended.
12163 * New random seeder for OpenVMS, using the system process statistics
12164 that are easily reachable.
12168 * Windows apparently can't transparently handle global
12169 variables defined in DLLs. Initialisations such as:
12171 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12173 won't compile. This is used by the any applications that need to
12174 declare their own ASN1 modules. This was fixed by adding the option
12175 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12176 needed for static libraries under Win32.
12180 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12181 setting of purpose and trust fields. New X509_STORE trust and
12182 purpose functions and tidy up setting in other SSL functions.
12186 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12187 structure. These are inherited by X509_STORE_CTX when it is
12188 initialised. This allows various defaults to be set in the
12189 X509_STORE structure (such as flags for CRL checking and custom
12190 purpose or trust settings) for functions which only use X509_STORE_CTX
12191 internally such as S/MIME.
12193 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12194 trust settings if they are not set in X509_STORE. This allows X509_STORE
12195 purposes and trust (in S/MIME for example) to override any set by default.
12197 Add command line options for CRL checking to smime, s_client and s_server
12202 * Initial CRL based revocation checking. If the CRL checking flag(s)
12203 are set then the CRL is looked up in the X509_STORE structure and
12204 its validity and signature checked, then if the certificate is found
12205 in the CRL the verify fails with a revoked error.
12207 Various new CRL related callbacks added to X509_STORE_CTX structure.
12209 Command line options added to 'verify' application to support this.
12211 This needs some additional work, such as being able to handle multiple
12212 CRLs with different times, extension based lookup (rather than just
12213 by subject name) and ultimately more complete V2 CRL extension
12218 * Add a general user interface API (crypto/ui/). This is designed
12219 to replace things like des_read_password and friends (backward
12220 compatibility functions using this new API are provided).
12221 The purpose is to remove prompting functions from the DES code
12222 section as well as provide for prompting through dialog boxes in
12223 a window system and the like.
12227 * Add "ex_data" support to ENGINE so implementations can add state at a
12228 per-structure level rather than having to store it globally.
12232 * Make it possible for ENGINE structures to be copied when retrieved by
12233 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12234 This causes the "original" ENGINE structure to act like a template,
12235 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12236 operational state can be localised to each ENGINE structure, despite the
12237 fact they all share the same "methods". New ENGINE structures returned in
12238 this case have no functional references and the return value is the single
12239 structural reference. This matches the single structural reference returned
12240 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12245 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12246 needs to match any other type at all we need to manually clear the
12251 * Changes to the "openssl engine" utility to include;
12252 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12253 about an ENGINE's available control commands.
12254 - executing control commands from command line arguments using the
12255 '-pre' and '-post' switches. '-post' is only used if '-t' is
12256 specified and the ENGINE is successfully initialised. The syntax for
12257 the individual commands are colon-separated, for example;
12258 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12262 * New dynamic control command support for ENGINEs. ENGINEs can now
12263 declare their own commands (numbers), names (strings), descriptions,
12264 and input types for run-time discovery by calling applications. A
12265 subset of these commands are implicitly classed as "executable"
12266 depending on their input type, and only these can be invoked through
12267 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12268 can be based on user input, config files, etc). The distinction is
12269 that "executable" commands cannot return anything other than a boolean
12270 result and can only support numeric or string input, whereas some
12271 discoverable commands may only be for direct use through
12272 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12273 pointers, or other custom uses. The "executable" commands are to
12274 support parameterisations of ENGINE behaviour that can be
12275 unambiguously defined by ENGINEs and used consistently across any
12276 OpenSSL-based application. Commands have been added to all the
12277 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12278 control over shared-library paths without source code alterations.
12282 * Changed all ENGINE implementations to dynamically allocate their
12283 ENGINEs rather than declaring them statically. Apart from this being
12284 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12285 this also allows the implementations to compile without using the
12286 internal engine_int.h header.
12290 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12291 'const' value. Any code that should be able to modify a RAND_METHOD
12292 should already have non-const pointers to it (ie. they should only
12293 modify their own ones).
12297 * Made a variety of little tweaks to the ENGINE code.
12298 - "atalla" and "ubsec" string definitions were moved from header files
12299 to C code. "nuron" string definitions were placed in variables
12300 rather than hard-coded - allowing parameterisation of these values
12301 later on via ctrl() commands.
12302 - Removed unused "#if 0"'d code.
12303 - Fixed engine list iteration code so it uses ENGINE_free() to release
12304 structural references.
12305 - Constified the RAND_METHOD element of ENGINE structures.
12306 - Constified various get/set functions as appropriate and added
12307 missing functions (including a catch-all ENGINE_cpy that duplicates
12308 all ENGINE values onto a new ENGINE except reference counts/state).
12309 - Removed NULL parameter checks in get/set functions. Setting a method
12310 or function to NULL is a way of cancelling out a previously set
12311 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12312 and doesn't justify the extra error symbols and code.
12313 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12314 flags from engine_int.h to engine.h.
12315 - Changed prototypes for ENGINE handler functions (init(), finish(),
12316 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12320 * Implement binary inversion algorithm for BN_mod_inverse in addition
12321 to the algorithm using long division. The binary algorithm can be
12322 used only if the modulus is odd. On 32-bit systems, it is faster
12323 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12324 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12325 up to 450 bits. In 64-bit environments, the binary algorithm
12326 appears to be advantageous for much longer moduli; here we use it
12327 for moduli up to 2048 bits.
12331 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12332 could not support the combine flag in choice fields.
12336 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12337 extensions from a certificate request to the certificate.
12341 * Allow multiple 'certopt' and 'nameopt' options to be separated
12342 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12343 file: this allows the display of the certificate about to be
12344 signed to be customised, to allow certain fields to be included
12345 or excluded and extension details. The old system didn't display
12346 multicharacter strings properly, omitted fields not in the policy
12347 and couldn't display additional details such as extensions.
12351 * Function EC_POINTs_mul for multiple scalar multiplication
12352 of an arbitrary number of elliptic curve points
12353 \sum scalars[i]*points[i],
12354 optionally including the generator defined for the EC_GROUP:
12355 scalar*generator + \sum scalars[i]*points[i].
12357 EC_POINT_mul is a simple wrapper function for the typical case
12358 that the point list has just one item (besides the optional
12363 * First EC_METHODs for curves over GF(p):
12365 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12366 operations and provides various method functions that can also
12367 operate with faster implementations of modular arithmetic.
12369 EC_GFp_mont_method() reuses most functions that are part of
12370 EC_GFp_simple_method, but uses Montgomery arithmetic.
12372 *Bodo Moeller; point addition and point doubling
12373 implementation directly derived from source code provided by
12374 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12376 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12377 crypto/ec/ec_lib.c):
12379 Curves are EC_GROUP objects (with an optional group generator)
12380 based on EC_METHODs that are built into the library.
12382 Points are EC_POINT objects based on EC_GROUP objects.
12384 Most of the framework would be able to handle curves over arbitrary
12385 finite fields, but as there are no obvious types for fields other
12386 than GF(p), some functions are limited to that for now.
12390 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12391 that the file contains a complete HTTP response.
12395 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12396 change the def and num file printf format specifier from "%-40sXXX"
12397 to "%-39s XXX". The latter will always guarantee a space after the
12398 field while the former will cause them to run together if the field
12399 is 40 of more characters long.
12403 * Constify the cipher and digest 'method' functions and structures
12404 and modify related functions to take constant EVP_MD and EVP_CIPHER
12409 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12410 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12414 * Modify `EVP_Digest*()` routines so they now return values. Although the
12415 internal software routines can never fail additional hardware versions
12420 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12422 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12423 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12426 ERR_R_NESTED_ASN1_ERROR
12428 ERR_R_MISSING_ASN1_EOS
12429 were 4 .. 9, conflicting with
12430 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12432 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12433 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12435 Add new error code 'ERR_R_INTERNAL_ERROR'.
12439 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12444 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12445 sets the subject name for a new request or supersedes the
12446 subject name in a given request. Formats that can be parsed are
12447 'CN=Some Name, OU=myOU, C=IT'
12449 'CN=Some Name/OU=myOU/C=IT'.
12451 Add options '-batch' and '-verbose' to 'openssl req'.
12453 *Massimiliano Pala <madwolf@hackmasters.net>*
12455 * Introduce the possibility to access global variables through
12456 functions on platform were that's the best way to handle exporting
12457 global variables in shared libraries. To enable this functionality,
12458 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12459 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12460 is normally done by Configure or something similar).
12462 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12463 in the source file (foo.c) like this:
12465 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12466 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12468 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12469 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12471 OPENSSL_DECLARE_GLOBAL(int,foo);
12472 #define foo OPENSSL_GLOBAL_REF(foo)
12473 OPENSSL_DECLARE_GLOBAL(double,bar);
12474 #define bar OPENSSL_GLOBAL_REF(bar)
12476 The #defines are very important, and therefore so is including the
12477 header file everywhere where the defined globals are used.
12479 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12480 of ASN.1 items, but that structure is a bit different.
12482 The largest change is in util/mkdef.pl which has been enhanced with
12483 better and easier to understand logic to choose which symbols should
12484 go into the Windows .def files as well as a number of fixes and code
12485 cleanup (among others, algorithm keywords are now sorted
12486 lexicographically to avoid constant rewrites).
12490 * In BN_div() keep a copy of the sign of 'num' before writing the
12491 result to 'rm' because if rm==num the value will be overwritten
12492 and produce the wrong result if 'num' is negative: this caused
12493 problems with BN_mod() and BN_nnmod().
12497 * Function OCSP_request_verify(). This checks the signature on an
12498 OCSP request and verifies the signer certificate. The signer
12499 certificate is just checked for a generic purpose and OCSP request
12504 * Add OCSP_check_validity() function to check the validity of OCSP
12505 responses. OCSP responses are prepared in real time and may only
12506 be a few seconds old. Simply checking that the current time lies
12507 between thisUpdate and nextUpdate max reject otherwise valid responses
12508 caused by either OCSP responder or client clock inaccuracy. Instead
12509 we allow thisUpdate and nextUpdate to fall within a certain period of
12510 the current time. The age of the response can also optionally be
12511 checked. Two new options -validity_period and -status_age added to
12516 * If signature or public key algorithm is unrecognized print out its
12517 OID rather that just UNKNOWN.
12521 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12522 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12523 ID to be generated from the issuer certificate alone which can then be
12524 passed to OCSP_id_issuer_cmp().
12528 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12529 ASN1 modules to export functions returning ASN1_ITEM pointers
12530 instead of the ASN1_ITEM structures themselves. This adds several
12531 new macros which allow the underlying ASN1 function/structure to
12532 be accessed transparently. As a result code should not use ASN1_ITEM
12533 references directly (such as &X509_it) but instead use the relevant
12534 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12535 use of the new ASN1 code on platforms where exporting structures
12536 is problematical (for example in shared libraries) but exporting
12537 functions returning pointers to structures is not.
12541 * Add support for overriding the generation of SSL/TLS session IDs.
12542 These callbacks can be registered either in an SSL_CTX or per SSL.
12543 The purpose of this is to allow applications to control, if they wish,
12544 the arbitrary values chosen for use as session IDs, particularly as it
12545 can be useful for session caching in multiple-server environments. A
12546 command-line switch for testing this (and any client code that wishes
12547 to use such a feature) has been added to "s_server".
12549 *Geoff Thorpe, Lutz Jaenicke*
12551 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12552 of the form `#if defined(...) || defined(...) || ...` and
12553 `#if !defined(...) && !defined(...) && ...`. This also avoids
12554 the growing number of special cases it was previously handling.
12558 * Make all configuration macros available for application by making
12559 sure they are available in opensslconf.h, by giving them names starting
12560 with `OPENSSL_` to avoid conflicts with other packages and by making
12561 sure e_os2.h will cover all platform-specific cases together with
12563 Additionally, it is now possible to define configuration/platform-
12564 specific names (called "system identities"). In the C code, these
12565 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12566 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12567 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12572 * New option -set_serial to 'req' and 'x509' this allows the serial
12573 number to use to be specified on the command line. Previously self
12574 signed certificates were hard coded with serial number 0 and the
12575 CA options of 'x509' had to use a serial number in a file which was
12580 * New options to 'ca' utility to support V2 CRL entry extensions.
12581 Currently CRL reason, invalidity date and hold instruction are
12582 supported. Add new CRL extensions to V3 code and some new objects.
12586 * New function EVP_CIPHER_CTX_set_padding() this is used to
12587 disable standard block padding (aka PKCS#5 padding) in the EVP
12588 API, which was previously mandatory. This means that the data is
12589 not padded in any way and so the total length much be a multiple
12590 of the block size, otherwise an error occurs.
12594 * Initial (incomplete) OCSP SSL support.
12598 * New function OCSP_parse_url(). This splits up a URL into its host,
12599 port and path components: primarily to parse OCSP URLs. New -url
12600 option to ocsp utility.
12604 * New nonce behavior. The return value of OCSP_check_nonce() now
12605 reflects the various checks performed. Applications can decide
12606 whether to tolerate certain situations such as an absent nonce
12607 in a response when one was present in a request: the ocsp application
12608 just prints out a warning. New function OCSP_add1_basic_nonce()
12609 this is to allow responders to include a nonce in a response even if
12610 the request is nonce-less.
12614 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12615 skipped when using openssl x509 multiple times on a single input file,
12616 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12620 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12621 set string type: to handle setting ASN1_TIME structures. Fix ca
12622 utility to correctly initialize revocation date of CRLs.
12626 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12627 the clients preferred ciphersuites and rather use its own preferences.
12628 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12629 Internet Explorer by ensuring unchanged hash method during stepup.
12630 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12634 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12635 to aes and add a new 'exist' option to print out symbols that don't
12640 * Additional options to ocsp utility to allow flags to be set and
12641 additional certificates supplied.
12645 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12646 OCSP client a number of certificate to only verify the response
12651 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12652 handle the new API. Currently only ECB, CBC modes supported. Add new
12655 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12656 Encryption Standard (AES) Ciphersuites for Transport Layer
12657 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12658 not enabled by default and were not part of the "ALL" ciphersuite
12659 alias because they were not yet official; they could be
12660 explicitly requested by specifying the "AESdraft" ciphersuite
12661 group alias. In the final release of OpenSSL 0.9.7, the group
12662 alias is called "AES" and is part of "ALL".)
12664 *Ben Laurie, Steve Henson, Bodo Moeller*
12666 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12667 request to response.
12671 * Functions for OCSP responders. OCSP_request_onereq_count(),
12672 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12673 extract information from a certificate request. OCSP_response_create()
12674 creates a response and optionally adds a basic response structure.
12675 OCSP_basic_add1_status() adds a complete single response to a basic
12676 response and returns the OCSP_SINGLERESP structure just added (to allow
12677 extensions to be included for example). OCSP_basic_add1_cert() adds a
12678 certificate to a basic response and OCSP_basic_sign() signs a basic
12679 response with various flags. New helper functions ASN1_TIME_check()
12680 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12681 (converts ASN1_TIME to GeneralizedTime).
12685 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12686 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12687 structure from a certificate. X509_pubkey_digest() digests the public_key
12688 contents: this is used in various key identifiers.
12692 * Make sk_sort() tolerate a NULL argument.
12694 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12696 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12697 passed by the function are trusted implicitly. If any of them signed the
12698 response then it is assumed to be valid and is not verified.
12702 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12703 to data. This was previously part of the PKCS7 ASN1 code. This
12704 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12705 *Steve Henson, reported by Kenneth R. Robinette
12706 <support@securenetterm.com>*
12708 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12709 routines: without these tracing memory leaks is very painful.
12710 Fix leaks in PKCS12 and PKCS7 routines.
12714 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12715 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12716 effectively meant GeneralizedTime would never be used. Now it
12717 is initialised to -1 but X509_time_adj() now has to check the value
12718 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12719 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12720 *Steve Henson, reported by Kenneth R. Robinette
12721 <support@securenetterm.com>*
12723 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12724 result in a zero length in the ASN1_INTEGER structure which was
12725 not consistent with the structure when d2i_ASN1_INTEGER() was used
12726 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12727 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12728 where it did not print out a minus for negative ASN1_INTEGER.
12732 * Add summary printout to ocsp utility. The various functions which
12733 convert status values to strings have been renamed to:
12734 OCSP_response_status_str(), OCSP_cert_status_str() and
12735 OCSP_crl_reason_str() and are no longer static. New options
12736 to verify nonce values and to disable verification. OCSP response
12737 printout format cleaned up.
12741 * Add additional OCSP certificate checks. These are those specified
12742 in RFC2560. This consists of two separate checks: the CA of the
12743 certificate being checked must either be the OCSP signer certificate
12744 or the issuer of the OCSP signer certificate. In the latter case the
12745 OCSP signer certificate must contain the OCSP signing extended key
12746 usage. This check is performed by attempting to match the OCSP
12747 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12748 in the OCSP_CERTID structures of the response.
12752 * Initial OCSP certificate verification added to OCSP_basic_verify()
12753 and related routines. This uses the standard OpenSSL certificate
12754 verify routines to perform initial checks (just CA validity) and
12755 to obtain the certificate chain. Then additional checks will be
12756 performed on the chain. Currently the root CA is checked to see
12757 if it is explicitly trusted for OCSP signing. This is used to set
12758 a root CA as a global signing root: that is any certificate that
12759 chains to that CA is an acceptable OCSP signing certificate.
12763 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12764 extensions from a separate configuration file.
12765 As when reading extensions from the main configuration file,
12766 the '-extensions ...' option may be used for specifying the
12769 *Massimiliano Pala <madwolf@comune.modena.it>*
12771 * New OCSP utility. Allows OCSP requests to be generated or
12772 read. The request can be sent to a responder and the output
12773 parsed, outputted or printed in text form. Not complete yet:
12774 still needs to check the OCSP response validity.
12778 * New subcommands for 'openssl ca':
12779 `openssl ca -status <serial>` prints the status of the cert with
12780 the given serial number (according to the index file).
12781 `openssl ca -updatedb` updates the expiry status of certificates
12784 *Massimiliano Pala <madwolf@comune.modena.it>*
12786 * New '-newreq-nodes' command option to CA.pl. This is like
12787 '-newreq', but calls 'openssl req' with the '-nodes' option
12788 so that the resulting key is not encrypted.
12790 *Damien Miller <djm@mindrot.org>*
12792 * New configuration for the GNU Hurd.
12794 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12796 * Initial code to implement OCSP basic response verify. This
12797 is currently incomplete. Currently just finds the signer's
12798 certificate and verifies the signature on the response.
12802 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12803 value of OPENSSLDIR. This is available via the new '-d' option
12804 to 'openssl version', and is also included in 'openssl version -a'.
12808 * Allowing defining memory allocation callbacks that will be given
12809 file name and line number information in additional arguments
12810 (a `const char*` and an int). The basic functionality remains, as
12811 well as the original possibility to just replace malloc(),
12812 realloc() and free() by functions that do not know about these
12813 additional arguments. To register and find out the current
12814 settings for extended allocation functions, the following
12815 functions are provided:
12817 CRYPTO_set_mem_ex_functions
12818 CRYPTO_set_locked_mem_ex_functions
12819 CRYPTO_get_mem_ex_functions
12820 CRYPTO_get_locked_mem_ex_functions
12822 These work the same way as CRYPTO_set_mem_functions and friends.
12823 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12824 extended allocation function is enabled.
12825 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12826 a conventional allocation function is enabled.
12828 *Richard Levitte, Bodo Moeller*
12830 * Finish off removing the remaining LHASH function pointer casts.
12831 There should no longer be any prototype-casting required when using
12832 the LHASH abstraction, and any casts that remain are "bugs". See
12833 the callback types and macros at the head of lhash.h for details
12834 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12838 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12839 If /dev/[u]random devices are not available or do not return enough
12840 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12842 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12843 /etc/entropy will be queried once each in this sequence, querying stops
12844 when enough entropy was collected without querying more sockets.
12848 * Change the Unix RAND_poll() variant to be able to poll several
12849 random devices, as specified by DEVRANDOM, until a sufficient amount
12850 of data has been collected. We spend at most 10 ms on each file
12851 (select timeout) and read in non-blocking mode. DEVRANDOM now
12852 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12853 (previously it was just the string "/dev/urandom"), so on typical
12854 platforms the 10 ms delay will never occur.
12855 Also separate out the Unix variant to its own file, rand_unix.c.
12856 For VMS, there's a currently-empty rand_vms.c.
12860 * Move OCSP client related routines to ocsp_cl.c. These
12861 provide utility functions which an application needing
12862 to issue a request to an OCSP responder and analyse the
12863 response will typically need: as opposed to those which an
12864 OCSP responder itself would need which will be added later.
12866 OCSP_request_sign() signs an OCSP request with an API similar
12867 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12868 response. OCSP_response_get1_basic() extracts basic response
12869 from response. OCSP_resp_find_status(): finds and extracts status
12870 information from an OCSP_CERTID structure (which will be created
12871 when the request structure is built). These are built from lower
12872 level functions which work on OCSP_SINGLERESP structures but
12873 won't normally be used unless the application wishes to examine
12874 extensions in the OCSP response for example.
12876 Replace nonce routines with a pair of functions.
12877 OCSP_request_add1_nonce() adds a nonce value and optionally
12878 generates a random value. OCSP_check_nonce() checks the
12879 validity of the nonce in an OCSP response.
12883 * Change function OCSP_request_add() to OCSP_request_add0_id().
12884 This doesn't copy the supplied OCSP_CERTID and avoids the
12885 need to free up the newly created id. Change return type
12886 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12887 This can then be used to add extensions to the request.
12888 Deleted OCSP_request_new(), since most of its functionality
12889 is now in OCSP_REQUEST_new() (and the case insensitive name
12890 clash) apart from the ability to set the request name which
12891 will be added elsewhere.
12895 * Update OCSP API. Remove obsolete extensions argument from
12896 various functions. Extensions are now handled using the new
12897 OCSP extension code. New simple OCSP HTTP function which
12898 can be used to send requests and parse the response.
12902 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12903 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12904 uses the special reorder version of SET OF to sort the attributes
12905 and reorder them to match the encoded order. This resolves a long
12906 standing problem: a verify on a PKCS7 structure just after signing
12907 it used to fail because the attribute order did not match the
12908 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12909 it uses the received order. This is necessary to tolerate some broken
12910 software that does not order SET OF. This is handled by encoding
12911 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12912 to produce the required SET OF.
12916 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12917 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12918 files to get correct declarations of the ASN.1 item variables.
12922 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12923 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12924 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12925 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12926 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12927 ASN1_ITEM and no wrapper functions.
12931 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12932 replace the old function pointer based I/O routines. Change most of
12933 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12937 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12938 lines, recognize more "algorithms" that can be deselected, and make
12939 it complain about algorithm deselection that isn't recognised.
12943 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12944 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12945 to use new functions. Add NO_ASN1_OLD which can be set to remove
12946 some old style ASN1 functions: this can be used to determine if old
12947 code will still work when these eventually go away.
12951 * New extension functions for OCSP structures, these follow the
12952 same conventions as certificates and CRLs.
12956 * New function X509V3_add1_i2d(). This automatically encodes and
12957 adds an extension. Its behaviour can be customised with various
12958 flags to append, replace or delete. Various wrappers added for
12959 certificates and CRLs.
12963 * Fix to avoid calling the underlying ASN1 print routine when
12964 an extension cannot be parsed. Correct a typo in the
12965 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12969 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12970 entries for variables.
12974 * Add functionality to apps/openssl.c for detecting locking
12975 problems: As the program is single-threaded, all we have
12976 to do is register a locking callback using an array for
12977 storing which locks are currently held by the program.
12981 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12982 SSL_get_ex_data_X509_STORE_idx(), which is used in
12983 ssl_verify_cert_chain() and thus can be called at any time
12984 during TLS/SSL handshakes so that thread-safety is essential.
12985 Unfortunately, the ex_data design is not at all suited
12986 for multi-threaded use, so it probably should be abolished.
12990 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12992 *Broadcom, tweaked and integrated by Geoff Thorpe*
12994 * Move common extension printing code to new function
12995 X509V3_print_extensions(). Reorganise OCSP print routines and
12996 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13000 * New function X509_signature_print() to remove duplication in some
13005 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13006 set (this was treated exactly the same as SET OF previously). This
13007 is used to reorder the STACK representing the structure to match the
13008 encoding. This will be used to get round a problem where a PKCS7
13009 structure which was signed could not be verified because the STACK
13010 order did not reflect the encoded order.
13014 * Reimplement the OCSP ASN1 module using the new code.
13018 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13019 for its ASN1 operations. The old style function pointers still exist
13020 for now but they will eventually go away.
13024 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13025 completely replaces the old ASN1 functionality with a table driven
13026 encoder and decoder which interprets an ASN1_ITEM structure describing
13027 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13028 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13029 has also been converted to the new form.
13033 * Change BN_mod_exp_recp so that negative moduli are tolerated
13034 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13035 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13036 for negative moduli.
13040 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13041 of not touching the result's sign bit.
13045 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13050 * Changed the LHASH code to use prototypes for callbacks, and created
13051 macros to declare and implement thin (optionally static) functions
13052 that provide type-safety and avoid function pointer casting for the
13053 type-specific callbacks.
13057 * Added Kerberos Cipher Suites to be used with TLS, as written in
13059 *Veers Staats <staatsvr@asc.hpc.mil>,
13060 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13062 * Reformat the FAQ so the different questions and answers can be divided
13063 in sections depending on the subject.
13067 * Have the zlib compression code load ZLIB.DLL dynamically under
13072 * New function BN_mod_sqrt for computing square roots modulo a prime
13073 (using the probabilistic Tonelli-Shanks algorithm unless
13074 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13075 be handled deterministically).
13077 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13079 * Make BN_mod_inverse faster by explicitly handling small quotients
13080 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13081 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13085 * New function BN_kronecker.
13089 * Fix BN_gcd so that it works on negative inputs; the result is
13090 positive unless both parameters are zero.
13091 Previously something reasonably close to an infinite loop was
13092 possible because numbers could be growing instead of shrinking
13093 in the implementation of Euclid's algorithm.
13097 * Fix BN_is_word() and BN_is_one() macros to take into account the
13098 sign of the number in question.
13100 Fix BN_is_word(a,w) to work correctly for w == 0.
13102 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13103 because its test if the absolute value of 'a' equals 'w'.
13104 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13105 it exists mostly for use in the implementations of BN_is_zero(),
13106 BN_is_one(), and BN_is_word().
13110 * New function BN_swap.
13114 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13115 the exponentiation functions are more likely to produce reasonable
13116 results on negative inputs.
13120 * Change BN_mod_mul so that the result is always non-negative.
13121 Previously, it could be negative if one of the factors was negative;
13122 I don't think anyone really wanted that behaviour.
13126 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13127 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13128 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13129 and add new functions:
13138 BN_mod_lshift1_quick
13140 BN_mod_lshift_quick
13142 These functions always generate non-negative results.
13144 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13145 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13147 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13148 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13149 be reduced modulo `m`.
13151 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13154 The following entry accidentally appeared in the CHANGES file
13155 distributed with OpenSSL 0.9.7. The modifications described in
13156 it do *not* apply to OpenSSL 0.9.7.
13158 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13159 was actually never needed) and in BN_mul(). The removal in BN_mul()
13160 required a small change in bn_mul_part_recursive() and the addition
13161 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13162 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13163 bn_sub_words() and bn_add_words() except they take arrays with
13169 * In 'openssl passwd', verify passwords read from the terminal
13170 unless the '-salt' option is used (which usually means that
13171 verification would just waste user's time since the resulting
13172 hash is going to be compared with some given password hash)
13173 or the new '-noverify' option is used.
13175 This is an incompatible change, but it does not affect
13176 non-interactive use of 'openssl passwd' (passwords on the command
13177 line, '-stdin' option, '-in ...' option) and thus should not
13178 cause any problems.
13182 * Remove all references to RSAref, since there's no more need for it.
13186 * Make DSO load along a path given through an environment variable
13187 (SHLIB_PATH) with shl_load().
13191 * Constify the ENGINE code as a result of BIGNUM constification.
13192 Also constify the RSA code and most things related to it. In a
13193 few places, most notable in the depth of the ASN.1 code, ugly
13194 casts back to non-const were required (to be solved at a later
13199 * Make it so the openssl application has all engines loaded by default.
13203 * Constify the BIGNUM routines a little more.
13207 * Add the following functions:
13209 ENGINE_load_cswift()
13211 ENGINE_load_atalla()
13212 ENGINE_load_nuron()
13213 ENGINE_load_builtin_engines()
13215 That way, an application can itself choose if external engines that
13216 are built-in in OpenSSL shall ever be used or not. The benefit is
13217 that applications won't have to be linked with libdl or other dso
13218 libraries unless it's really needed.
13220 Changed 'openssl engine' to load all engines on demand.
13221 Changed the engine header files to avoid the duplication of some
13222 declarations (they differed!).
13226 * 'openssl engine' can now list capabilities.
13230 * Better error reporting in 'openssl engine'.
13234 * Never call load_dh_param(NULL) in s_server.
13238 * Add engine application. It can currently list engines by name and
13239 identity, and test if they are actually available.
13243 * Improve RPM specification file by forcing symbolic linking and making
13244 sure the installed documentation is also owned by root.root.
13246 *Damien Miller <djm@mindrot.org>*
13248 * Give the OpenSSL applications more possibilities to make use of
13249 keys (public as well as private) handled by engines.
13253 * Add OCSP code that comes from CertCo.
13257 * Add VMS support for the Rijndael code.
13261 * Added untested support for Nuron crypto accelerator.
13265 * Add support for external cryptographic devices. This code was
13266 previously distributed separately as the "engine" branch.
13268 *Geoff Thorpe, Richard Levitte*
13270 * Rework the filename-translation in the DSO code. It is now possible to
13271 have far greater control over how a "name" is turned into a filename
13272 depending on the operating environment and any oddities about the
13273 different shared library filenames on each system.
13277 * Support threads on FreeBSD-elf in Configure.
13281 * Fix for SHA1 assembly problem with MASM: it produces
13282 warnings about corrupt line number information when assembling
13283 with debugging information. This is caused by the overlapping
13286 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13289 NCONF_get_number() has no error checking at all. As a replacement,
13290 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13291 promoted strongly. The old NCONF_get_number is kept around for
13292 binary backward compatibility.
13293 Make it possible for methods to load from something other than a BIO,
13294 by providing a function pointer that is given a name instead of a BIO.
13295 For example, this could be used to load configuration data from an
13300 * Fix for non blocking accept BIOs. Added new I/O special reason
13301 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13302 with non blocking I/O was not possible because no retry code was
13303 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13308 * Added the beginnings of Rijndael support.
13312 * Fix for bug in DirectoryString mask setting. Add support for
13313 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13314 to allow certificate printing to more controllable, additional
13315 'certopt' option to 'x509' to allow new printing options to be
13320 * Clean old EAY MD5 hack from e_os.h.
13324 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13326 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13327 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13329 *Joe Orton, Steve Henson*
13331 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13333 * Fix additional bug revealed by the NISCC test suite:
13335 Stop bug triggering large recursion when presented with
13336 certain ASN.1 tags ([CVE-2003-0851])
13340 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13342 * Fix various bugs revealed by running the NISCC test suite:
13344 Stop out of bounds reads in the ASN1 code when presented with
13345 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13347 If verify callback ignores invalid public key errors don't try to check
13348 certificate signature with the NULL public key.
13352 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13353 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13358 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13359 extra data after the compression methods not only for TLS 1.0
13360 but also for SSL 3.0 (as required by the specification).
13362 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13364 * Change X509_certificate_type() to mark the key as exported/exportable
13365 when it's 512 *bits* long, not 512 bytes.
13369 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13371 * Countermeasure against the Klima-Pokorny-Rosa extension of
13372 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13373 a protocol version number mismatch like a decryption error
13374 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13378 * Turn on RSA blinding by default in the default implementation
13379 to avoid a timing attack. Applications that don't want it can call
13380 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13381 They would be ill-advised to do so in most cases.
13383 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13385 * Change RSA blinding code so that it works when the PRNG is not
13386 seeded (in this case, the secret RSA exponent is abused as
13387 an unpredictable seed -- if it is not unpredictable, there
13388 is no point in blinding anyway). Make RSA blinding thread-safe
13389 by remembering the creator's thread ID in rsa->blinding and
13390 having all other threads use local one-time blinding factors
13391 (this requires more computation than sharing rsa->blinding, but
13392 avoids excessive locking; and if an RSA object is not shared
13393 between threads, blinding will still be very fast).
13397 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13399 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13400 via timing by performing a MAC computation even if incorrect
13401 block cipher padding has been found. This is a countermeasure
13402 against active attacks where the attacker has to distinguish
13403 between bad padding and a MAC verification error. ([CVE-2003-0078])
13405 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13406 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13407 Martin Vuagnoux (EPFL, Ilion)*
13409 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13411 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13412 memory from its contents. This is done with a counter that will
13413 place alternating values in each byte. This can be used to solve
13414 two issues: 1) the removal of calls to memset() by highly optimizing
13415 compilers, and 2) cleansing with other values than 0, since those can
13416 be read through on certain media, for example a swap space on disk.
13420 * Bugfix: client side session caching did not work with external caching,
13421 because the session->cipher setting was not restored when reloading
13422 from the external cache. This problem was masked, when
13423 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13424 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13428 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13429 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13431 *Zeev Lieber <zeev-l@yahoo.com>*
13433 * Undo an undocumented change introduced in 0.9.6e which caused
13434 repeated calls to OpenSSL_add_all_ciphers() and
13435 OpenSSL_add_all_digests() to be ignored, even after calling
13440 * Change the default configuration reader to deal with last line not
13441 being properly terminated.
13445 * Change X509_NAME_cmp() so it applies the special rules on handling
13446 DN values that are of type PrintableString, as well as RDNs of type
13447 emailAddress where the value has the type ia5String.
13449 *stefank@valicert.com via Richard Levitte*
13451 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13452 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13453 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13454 the bitwise-OR of the two for use by the majority of applications
13455 wanting this behaviour, and update the docs. The documented
13456 behaviour and actual behaviour were inconsistent and had been
13457 changing anyway, so this is more a bug-fix than a behavioural
13460 *Geoff Thorpe, diagnosed by Nadav Har'El*
13462 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13463 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13467 * Fix initialization code race conditions in
13468 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13469 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13470 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13471 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13472 ssl2_get_cipher_by_char(),
13473 ssl3_get_cipher_by_char().
13475 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13477 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13478 the cached sessions are flushed, as the remove_cb() might use ex_data
13479 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13480 (see [openssl.org #212]).
13482 *Geoff Thorpe, Lutz Jaenicke*
13484 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13485 length, instead of the encoding length to d2i_ASN1_OBJECT.
13489 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13491 * [In 0.9.6g-engine release:]
13492 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13494 *Lynn Gazis <lgazis@rainbow.com>*
13496 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13498 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13499 and get fix the header length calculation.
13500 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13501 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13503 * Use proper error handling instead of 'assertions' in buffer
13504 overflow checks added in 0.9.6e. This prevents DoS (the
13505 assertions could call abort()).
13507 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13509 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13511 * Add various sanity checks to asn1_get_length() to reject
13512 the ASN1 length bytes if they exceed sizeof(long), will appear
13513 negative or the content length exceeds the length of the
13516 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13518 * Fix cipher selection routines: ciphers without encryption had no flags
13519 for the cipher strength set and where therefore not handled correctly
13520 by the selection routines (PR #130).
13524 * Fix EVP_dsa_sha macro.
13529 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13530 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13531 that was added in OpenSSL 0.9.6d.
13533 As the countermeasure turned out to be incompatible with some
13534 broken SSL implementations, the new option is part of SSL_OP_ALL.
13535 SSL_OP_ALL is usually employed when compatibility with weird SSL
13536 implementations is desired (e.g. '-bugs' option to 's_client' and
13537 's_server'), so the new option is automatically set in many
13542 * Changes in security patch:
13544 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13545 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13546 Air Force Materiel Command, USAF, under agreement number
13549 * Add various sanity checks to asn1_get_length() to reject
13550 the ASN1 length bytes if they exceed sizeof(long), will appear
13551 negative or the content length exceeds the length of the
13552 supplied buffer. ([CVE-2002-0659])
13554 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13556 * Assertions for various potential buffer overflows, not known to
13557 happen in practice.
13559 *Ben Laurie (CHATS)*
13561 * Various temporary buffers to hold ASCII versions of integers were
13562 too small for 64 bit platforms. ([CVE-2002-0655])
13563 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13565 * Remote buffer overflow in SSL3 protocol - an attacker could
13566 supply an oversized session ID to a client. ([CVE-2002-0656])
13568 *Ben Laurie (CHATS)*
13570 * Remote buffer overflow in SSL2 protocol - an attacker could
13571 supply an oversized client master key. ([CVE-2002-0656])
13573 *Ben Laurie (CHATS)*
13575 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13577 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13578 encoded as NULL) with id-dsa-with-sha1.
13580 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13582 * Check various `X509_...()` return values in apps/req.c.
13584 *Nils Larsch <nla@trustcenter.de>*
13586 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13587 an end-of-file condition would erroneously be flagged, when the CRLF
13588 was just at the end of a processed block. The bug was discovered when
13589 processing data through a buffering memory BIO handing the data to a
13590 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13591 <ptsekov@syntrex.com> and Nedelcho Stanev.
13595 * Implement a countermeasure against a vulnerability recently found
13596 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13597 before application data chunks to avoid the use of known IVs
13598 with data potentially chosen by the attacker.
13602 * Fix length checks in ssl3_get_client_hello().
13606 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13607 to prevent ssl3_read_internal() from incorrectly assuming that
13608 ssl3_read_bytes() found application data while handshake
13609 processing was enabled when in fact s->s3->in_read_app_data was
13610 merely automatically cleared during the initial handshake.
13612 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13614 * Fix object definitions for Private and Enterprise: they were not
13615 recognized in their shortname (=lowercase) representation. Extend
13616 obj_dat.pl to issue an error when using undefined keywords instead
13617 of silently ignoring the problem (Svenning Sorensen
13618 <sss@sss.dnsalias.net>).
13622 * Fix DH_generate_parameters() so that it works for 'non-standard'
13623 generators, i.e. generators other than 2 and 5. (Previously, the
13624 code did not properly initialise the 'add' and 'rem' values to
13625 BN_generate_prime().)
13627 In the new general case, we do not insist that 'generator' is
13628 actually a primitive root: This requirement is rather pointless;
13629 a generator of the order-q subgroup is just as good, if not
13634 * Map new X509 verification errors to alerts. Discovered and submitted by
13635 Tom Wu <tom@arcot.com>.
13639 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13640 returning non-zero before the data has been completely received
13641 when using non-blocking I/O.
13643 *Bodo Moeller; problem pointed out by John Hughes*
13645 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13647 *Ben Laurie, Lutz Jaenicke*
13649 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13650 Yoram Zahavi <YoramZ@gilian.com>).
13654 * Add information about CygWin 1.3 and on, and preserve proper
13655 configuration for the versions before that.
13657 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13659 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13660 check whether we deal with a copy of a session and do not delete from
13661 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13662 <izhar@checkpoint.com>.
13666 * Do not store session data into the internal session cache, if it
13667 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13668 flag is set). Proposed by Aslam <aslam@funk.com>.
13672 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13677 * [In 0.9.6d-engine release:]
13678 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13680 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13682 * Add the configuration target linux-s390x.
13684 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13686 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13687 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13688 variable as an indication that a ClientHello message has been
13689 received. As the flag value will be lost between multiple
13690 invocations of ssl3_accept when using non-blocking I/O, the
13691 function may not be aware that a handshake has actually taken
13692 place, thus preventing a new session from being added to the
13695 To avoid this problem, we now set s->new_session to 2 instead of
13696 using a local variable.
13698 *Lutz Jaenicke, Bodo Moeller*
13700 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13701 if the SSL_R_LENGTH_MISMATCH error is detected.
13703 *Geoff Thorpe, Bodo Moeller*
13705 * New 'shared_ldflag' column in Configure platform table.
13709 * Fix EVP_CIPHER_mode macro.
13711 *"Dan S. Camper" <dan@bti.net>*
13713 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13714 type, we must throw them away by setting rr->length to 0.
13716 *D P Chang <dpc@qualys.com>*
13718 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13720 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13721 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13722 worked incorrectly for those cases where range = `10..._2` and
13723 `3*range` is two bits longer than range.)
13727 * Only add signing time to PKCS7 structures if it is not already
13732 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13733 OBJ_ld_ce should be OBJ_id_ce.
13734 Also some ip-pda OIDs in crypto/objects/objects.txt were
13735 incorrect (cf. RFC 3039).
13737 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13739 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13740 returns early because it has nothing to do.
13742 *Andy Schneider <andy.schneider@bjss.co.uk>*
13744 * [In 0.9.6c-engine release:]
13745 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13747 *Andy Schneider <andy.schneider@bjss.co.uk>*
13749 * [In 0.9.6c-engine release:]
13750 Add support for Cryptographic Appliance's keyserver technology.
13751 (Use engine 'keyclient')
13753 *Cryptographic Appliances and Geoff Thorpe*
13755 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13756 is called via tools/c89.sh because arguments have to be
13757 rearranged (all '-L' options must appear before the first object
13760 *Richard Shapiro <rshapiro@abinitio.com>*
13762 * [In 0.9.6c-engine release:]
13763 Add support for Broadcom crypto accelerator cards, backported
13766 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13768 * [In 0.9.6c-engine release:]
13769 Add support for SureWare crypto accelerator cards from
13770 Baltimore Technologies. (Use engine 'sureware')
13772 *Baltimore Technologies and Mark Cox*
13774 * [In 0.9.6c-engine release:]
13775 Add support for crypto accelerator cards from Accelerated
13776 Encryption Processing, www.aep.ie. (Use engine 'aep')
13778 *AEP Inc. and Mark Cox*
13780 * Add a configuration entry for gcc on UnixWare.
13782 *Gary Benson <gbenson@redhat.com>*
13784 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13785 messages are stored in a single piece (fixed-length part and
13786 variable-length part combined) and fix various bugs found on the way.
13790 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13791 instead. BIO_gethostbyname() does not know what timeouts are
13792 appropriate, so entries would stay in cache even when they have
13794 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13796 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13797 faced with a pathologically small ClientHello fragment that does
13798 not contain client_version: Instead of aborting with an error,
13799 simply choose the highest available protocol version (i.e.,
13800 TLS 1.0 unless it is disabled). In practice, ClientHello
13801 messages are never sent like this, but this change gives us
13802 strictly correct behaviour at least for TLS.
13806 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13807 never resets s->method to s->ctx->method when called from within
13808 one of the SSL handshake functions.
13810 *Bodo Moeller; problem pointed out by Niko Baric*
13812 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13813 (sent using the client's version number) if client_version is
13814 smaller than the protocol version in use. Also change
13815 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13816 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13817 the client will at least see that alert.
13821 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13826 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13827 client receives HelloRequest while in a handshake.
13829 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13831 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13832 should end in 'break', not 'goto end' which circumvents various
13833 cleanups done in state SSL_ST_OK. But session related stuff
13834 must be disabled for SSL_ST_OK in the case that we just sent a
13837 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13838 before just sending a HelloRequest.
13840 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13842 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13843 reveal whether illegal block cipher padding was found or a MAC
13844 verification error occurred. (Neither SSLerr() codes nor alerts
13845 are directly visible to potential attackers, but the information
13846 may leak via logfiles.)
13848 Similar changes are not required for the SSL 2.0 implementation
13849 because the number of padding bytes is sent in clear for SSL 2.0,
13850 and the extra bytes are just ignored. However ssl/s2_pkt.c
13851 failed to verify that the purported number of padding bytes is in
13856 * Add OpenUNIX-8 support including shared libraries
13857 (Boyd Lynn Gerber <gerberb@zenez.com>).
13861 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13862 'wristwatch attack' using huge encoding parameters (cf.
13863 James H. Manger's CRYPTO 2001 paper). Note that the
13864 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13865 encoding parameters and hence was not vulnerable.
13869 * BN_sqr() bug fix.
13871 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13873 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13874 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13875 followed by modular reduction.
13877 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13879 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13880 equivalent based on BN_pseudo_rand() instead of BN_rand().
13884 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13885 This function was broken, as the check for a new client hello message
13886 to handle SGC did not allow these large messages.
13887 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13891 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13895 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13896 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13900 * Rework the configuration and shared library support for Tru64 Unix.
13901 The configuration part makes use of modern compiler features and
13902 still retains old compiler behavior for those that run older versions
13903 of the OS. The shared library support part includes a variant that
13904 uses the RPATH feature, and is available through the special
13905 configuration target "alpha-cc-rpath", which will never be selected
13908 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13910 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13911 with the same message size as in ssl3_get_certificate_request().
13912 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13913 messages might inadvertently be reject as too long.
13915 *Petr Lampa <lampa@fee.vutbr.cz>*
13917 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13921 * Modified SSL library such that the verify_callback that has been set
13922 specifically for an SSL object with SSL_set_verify() is actually being
13923 used. Before the change, a verify_callback set with this function was
13924 ignored and the verify_callback() set in the SSL_CTX at the time of
13925 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13926 to allow the necessary settings.
13930 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13931 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13932 done automatically (in contradiction to the requirements of the C
13933 standard). This made problems when used from OpenSSH.
13937 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13938 dh->length and always used
13940 BN_rand_range(priv_key, dh->p).
13942 BN_rand_range() is not necessary for Diffie-Hellman, and this
13943 specific range makes Diffie-Hellman unnecessarily inefficient if
13944 dh->length (recommended exponent length) is much smaller than the
13945 length of dh->p. We could use BN_rand_range() if the order of
13946 the subgroup was stored in the DH structure, but we only have
13951 BN_rand(priv_key, l, ...)
13953 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13960 RSA_eay_public_encrypt
13961 RSA_eay_private_decrypt
13962 RSA_eay_private_encrypt (signing)
13963 RSA_eay_public_decrypt (signature verification)
13965 (default implementations for RSA_public_encrypt,
13966 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13967 always reject numbers >= n.
13971 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13972 to synchronize access to 'locking_thread'. This is necessary on
13973 systems where access to 'locking_thread' (an 'unsigned long'
13974 variable) is not atomic.
13978 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13979 *before* setting the 'crypto_lock_rand' flag. The previous code had
13980 a race condition if 0 is a valid thread ID.
13982 *Travis Vitek <vitek@roguewave.com>*
13984 * Add support for shared libraries under Irix.
13986 *Albert Chin-A-Young <china@thewrittenword.com>*
13988 * Add configuration option to build on Linux on both big-endian and
13989 little-endian MIPS.
13991 *Ralf Baechle <ralf@uni-koblenz.de>*
13993 * Add the possibility to create shared libraries on HP-UX.
13997 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13999 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14000 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14001 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14002 PRNG state recovery was possible based on the output of
14003 one PRNG request appropriately sized to gain knowledge on
14004 'md' followed by enough consecutive 1-byte PRNG requests
14005 to traverse all of 'state'.
14007 1. When updating 'md_local' (the current thread's copy of 'md')
14008 during PRNG output generation, hash all of the previous
14009 'md_local' value, not just the half used for PRNG output.
14011 2. Make the number of bytes from 'state' included into the hash
14012 independent from the number of PRNG bytes requested.
14014 The first measure alone would be sufficient to avoid
14015 Markku-Juhani's attack. (Actually it had never occurred
14016 to me that the half of 'md_local' used for chaining was the
14017 half from which PRNG output bytes were taken -- I had always
14018 assumed that the secret half would be used.) The second
14019 measure makes sure that additional data from 'state' is never
14020 mixed into 'md_local' in small portions; this heuristically
14021 further strengthens the PRNG.
14025 * Fix crypto/bn/asm/mips3.s.
14029 * When only the key is given to "enc", the IV is undefined. Print out
14030 an error message in this case.
14034 * Handle special case when X509_NAME is empty in X509 printing routines.
14038 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14039 positive and less than q.
14043 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14044 used: it isn't thread safe and the add_lock_callback should handle
14047 *Paul Rose <Paul.Rose@bridge.com>*
14049 * Verify that incoming data obeys the block size in
14050 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14056 *Ulf Möller, Bodo Möller*
14058 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14059 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14060 when fixing the server behaviour for backwards-compatible 'client
14061 hello' messages. (Note that the attack is impractical against
14062 SSL 3.0 and TLS 1.0 anyway because length and version checking
14063 means that the probability of guessing a valid ciphertext is
14064 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14067 Before 0.9.5, the countermeasure (hide the error by generating a
14068 random 'decryption result') did not work properly because
14069 ERR_clear_error() was missing, meaning that SSL_get_error() would
14070 detect the supposedly ignored error.
14072 Both problems are now fixed.
14076 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14077 (previously it was 1024).
14081 * Fix for compatibility mode trust settings: ignore trust settings
14082 unless some valid trust or reject settings are present.
14086 * Fix for blowfish EVP: its a variable length cipher.
14090 * Fix various bugs related to DSA S/MIME verification. Handle missing
14091 parameters in DSA public key structures and return an error in the
14092 DSA routines if parameters are absent.
14096 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14097 in the current directory if neither $RANDFILE nor $HOME was set.
14098 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14099 caused some confusion to Windows users who haven't defined $HOME.
14100 Thus RAND_file_name() is changed again: e_os.h can define a
14101 DEFAULT_HOME, which will be used if $HOME is not set.
14102 For Windows, we use "C:"; on other platforms, we still require
14103 environment variables.
14105 * Move 'if (!initialized) RAND_poll()' into regions protected by
14106 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14107 having multiple threads call RAND_poll() concurrently.
14111 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14112 combination of a flag and a thread ID variable.
14113 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14114 flag), *other* threads can enter ssleay_add_bytes without obeying
14115 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14116 that they do not hold after the first thread unsets add_do_not_lock).
14120 * Change bctest again: '-x' expressions are not available in all
14121 versions of 'test'.
14125 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14127 * Fix a couple of memory leaks in PKCS7_dataDecode()
14129 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14131 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14132 the default extension for executables, if any. Also, make the perl
14133 scripts that use symlink() to test if it really exists and use "cp"
14134 if it doesn't. All this made OpenSSL compilable and installable in
14139 * Fix for asn1_GetSequence() for indefinite length constructed data.
14140 If SEQUENCE is length is indefinite just set c->slen to the total
14141 amount of data available.
14143 *Steve Henson, reported by shige@FreeBSD.org*
14145 *This change does not apply to 0.9.7.*
14147 * Change bctest to avoid here-documents inside command substitution
14148 (workaround for FreeBSD /bin/sh bug).
14149 For compatibility with Ultrix, avoid shell functions (introduced
14150 in the bctest version that searches along $PATH).
14154 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14155 with des_encrypt() defined on some operating systems, like Solaris
14160 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14161 On the Importance of Eliminating Errors in Cryptographic
14162 Computations, J. Cryptology 14 (2001) 2, 101-119,
14163 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14167 * MIPS assembler BIGNUM division bug fix.
14171 * Disabled incorrect Alpha assembler code.
14175 * Fix PKCS#7 decode routines so they correctly update the length
14176 after reading an EOC for the EXPLICIT tag.
14180 *This change does not apply to 0.9.7.*
14182 * Fix bug in PKCS#12 key generation routines. This was triggered
14183 if a 3DES key was generated with a 0 initial byte. Include
14184 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14185 (but broken) behaviour.
14189 * Enhance bctest to search for a working bc along $PATH and print
14192 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14194 * Fix memory leaks in err.c: free err_data string if necessary;
14195 don't write to the wrong index in ERR_set_error_data.
14199 * Implement ssl23_peek (analogous to ssl23_read), which previously
14204 * Replace rdtsc with `_emit` statements for VC++ version 5.
14206 *Jeremy Cooper <jeremy@baymoo.org>*
14208 * Make it possible to reuse SSLv2 sessions.
14212 * In copy_email() check for >= 0 as a return value for
14213 X509_NAME_get_index_by_NID() since 0 is a valid index.
14215 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14217 * Avoid coredump with unsupported or invalid public keys by checking if
14218 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14219 PKCS7_verify() fails with non detached data.
14223 * Don't use getenv in library functions when run as setuid/setgid.
14224 New function OPENSSL_issetugid().
14228 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14229 due to incorrect handling of multi-threading:
14231 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14233 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14235 3. Count how many times MemCheck_off() has been called so that
14236 nested use can be treated correctly. This also avoids
14237 inband-signalling in the previous code (which relied on the
14238 assumption that thread ID 0 is impossible).
14242 * Add "-rand" option also to s_client and s_server.
14246 * Fix CPU detection on Irix 6.x.
14247 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14248 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14250 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14255 *This change does not apply to 0.9.7.*
14257 * Use the cached encoding of an X509_NAME structure rather than
14258 copying it. This is apparently the reason for the libsafe "errors"
14259 but the code is actually correct.
14263 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14264 Bleichenbacher's DSA attack.
14265 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14266 to be set and top=0 forces the highest bit to be set; top=-1 is new
14267 and leaves the highest bit random.
14269 *Ulf Moeller, Bodo Moeller*
14271 * In the `NCONF_...`-based implementations for `CONF_...` queries
14272 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14273 a temporary CONF structure with the data component set to NULL
14274 (which gives segmentation faults in lh_retrieve).
14275 Instead, use NULL for the CONF pointer in CONF_get_string and
14276 CONF_get_number (which may use environment variables) and directly
14277 return NULL from CONF_get_section.
14281 * Fix potential buffer overrun for EBCDIC.
14285 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14286 keyUsage if basicConstraints absent for a CA.
14290 * Make SMIME_write_PKCS7() write mail header values with a format that
14291 is more generally accepted (no spaces before the semicolon), since
14292 some programs can't parse those values properly otherwise. Also make
14293 sure BIO's that break lines after each write do not create invalid
14298 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14299 macros previously used would not encode an empty SEQUENCE OF
14300 and break the signature.
14304 *This change does not apply to 0.9.7.*
14306 * Zero the premaster secret after deriving the master secret in
14311 * Add some EVP_add_digest_alias registrations (as found in
14312 OpenSSL_add_all_digests()) to SSL_library_init()
14313 aka OpenSSL_add_ssl_algorithms(). This provides improved
14314 compatibility with peers using X.509 certificates
14315 with unconventional AlgorithmIdentifier OIDs.
14319 * Fix for Irix with NO_ASM.
14321 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14323 * ./config script fixes.
14325 *Ulf Moeller, Richard Levitte*
14327 * Fix 'openssl passwd -1'.
14331 * Change PKCS12_key_gen_asc() so it can cope with non null
14332 terminated strings whose length is passed in the passlen
14333 parameter, for example from PEM callbacks. This was done
14334 by adding an extra length parameter to asc2uni().
14336 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14338 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14339 call failed, free the DSA structure.
14343 * Fix to uni2asc() to cope with zero length Unicode strings.
14344 These are present in some PKCS#12 files.
14348 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14349 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14350 when writing a 32767 byte record.
14352 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14354 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14355 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14357 (RSA objects have a reference count access to which is protected
14358 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14359 so they are meant to be shared between threads.)
14360 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14361 "Reddie, Steven" <Steven.Reddie@ca.com>*
14363 * Fix a deadlock in CRYPTO_mem_leaks().
14367 * Use better test patterns in bntest.
14371 * rand_win.c fix for Borland C.
14375 * BN_rshift bugfix for n == 0.
14379 * Add a 'bctest' script that checks for some known 'bc' bugs
14380 so that 'make test' does not abort just because 'bc' is broken.
14384 * Store verify_result within SSL_SESSION also for client side to
14385 avoid potential security hole. (Re-used sessions on the client side
14386 always resulted in verify_result==X509_V_OK, not using the original
14387 result of the server certificate verification.)
14391 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14392 SSL3_RT_APPLICATION_DATA, return 0.
14393 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14398 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14399 releases, have been re-implemented by renaming the previous
14400 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14401 and ssl3_read_internal, respectively, and adding 'peek' parameters
14402 to them. The new ssl[23]_{read,peek} functions are calls to
14403 ssl[23]_read_internal with the 'peek' flag set appropriately.
14404 A 'peek' parameter has also been added to ssl3_read_bytes, which
14405 does the actual work for ssl3_read_internal.
14409 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14410 the method-specific "init()" handler. Also clean up ex_data after
14411 calling the method-specific "finish()" handler. Previously, this was
14412 happening the other way round.
14416 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14417 The previous value, 12, was not always sufficient for BN_mod_exp().
14421 * Make sure that shared libraries get the internal name engine with
14422 the full version number and not just 0. This should mark the
14423 shared libraries as not backward compatible. Of course, this should
14424 be changed again when we can guarantee backward binary compatibility.
14428 * Fix typo in get_cert_by_subject() in by_dir.c
14430 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14432 * Rework the system to generate shared libraries:
14434 - Make note of the expected extension for the shared libraries and
14435 if there is a need for symbolic links from for example libcrypto.so.0
14436 to libcrypto.so.0.9.7. There is extended info in Configure for
14439 - Make as few rebuilds of the shared libraries as possible.
14441 - Still avoid linking the OpenSSL programs with the shared libraries.
14443 - When installing, install the shared libraries separately from the
14448 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14450 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14451 and not in SSL_clear because the latter is also used by the
14452 accept/connect functions; previously, the settings made by
14453 SSL_set_read_ahead would be lost during the handshake.
14455 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14457 * Correct util/mkdef.pl to be selective about disabled algorithms.
14458 Previously, it would create entries for disabled algorithms no
14463 * Added several new manual pages for SSL_* function.
14467 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14469 * In ssl23_get_client_hello, generate an error message when faced
14470 with an initial SSL 3.0/TLS record that is too small to contain the
14471 first two bytes of the ClientHello message, i.e. client_version.
14472 (Note that this is a pathologic case that probably has never happened
14473 in real life.) The previous approach was to use the version number
14474 from the record header as a substitute; but our protocol choice
14475 should not depend on that one because it is not authenticated
14476 by the Finished messages.
14480 * More robust randomness gathering functions for Windows.
14482 *Jeffrey Altman <jaltman@columbia.edu>*
14484 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14485 not set then we don't setup the error code for issuer check errors
14486 to avoid possibly overwriting other errors which the callback does
14487 handle. If an application does set the flag then we assume it knows
14488 what it is doing and can handle the new informational codes
14493 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14494 a general "ANY" type, as such it should be able to decode anything
14495 including tagged types. However it didn't check the class so it would
14496 wrongly interpret tagged types in the same way as their universal
14497 counterpart and unknown types were just rejected. Changed so that the
14498 tagged and unknown types are handled in the same way as a SEQUENCE:
14499 that is the encoding is stored intact. There is also a new type
14500 "V_ASN1_OTHER" which is used when the class is not universal, in this
14501 case we have no idea what the actual type is so we just lump them all
14506 * On VMS, stdout may very well lead to a file that is written to
14507 in a record-oriented fashion. That means that every write() will
14508 write a separate record, which will be read separately by the
14509 programs trying to read from it. This can be very confusing.
14511 The solution is to put a BIO filter in the way that will buffer
14512 text until a linefeed is reached, and then write everything a
14513 line at a time, so every record written will be an actual line,
14514 not chunks of lines and not (usually doesn't happen, but I've
14515 seen it once) several lines in one record. BIO_f_linebuffer() is
14518 Currently, it's a VMS-only method, because that's where it has
14519 been tested well enough.
14523 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14524 it can return incorrect results.
14525 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14526 but it was in 0.9.6-beta[12].)
14530 * Disable the check for content being present when verifying detached
14531 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14532 include zero length content when signing messages.
14536 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14537 BIO_ctrl (for BIO pairs).
14541 * Add DSO method for VMS.
14545 * Bug fix: Montgomery multiplication could produce results with the
14550 * Add RPM specification openssl.spec and modify it to build three
14551 packages. The default package contains applications, application
14552 documentation and run-time libraries. The devel package contains
14553 include files, static libraries and function documentation. The
14554 doc package contains the contents of the doc directory. The original
14555 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14559 * Add a large number of documentation files for many SSL routines.
14561 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14563 * Add a configuration entry for Sony News 4.
14565 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14567 * Don't set the two most significant bits to one when generating a
14568 random number < q in the DSA library.
14572 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14573 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14574 the underlying transport is blocking) if a handshake took place.
14575 (The default behaviour is needed by applications such as s_client
14576 and s_server that use select() to determine when to use SSL_read;
14577 but for applications that know in advance when to expect data, it
14578 just makes things more complicated.)
14582 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14587 * Add a few more EBCDIC conditionals that make `req` and `x509`
14588 work better on such systems.
14590 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14592 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14593 Update PKCS12_parse() so it copies the friendlyName and the
14594 keyid to the certificates aux info.
14598 * Fix bug in PKCS7_verify() which caused an infinite loop
14599 if there was more than one signature.
14601 *Sven Uszpelkat <su@celocom.de>*
14603 * Major change in util/mkdef.pl to include extra information
14604 about each symbol, as well as presenting variables as well
14605 as functions. This change means that there's n more need
14606 to rebuild the .num files when some algorithms are excluded.
14610 * Allow the verify time to be set by an application,
14611 rather than always using the current time.
14615 * Phase 2 verify code reorganisation. The certificate
14616 verify code now looks up an issuer certificate by a
14617 number of criteria: subject name, authority key id
14618 and key usage. It also verifies self signed certificates
14619 by the same criteria. The main comparison function is
14620 X509_check_issued() which performs these checks.
14622 Lot of changes were necessary in order to support this
14623 without completely rewriting the lookup code.
14625 Authority and subject key identifier are now cached.
14627 The LHASH 'certs' is X509_STORE has now been replaced
14628 by a STACK_OF(X509_OBJECT). This is mainly because an
14629 LHASH can't store or retrieve multiple objects with
14630 the same hash value.
14632 As a result various functions (which were all internal
14633 use only) have changed to handle the new X509_STORE
14634 structure. This will break anything that messed round
14635 with X509_STORE internally.
14637 The functions X509_STORE_add_cert() now checks for an
14638 exact match, rather than just subject name.
14640 The X509_STORE API doesn't directly support the retrieval
14641 of multiple certificates matching a given criteria, however
14642 this can be worked round by performing a lookup first
14643 (which will fill the cache with candidate certificates)
14644 and then examining the cache for matches. This is probably
14645 the best we can do without throwing out X509_LOOKUP
14646 entirely (maybe later...).
14648 The X509_VERIFY_CTX structure has been enhanced considerably.
14650 All certificate lookup operations now go via a get_issuer()
14651 callback. Although this currently uses an X509_STORE it
14652 can be replaced by custom lookups. This is a simple way
14653 to bypass the X509_STORE hackery necessary to make this
14654 work and makes it possible to use more efficient techniques
14655 in future. A very simple version which uses a simple
14656 STACK for its trusted certificate store is also provided
14657 using X509_STORE_CTX_trusted_stack().
14659 The verify_cb() and verify() callbacks now have equivalents
14660 in the X509_STORE_CTX structure.
14662 X509_STORE_CTX also has a 'flags' field which can be used
14663 to customise the verify behaviour.
14667 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14668 excludes S/MIME capabilities.
14672 * When a certificate request is read in keep a copy of the
14673 original encoding of the signed data and use it when outputting
14674 again. Signatures then use the original encoding rather than
14675 a decoded, encoded version which may cause problems if the
14676 request is improperly encoded.
14680 * For consistency with other BIO_puts implementations, call
14681 buffer_write(b, ...) directly in buffer_puts instead of calling
14684 In BIO_puts, increment b->num_write as in BIO_write.
14686 *Peter.Sylvester@EdelWeb.fr*
14688 * Fix BN_mul_word for the case where the word is 0. (We have to use
14689 BN_zero, we may not return a BIGNUM with an array consisting of
14690 words set to zero.)
14694 * Avoid calling abort() from within the library when problems are
14695 detected, except if preprocessor symbols have been defined
14696 (such as REF_CHECK, BN_DEBUG etc.).
14700 * New openssl application 'rsautl'. This utility can be
14701 used for low level RSA operations. DER public key
14702 BIO/fp routines also added.
14706 * New Configure entry and patches for compiling on QNX 4.
14708 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14710 * A demo state-machine implementation was sponsored by
14711 Nuron (<http://www.nuron.com/>) and is now available in
14712 demos/state_machine.
14716 * New options added to the 'dgst' utility for signature
14717 generation and verification.
14721 * Unrecognized PKCS#7 content types are now handled via a
14722 catch all ASN1_TYPE structure. This allows unsupported
14723 types to be stored as a "blob" and an application can
14724 encode and decode it manually.
14728 * Fix various signed/unsigned issues to make a_strex.c
14729 compile under VC++.
14731 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14733 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14734 length if passed a buffer. ASN1_INTEGER_to_BN failed
14735 if passed a NULL BN and its argument was negative.
14737 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14739 * Modification to PKCS#7 encoding routines to output definite
14740 length encoding. Since currently the whole structures are in
14741 memory there's not real point in using indefinite length
14742 constructed encoding. However if OpenSSL is compiled with
14743 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14747 * Added BIO_vprintf() and BIO_vsnprintf().
14751 * Added more prefixes to parse for in the strings written
14752 through a logging bio, to cover all the levels that are available
14753 through syslog. The prefixes are now:
14755 PANIC, EMERG, EMR => LOG_EMERG
14756 ALERT, ALR => LOG_ALERT
14757 CRIT, CRI => LOG_CRIT
14758 ERROR, ERR => LOG_ERR
14759 WARNING, WARN, WAR => LOG_WARNING
14760 NOTICE, NOTE, NOT => LOG_NOTICE
14761 INFO, INF => LOG_INFO
14762 DEBUG, DBG => LOG_DEBUG
14764 and as before, if none of those prefixes are present at the
14765 beginning of the string, LOG_ERR is chosen.
14767 On Win32, the `LOG_*` levels are mapped according to this:
14769 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14770 LOG_WARNING => EVENTLOG_WARNING_TYPE
14771 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14775 * Made it possible to reconfigure with just the configuration
14776 argument "reconf" or "reconfigure". The command line arguments
14777 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14778 and are retrieved from there when reconfiguring.
14784 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14786 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14790 * The obj_dat.pl script was messing up the sorting of object
14791 names. The reason was that it compared the quoted version
14792 of strings as a result "OCSP" > "OCSP Signing" because
14793 " > SPACE. Changed script to store unquoted versions of
14794 names and add quotes on output. It was also omitting some
14795 names from the lookup table if they were given a default
14796 value (that is if SN is missing it is given the same
14797 value as LN and vice versa), these are now added on the
14798 grounds that if an object has a name we should be able to
14799 look it up. Finally added warning output when duplicate
14800 short or long names are found.
14804 * Changes needed for Tandem NSK.
14806 *Scott Uroff <scott@xypro.com>*
14808 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14809 RSA_padding_check_SSLv23(), special padding was never detected
14810 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14811 version rollback attacks was not effective.
14813 In s23_clnt.c, don't use special rollback-attack detection padding
14814 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14815 client; similarly, in s23_srvr.c, don't do the rollback check if
14816 SSL 2.0 is the only protocol enabled in the server.
14820 * Make it possible to get hexdumps of unprintable data with 'openssl
14821 asn1parse'. By implication, the functions ASN1_parse_dump() and
14822 BIO_dump_indent() are added.
14826 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14827 these print out strings and name structures based on various
14828 flags including RFC2253 support and proper handling of
14829 multibyte characters. Added options to the 'x509' utility
14830 to allow the various flags to be set.
14834 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14835 Also change the functions X509_cmp_current_time() and
14836 X509_gmtime_adj() work with an ASN1_TIME structure,
14837 this will enable certificates using GeneralizedTime in validity
14838 dates to be checked.
14842 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14843 negative public key encodings) on by default,
14844 NO_NEG_PUBKEY_BUG can be set to disable it.
14848 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14849 content octets. An i2c_ASN1_OBJECT is unnecessary because
14850 the encoding can be trivially obtained from the structure.
14854 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14855 not read locks (`CRYPTO_r_[un]lock`).
14859 * A first attempt at creating official support for shared
14860 libraries through configuration. I've kept it so the
14861 default is static libraries only, and the OpenSSL programs
14862 are always statically linked for now, but there are
14863 preparations for dynamic linking in place.
14864 This has been tested on Linux and Tru64.
14868 * Randomness polling function for Win9x, as described in:
14869 Peter Gutmann, Software Generation of Practically Strong
14874 * Fix so PRNG is seeded in req if using an already existing
14879 * New options to smime application. -inform and -outform
14880 allow alternative formats for the S/MIME message including
14881 PEM and DER. The -content option allows the content to be
14882 specified separately. This should allow things like Netscape
14883 form signing output easier to verify.
14887 * Fix the ASN1 encoding of tags using the 'long form'.
14891 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14892 STRING types. These convert content octets to and from the
14893 underlying type. The actual tag and length octets are
14894 already assumed to have been read in and checked. These
14895 are needed because all other string types have virtually
14896 identical handling apart from the tag. By having versions
14897 of the ASN1 functions that just operate on content octets
14898 IMPLICIT tagging can be handled properly. It also allows
14899 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14900 and ASN1_INTEGER are identical apart from the tag.
14904 * Change the handling of OID objects as follows:
14906 - New object identifiers are inserted in objects.txt, following
14907 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14908 - objects.pl is used to process obj_mac.num and create a new
14910 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14913 This is currently kind of a hack, and the perl code in objects.pl
14914 isn't very elegant, but it works as I intended. The simplest way
14915 to check that it worked correctly is to look in obj_dat.h and
14916 check the array nid_objs and make sure the objects haven't moved
14917 around (this is important!). Additions are OK, as well as
14918 consistent name changes.
14922 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14926 * Addition of the command line parameter '-rand file' to 'openssl req'.
14927 The given file adds to whatever has already been seeded into the
14928 random pool through the RANDFILE configuration file option or
14929 environment variable, or the default random state file.
14933 * mkstack.pl now sorts each macro group into lexical order.
14934 Previously the output order depended on the order the files
14935 appeared in the directory, resulting in needless rewriting
14940 * Patches to make OpenSSL compile under Win32 again. Mostly
14941 work arounds for the VC++ problem that it treats func() as
14942 func(void). Also stripped out the parts of mkdef.pl that
14943 added extra typesafe functions: these no longer exist.
14947 * Reorganisation of the stack code. The macros are now all
14948 collected in safestack.h . Each macro is defined in terms of
14949 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14950 DEBUG_SAFESTACK is now handled in terms of function casts,
14951 this has the advantage of retaining type safety without the
14952 use of additional functions. If DEBUG_SAFESTACK is not defined
14953 then the non typesafe macros are used instead. Also modified the
14954 mkstack.pl script to handle the new form. Needs testing to see
14955 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14956 the default if no major problems. Similar behaviour for ASN1_SET_OF
14957 and PKCS12_STACK_OF.
14961 * When some versions of IIS use the 'NET' form of private key the
14962 key derivation algorithm is different. Normally MD5(password) is
14963 used as a 128 bit RC4 key. In the modified case
14964 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14965 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14966 as the old Netscape_RSA functions except they have an additional
14967 'sgckey' parameter which uses the modified algorithm. Also added
14968 an -sgckey command line option to the rsa utility. Thanks to
14969 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14970 algorithm to openssl-dev.
14974 * The evp_local.h macros were using 'c.##kname' which resulted in
14975 invalid expansion on some systems (SCO 5.0.5 for example).
14976 Corrected to 'c.kname'.
14978 *Phillip Porch <root@theporch.com>*
14980 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14981 a STACK of email addresses from a certificate or request, these look
14982 in the subject name and the subject alternative name extensions and
14983 omit any duplicate addresses.
14987 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14988 This makes DSA verification about 2 % faster.
14992 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14993 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14994 plus overhead for 1024 bit moduli).
14995 This makes exponentiations about 0.5 % faster for 1024 bit
14996 exponents (as measured by "openssl speed rsa2048").
15000 * Rename memory handling macros to avoid conflicts with other
15002 Malloc => OPENSSL_malloc
15003 Malloc_locked => OPENSSL_malloc_locked
15004 Realloc => OPENSSL_realloc
15005 Free => OPENSSL_free
15009 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15010 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15014 * CygWin32 support.
15016 *John Jarvie <jjarvie@newsguy.com>*
15018 * The type-safe stack code has been rejigged. It is now only compiled
15019 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15020 by default all type-specific stack functions are "#define"d back to
15021 standard stack functions. This results in more streamlined output
15022 but retains the type-safety checking possibilities of the original
15027 * The STACK code has been cleaned up, and certain type declarations
15028 that didn't make a lot of sense have been brought in line. This has
15029 also involved a cleanup of sorts in safestack.h to more correctly
15030 map type-safe stack functions onto their plain stack counterparts.
15031 This work has also resulted in a variety of "const"ifications of
15032 lots of the code, especially `_cmp` operations which should normally
15033 be prototyped with "const" parameters anyway.
15037 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15038 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15039 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15040 where all of 'md' is used each time the PRNG is used, but 'state'
15041 is used only indexed by a cyclic counter. As entropy may not be
15042 well distributed from the beginning, 'md' is important as a
15043 chaining variable. However, the output function chains only half
15044 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15045 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15046 in all of 'state' being rewritten, with the new values depending
15047 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15051 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15052 the handshake is continued after ssl_verify_cert_chain();
15053 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15054 can lead to 'unexplainable' connection aborts later.
15056 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15058 * Major EVP API cipher revision.
15059 Add hooks for extra EVP features. This allows various cipher
15060 parameters to be set in the EVP interface. Support added for variable
15061 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15062 setting of RC2 and RC5 parameters.
15064 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15067 Remove lots of duplicated code from the EVP library. For example *every*
15068 cipher init() function handles the 'iv' in the same way according to the
15069 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15070 for CFB and OFB modes they zero ctx->num.
15072 New functionality allows removal of S/MIME code RC2 hack.
15074 Most of the routines have the same form and so can be declared in terms
15077 By shifting this to the top level EVP_CipherInit() it can be removed from
15078 all individual ciphers. If the cipher wants to handle IVs or keys
15079 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15082 Change lots of functions like EVP_EncryptUpdate() to now return a
15083 value: although software versions of the algorithms cannot fail
15084 any installed hardware versions can.
15088 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15089 this option is set, tolerate broken clients that send the negotiated
15090 protocol version number instead of the requested protocol version
15095 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15096 i.e. non-zero for export ciphersuites, zero otherwise.
15097 Previous versions had this flag inverted, inconsistent with
15098 rsa_tmp_cb (..._TMP_RSA_CB).
15100 *Bodo Moeller; problem reported by Amit Chopra*
15102 * Add missing DSA library text string. Work around for some IIS
15103 key files with invalid SEQUENCE encoding.
15107 * Add a document (doc/standards.txt) that list all kinds of standards
15108 and so on that are implemented in OpenSSL.
15112 * Enhance c_rehash script. Old version would mishandle certificates
15113 with the same subject name hash and wouldn't handle CRLs at all.
15114 Added -fingerprint option to crl utility, to support new c_rehash
15119 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15123 * Fix for SSL server purpose checking. Server checking was
15124 rejecting certificates which had extended key usage present
15125 but no ssl client purpose.
15127 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15129 * Make PKCS#12 code work with no password. The PKCS#12 spec
15130 is a little unclear about how a blank password is handled.
15131 Since the password in encoded as a BMPString with terminating
15132 double NULL a zero length password would end up as just the
15133 double NULL. However no password at all is different and is
15134 handled differently in the PKCS#12 key generation code. NS
15135 treats a blank password as zero length. MSIE treats it as no
15136 password on export: but it will try both on import. We now do
15137 the same: PKCS12_parse() tries zero length and no password if
15138 the password is set to "" or NULL (NULL is now a valid password:
15139 it wasn't before) as does the pkcs12 application.
15143 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15144 perror when PEM_read_bio_X509_REQ fails, the error message must
15145 be obtained from the error queue.
15149 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15150 it in ERR_remove_state if appropriate, and change ERR_get_state
15151 accordingly to avoid race conditions (this is necessary because
15152 thread_hash is no longer constant once set).
15156 * Bugfix for linux-elf makefile.one.
15160 * RSA_get_default_method() will now cause a default
15161 RSA_METHOD to be chosen if one doesn't exist already.
15162 Previously this was only set during a call to RSA_new()
15163 or RSA_new_method(NULL) meaning it was possible for
15164 RSA_get_default_method() to return NULL.
15168 * Added native name translation to the existing DSO code
15169 that will convert (if the flag to do so is set) filenames
15170 that are sufficiently small and have no path information
15171 into a canonical native form. Eg. "blah" converted to
15172 "libblah.so" or "blah.dll" etc.
15176 * New function ERR_error_string_n(e, buf, len) which is like
15177 ERR_error_string(e, buf), but writes at most 'len' bytes
15178 including the 0 terminator. For ERR_error_string_n, 'buf'
15181 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15183 * CONF library reworked to become more general. A new CONF
15184 configuration file reader "class" is implemented as well as a
15185 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15186 old `CONF_*` functions are still there, but are reimplemented to
15187 work in terms of the new functions. Also, a set of functions
15188 to handle the internal storage of the configuration data is
15189 provided to make it easier to write new configuration file
15190 reader "classes" (I can definitely see something reading a
15191 configuration file in XML format, for example), called `_CONF_*`,
15192 or "the configuration storage API"...
15194 The new configuration file reading functions are:
15196 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15197 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15199 NCONF_default, NCONF_WIN32
15201 NCONF_dump_fp, NCONF_dump_bio
15203 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15204 NCONF_new creates a new CONF object. This works in the same way
15205 as other interfaces in OpenSSL, like the BIO interface.
15206 `NCONF_dump_*` dump the internal storage of the configuration file,
15207 which is useful for debugging. All other functions take the same
15208 arguments as the old `CONF_*` functions with the exception of the
15209 first that must be a `CONF *` instead of a `LHASH *`.
15211 To make it easier to use the new classes with the old `CONF_*` functions,
15212 the function CONF_set_default_method is provided.
15216 * Add '-tls1' option to 'openssl ciphers', which was already
15217 mentioned in the documentation but had not been implemented.
15218 (This option is not yet really useful because even the additional
15219 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15223 * Initial DSO code added into libcrypto for letting OpenSSL (and
15224 OpenSSL-based applications) load shared libraries and bind to
15225 them in a portable way.
15227 *Geoff Thorpe, with contributions from Richard Levitte*
15229 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15231 * Make sure _lrotl and _lrotr are only used with MSVC.
15233 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15234 (the default implementation of RAND_status).
15236 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15237 to '-clrext' (= clear extensions), as intended and documented.
15238 *Bodo Moeller; inconsistency pointed out by Michael Attili
15239 <attili@amaxo.com>*
15241 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15242 was larger than the MD block size.
15244 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15246 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15247 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15248 using the passed key: if the passed key was a private key the result
15249 of X509_print(), for example, would be to print out all the private key
15254 * des_quad_cksum() byte order bug fix.
15255 *Ulf Möller, using the problem description in krb4-0.9.7, where
15256 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15258 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15261 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15263 * For easily testing in shell scripts whether some command
15264 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15265 returns with exit code 0 iff no command of the given name is available.
15266 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15267 the output goes to stdout and nothing is printed to stderr.
15268 Additional arguments are always ignored.
15270 Since for each cipher there is a command of the same name,
15271 the 'no-cipher' compilation switches can be tested this way.
15273 ('openssl no-XXX' is not able to detect pseudo-commands such
15274 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15278 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15282 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15283 is set; it will be thrown away anyway because each handshake creates
15285 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15286 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15287 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15288 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15292 * New s_client option -ign_eof: EOF at stdin is ignored, and
15293 'Q' and 'R' lose their special meanings (quit/renegotiate).
15294 This is part of what -quiet does; unlike -quiet, -ign_eof
15295 does not suppress any output.
15299 * Add compatibility options to the purpose and trust code. The
15300 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15301 accepts a certificate or CA, this was the previous behaviour,
15302 with all the associated security issues.
15304 X509_TRUST_COMPAT is the old trust behaviour: only and
15305 automatically trust self signed roots in certificate store. A
15306 new trust setting X509_TRUST_DEFAULT is used to specify that
15307 a purpose has no associated trust setting and it should instead
15308 use the value in the default purpose.
15312 * Fix the PKCS#8 DSA private key code so it decodes keys again
15313 and fix a memory leak.
15317 * In util/mkerr.pl (which implements 'make errors'), preserve
15318 reason strings from the previous version of the .c file, as
15319 the default to have only downcase letters (and digits) in
15320 automatically generated reasons codes is not always appropriate.
15324 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15325 using strerror. Previously, ERR_reason_error_string() returned
15326 library names as reason strings for SYSerr; but SYSerr is a special
15327 case where small numbers are errno values, not library numbers.
15331 * Add '-dsaparam' option to 'openssl dhparam' application. This
15332 converts DSA parameters into DH parameters. (When creating parameters,
15333 DSA_generate_parameters is used.)
15337 * Include 'length' (recommended exponent length) in C code generated
15338 by 'openssl dhparam -C'.
15342 * The second argument to set_label in perlasm was already being used
15343 so couldn't be used as a "file scope" flag. Moved to third argument
15348 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15349 instead of RAND_bytes for encryption IVs and salts.
15353 * Include RAND_status() into RAND_METHOD instead of implementing
15354 it only for md_rand.c Otherwise replacing the PRNG by calling
15355 RAND_set_rand_method would be impossible.
15359 * Don't let DSA_generate_key() enter an infinite loop if the random
15360 number generation fails.
15364 * New 'rand' application for creating pseudo-random output.
15368 * Added configuration support for Linux/IA64
15370 *Rolf Haberrecker <rolf@suse.de>*
15372 * Assembler module support for Mingw32.
15376 * Shared library support for HPUX (in shlib/).
15378 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15380 * Shared library support for Solaris gcc.
15382 *Lutz Behnke <behnke@trustcenter.de>*
15384 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15386 * PKCS7_encrypt() was adding text MIME headers twice because they
15387 were added manually and by SMIME_crlf_copy().
15391 * In bntest.c don't call BN_rand with zero bits argument.
15393 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15395 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15396 case was implemented. This caused BN_div_recp() to fail occasionally.
15400 * Add an optional second argument to the set_label() in the perl
15401 assembly language builder. If this argument exists and is set
15402 to 1 it signals that the assembler should use a symbol whose
15403 scope is the entire file, not just the current function. This
15404 is needed with MASM which uses the format label:: for this scope.
15406 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15408 * Change the ASN1 types so they are typedefs by default. Before
15409 almost all types were #define'd to ASN1_STRING which was causing
15410 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15415 * Change names of new functions to the new get1/get0 naming
15416 convention: After 'get1', the caller owns a reference count
15417 and has to call `..._free`; 'get0' returns a pointer to some
15418 data structure without incrementing reference counters.
15419 (Some of the existing 'get' functions increment a reference
15420 counter, some don't.)
15421 Similarly, 'set1' and 'add1' functions increase reference
15422 counters or duplicate objects.
15426 * Allow for the possibility of temp RSA key generation failure:
15427 the code used to assume it always worked and crashed on failure.
15431 * Fix potential buffer overrun problem in BIO_printf().
15432 *Ulf Möller, using public domain code by Patrick Powell; problem
15433 pointed out by David Sacerdote <das33@cornell.edu>*
15435 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15436 RAND_egd() and RAND_status(). In the command line application,
15437 the EGD socket can be specified like a seed file using RANDFILE
15442 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15443 Some CAs (e.g. Verisign) distribute certificates in this form.
15447 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15448 list to exclude them. This means that no special compilation option
15449 is needed to use anonymous DH: it just needs to be included in the
15454 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15455 EVP_MD_type. The old functionality is available in a new macro called
15456 EVP_MD_md(). Change code that uses it and update docs.
15460 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15461 where the `void *` argument is replaced by a function pointer argument.
15462 Previously `void *` was abused to point to functions, which works on
15463 many platforms, but is not correct. As these functions are usually
15464 called by macros defined in OpenSSL header files, most source code
15465 should work without changes.
15469 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15470 sections with information on -D... compiler switches used for
15471 compiling the library so that applications can see them. To enable
15472 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15473 must be defined. E.g.,
15474 #define OPENSSL_ALGORITHM_DEFINES
15475 #include <openssl/opensslconf.h>
15476 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15478 *Richard Levitte, Ulf and Bodo Möller*
15480 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15485 * Change the 'other' type in certificate aux info to a STACK_OF
15486 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15487 the required ASN1 format: arbitrary types determined by an OID.
15491 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15492 argument to 'req'. This is not because the function is newer or
15493 better than others it just uses the work 'NEW' in the certificate
15494 request header lines. Some software needs this.
15498 * Reorganise password command line arguments: now passwords can be
15499 obtained from various sources. Delete the PEM_cb function and make
15500 it the default behaviour: i.e. if the callback is NULL and the
15501 usrdata argument is not NULL interpret it as a null terminated pass
15502 phrase. If usrdata and the callback are NULL then the pass phrase
15503 is prompted for as usual.
15507 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15508 the support is automatically enabled. The resulting binaries will
15509 autodetect the card and use it if present.
15511 *Ben Laurie and Compaq Inc.*
15513 * Work around for Netscape hang bug. This sends certificate request
15514 and server done in one record. Since this is perfectly legal in the
15515 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15516 the bugs/SSLv3 entry for more info.
15520 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15524 * Add -rand argument to smime and pkcs12 applications and read/write
15529 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15533 * Add command line password options to the remaining applications.
15537 * Bug fix for BN_div_recp() for numerators with an even number of
15542 * More tests in bntest.c, and changed test_bn output.
15546 * ./config recognizes MacOS X now.
15550 * Bug fix for BN_div() when the first words of num and divisor are
15551 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15555 * Add support for various broken PKCS#8 formats, and command line
15556 options to produce them.
15560 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15561 get temporary BIGNUMs from a BN_CTX.
15565 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15570 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15571 include a #define from the old name to the new. The original intent
15572 was that statically linked binaries could for example just call
15573 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15574 link with digests. This never worked because SSLeay_add_all_digests()
15575 and SSLeay_add_all_ciphers() were in the same source file so calling
15576 one would link with the other. They are now in separate source files.
15580 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15584 * Use a less unusual form of the Miller-Rabin primality test (it used
15585 a binary algorithm for exponentiation integrated into the Miller-Rabin
15586 loop, our standard modexp algorithms are faster).
15590 * Support for the EBCDIC character set completed.
15592 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15594 * Source code cleanups: use const where appropriate, eliminate casts,
15595 use `void *` instead of `char *` in lhash.
15599 * Bugfix: ssl3_send_server_key_exchange was not restartable
15600 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15601 this the server could overwrite ephemeral keys that the client
15606 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15607 using 50 iterations of the Rabin-Miller test.
15609 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15610 iterations of the Rabin-Miller test as required by the appendix
15611 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15612 As BN_is_prime_fasttest includes trial division, DSA parameter
15613 generation becomes much faster.
15615 This implies a change for the callback functions in DSA_is_prime
15616 and DSA_generate_parameters: The callback function is called once
15617 for each positive witness in the Rabin-Miller test, not just
15618 occasionally in the inner loop; and the parameters to the
15619 callback function now provide an iteration count for the outer
15620 loop rather than for the current invocation of the inner loop.
15621 DSA_generate_parameters additionally can call the callback
15622 function with an 'iteration count' of -1, meaning that a
15623 candidate has passed the trial division test (when q is generated
15624 from an application-provided seed, trial division is skipped).
15628 * New function BN_is_prime_fasttest that optionally does trial
15629 division before starting the Rabin-Miller test and has
15630 an additional BN_CTX * argument (whereas BN_is_prime always
15631 has to allocate at least one BN_CTX).
15632 'callback(1, -1, cb_arg)' is called when a number has passed the
15633 trial division stage.
15637 * Fix for bug in CRL encoding. The validity dates weren't being handled
15642 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15646 * New function BN_pseudo_rand().
15650 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15651 bignum version of BN_from_montgomery() with the working code from
15652 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15657 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15658 made it impossible to use the same SSL_SESSION data structure in
15659 SSL2 clients in multiple threads.
15663 * The return value of RAND_load_file() no longer counts bytes obtained
15664 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15665 to seed the PRNG (previously an explicit byte count was required).
15667 *Ulf Möller, Bodo Möller*
15669 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15670 used `char *` instead of `void *` and had casts all over the place.
15674 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15678 * Retain source code compatibility for BN_prime_checks macro:
15679 BN_is_prime(..., BN_prime_checks, ...) now uses
15680 BN_prime_checks_for_size to determine the appropriate number of
15681 Rabin-Miller iterations.
15685 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15686 DH_CHECK_P_NOT_SAFE_PRIME.
15687 (Check if this is true? OpenPGP calls them "strong".)
15691 * Merge the functionality of "dh" and "gendh" programs into a new program
15692 "dhparam". The old programs are retained for now but will handle DH keys
15693 (instead of parameters) in future.
15697 * Make the ciphers, s_server and s_client programs check the return values
15698 when a new cipher list is set.
15702 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15703 ciphers. Before when the 56bit ciphers were enabled the sorting was
15706 The syntax for the cipher sorting has been extended to support sorting by
15707 cipher-strength (using the strength_bits hard coded in the tables).
15708 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15710 Fix a bug in the cipher-command parser: when supplying a cipher command
15711 string with an "undefined" symbol (neither command nor alphanumeric
15712 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15713 an error is flagged.
15715 Due to the strength-sorting extension, the code of the
15716 ssl_create_cipher_list() function was completely rearranged. I hope that
15717 the readability was also increased :-)
15719 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15721 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15722 for the first serial number and places 2 in the serial number file. This
15723 avoids problems when the root CA is created with serial number zero and
15724 the first user certificate has the same issuer name and serial number
15729 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15730 the new code. Add documentation for this stuff.
15734 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15735 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15736 structures and behave in an analogous way to the X509v3 functions:
15737 they shouldn't be called directly but wrapper functions should be used
15740 So we also now have some wrapper functions that call the X509at functions
15741 when passed certificate requests. (TO DO: similar things can be done with
15742 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15743 things. Some of these need some d2i or i2d and print functionality
15744 because they handle more complex structures.)
15748 * Add missing #ifndefs that caused missing symbols when building libssl
15749 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15750 NO_RSA in `ssl/s2*.c`.
15752 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15754 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15755 has a return value which indicates the quality of the random data
15756 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15757 error queue. New function RAND_pseudo_bytes() generates output that is
15758 guaranteed to be unique but not unpredictable. RAND_add is like
15759 RAND_seed, but takes an extra argument for an entropy estimate
15760 (RAND_seed always assumes full entropy).
15764 * Do more iterations of Rabin-Miller probable prime test (specifically,
15765 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15766 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15767 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15768 false-positive rate of at most 2^-80 for random input.
15772 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15776 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15777 in the 0.9.5 release), this returns the chain
15778 from an X509_CTX structure with a dup of the stack and all
15779 the X509 reference counts upped: so the stack will exist
15780 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15783 Also make SSL_SESSION_print() print out the verify return
15788 * Add manpage for the pkcs12 command. Also change the default
15789 behaviour so MAC iteration counts are used unless the new
15790 -nomaciter option is used. This improves file security and
15791 only older versions of MSIE (4.0 for example) need it.
15795 * Honor the no-xxx Configure options when creating .DEF files.
15799 * Add PKCS#10 attributes to field table: challengePassword,
15800 unstructuredName and unstructuredAddress. These are taken from
15801 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15802 international characters are used.
15804 More changes to X509_ATTRIBUTE code: allow the setting of types
15805 based on strings. Remove the 'loc' parameter when adding
15806 attributes because these will be a SET OF encoding which is sorted
15811 * Initial changes to the 'req' utility to allow request generation
15812 automation. This will allow an application to just generate a template
15813 file containing all the field values and have req construct the
15816 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15817 used all over the place including certificate requests and PKCS#7
15818 structures. They are currently handled manually where necessary with
15819 some primitive wrappers for PKCS#7. The new functions behave in a
15820 manner analogous to the X509 extension functions: they allow
15821 attributes to be looked up by NID and added.
15823 Later something similar to the X509V3 code would be desirable to
15824 automatically handle the encoding, decoding and printing of the
15825 more complex types. The string types like challengePassword can
15826 be handled by the string table functions.
15828 Also modified the multi byte string table handling. Now there is
15829 a 'global mask' which masks out certain types. The table itself
15830 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15831 is useful when for example there is only one permissible type
15832 (as in countryName) and using the mask might result in no valid
15837 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15838 SSL_get_peer_finished to allow applications to obtain the latest
15839 Finished messages sent to the peer or expected from the peer,
15840 respectively. (SSL_get_peer_finished is usually the Finished message
15841 actually received from the peer, otherwise the protocol will be aborted.)
15843 As the Finished message are message digests of the complete handshake
15844 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15845 be used for external authentication procedures when the authentication
15846 provided by SSL/TLS is not desired or is not enough.
15850 * Enhanced support for Alpha Linux is added. Now ./config checks if
15851 the host supports BWX extension and if Compaq C is present on the
15852 $PATH. Just exploiting of the BWX extension results in 20-30%
15853 performance kick for some algorithms, e.g. DES and RC4 to mention
15854 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15859 * Add support for MS "fast SGC". This is arguably a violation of the
15860 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15861 weak crypto and after checking the certificate is SGC a second one
15862 with strong crypto. MS SGC stops the first handshake after receiving
15863 the server certificate message and sends a second client hello. Since
15864 a server will typically do all the time consuming operations before
15865 expecting any further messages from the client (server key exchange
15866 is the most expensive) there is little difference between the two.
15868 To get OpenSSL to support MS SGC we have to permit a second client
15869 hello message after we have sent server done. In addition we have to
15870 reset the MAC if we do get this second client hello.
15874 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15875 if a DER encoded private key is RSA or DSA traditional format. Changed
15876 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15877 format DER encoded private key. Newer code should use PKCS#8 format which
15878 has the key type encoded in the ASN1 structure. Added DER private key
15879 support to pkcs8 application.
15883 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15884 ciphersuites has been selected (as required by the SSL 3/TLS 1
15885 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15886 is set, we interpret this as a request to violate the specification
15887 (the worst that can happen is a handshake failure, and 'correct'
15888 behaviour would result in a handshake failure anyway).
15892 * In SSL_CTX_add_session, take into account that there might be multiple
15893 SSL_SESSION structures with the same session ID (e.g. when two threads
15894 concurrently obtain them from an external cache).
15895 The internal cache can handle only one SSL_SESSION with a given ID,
15896 so if there's a conflict, we now throw out the old one to achieve
15901 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15902 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15903 some routines that use cipher OIDs: some ciphers do not have OIDs
15904 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15909 * Simplify the trust setting structure and code. Now we just have
15910 two sequences of OIDs for trusted and rejected settings. These will
15911 typically have values the same as the extended key usage extension
15912 and any application specific purposes.
15914 The trust checking code now has a default behaviour: it will just
15915 check for an object with the same NID as the passed id. Functions can
15916 be provided to override either the default behaviour or the behaviour
15917 for a given id. SSL client, server and email already have functions
15918 in place for compatibility: they check the NID and also return "trusted"
15919 if the certificate is self signed.
15923 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15924 traditional format into an EVP_PKEY structure.
15928 * Add a password callback function PEM_cb() which either prompts for
15929 a password if usr_data is NULL or otherwise assumes it is a null
15930 terminated password. Allow passwords to be passed on command line
15931 environment or config files in a few more utilities.
15935 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15936 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15937 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15938 Update documentation.
15942 * Support for ASN1 "NULL" type. This could be handled before by using
15943 ASN1_TYPE but there wasn't any function that would try to read a NULL
15944 and produce an error if it couldn't. For compatibility we also have
15945 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15946 don't allocate anything because they don't need to.
15950 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15953 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15955 * Rebuild of the memory allocation routines used by OpenSSL code and
15956 possibly others as well. The purpose is to make an interface that
15957 provide hooks so anyone can build a separate set of allocation and
15958 deallocation routines to be used by OpenSSL, for example memory
15959 pool implementations, or something else, which was previously hard
15960 since Malloc(), Realloc() and Free() were defined as macros having
15961 the values malloc, realloc and free, respectively (except for Win32
15962 compilations). The same is provided for memory debugging code.
15963 OpenSSL already comes with functionality to find memory leaks, but
15964 this gives people a chance to debug other memory problems.
15966 With these changes, a new set of functions and macros have appeared:
15968 CRYPTO_set_mem_debug_functions() [F]
15969 CRYPTO_get_mem_debug_functions() [F]
15970 CRYPTO_dbg_set_options() [F]
15971 CRYPTO_dbg_get_options() [F]
15972 CRYPTO_malloc_debug_init() [M]
15974 The memory debug functions are NULL by default, unless the library
15975 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15976 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15977 gives the standard debugging functions that come with OpenSSL) or
15978 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15979 provided by the library user) must be used. When the standard
15980 debugging functions are used, CRYPTO_dbg_set_options can be used to
15981 request additional information:
15982 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15983 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15985 Also, things like CRYPTO_set_mem_functions will always give the
15986 expected result (the new set of functions is used for allocation
15987 and deallocation) at all times, regardless of platform and compiler
15990 To finish it up, some functions that were never use in any other
15991 way than through macros have a new API and new semantic:
15993 CRYPTO_dbg_malloc()
15994 CRYPTO_dbg_realloc()
15997 All macros of value have retained their old syntax.
15999 *Richard Levitte and Bodo Moeller*
16001 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16002 ordering of SMIMECapabilities wasn't in "strength order" and there
16003 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16008 * Some ASN1 types with illegal zero length encoding (INTEGER,
16009 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16011 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16013 * Merge in my S/MIME library for OpenSSL. This provides a simple
16014 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16015 functionality to handle multipart/signed properly) and a utility
16016 called 'smime' to call all this stuff. This is based on code I
16017 originally wrote for Celo who have kindly allowed it to be
16018 included in OpenSSL.
16022 * Add variants des_set_key_checked and des_set_key_unchecked of
16023 des_set_key (aka des_key_sched). Global variable des_check_key
16024 decides which of these is called by des_set_key; this way
16025 des_check_key behaves as it always did, but applications and
16026 the library itself, which was buggy for des_check_key == 1,
16027 have a cleaner way to pick the version they need.
16031 * New function PKCS12_newpass() which changes the password of a
16036 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16037 dynamic mix. In both cases the ids can be used as an index into the
16038 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16039 functions so they accept a list of the field values and the
16040 application doesn't need to directly manipulate the X509_TRUST
16045 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16050 * Modify the way the V3 extension code looks up extensions. This now
16051 works in a similar way to the object code: we have some "standard"
16052 extensions in a static table which is searched with OBJ_bsearch()
16053 and the application can add dynamic ones if needed. The file
16054 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16055 updated whenever a new extension is added to the core code and kept
16056 in ext_nid order. There is a simple program 'tabtest.c' which checks
16057 this. New extensions are not added too often so this file can readily
16058 be maintained manually.
16060 There are two big advantages in doing things this way. The extensions
16061 can be looked up immediately and no longer need to be "added" using
16062 X509V3_add_standard_extensions(): this function now does nothing.
16063 Side note: I get *lots* of email saying the extension code doesn't
16064 work because people forget to call this function.
16065 Also no dynamic allocation is done unless new extensions are added:
16066 so if we don't add custom extensions there is no need to call
16067 X509V3_EXT_cleanup().
16071 * Modify enc utility's salting as follows: make salting the default. Add a
16072 magic header, so unsalted files fail gracefully instead of just decrypting
16073 to garbage. This is because not salting is a big security hole, so people
16074 should be discouraged from doing it.
16078 * Fixes and enhancements to the 'x509' utility. It allowed a message
16079 digest to be passed on the command line but it only used this
16080 parameter when signing a certificate. Modified so all relevant
16081 operations are affected by the digest parameter including the
16082 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16083 DSA key was used because it didn't fix the digest.
16087 * Initial certificate chain verify code. Currently tests the untrusted
16088 certificates for consistency with the verify purpose (which is set
16089 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16091 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16092 this is because it will reject chains with invalid extensions whereas
16093 every previous version of OpenSSL and SSLeay made no checks at all.
16095 Trust code: checks the root CA for the relevant trust settings. Trust
16096 settings have an initial value consistent with the verify purpose: e.g.
16097 if the verify purpose is for SSL client use it expects the CA to be
16098 trusted for SSL client use. However the default value can be changed to
16099 permit custom trust settings: one example of this would be to only trust
16100 certificates from a specific "secure" set of CAs.
16102 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16103 which should be used for version portability: especially since the
16104 verify structure is likely to change more often now.
16106 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16107 to set them. If not set then assume SSL clients will verify SSL servers
16110 Two new options to the verify program: -untrusted allows a set of
16111 untrusted certificates to be passed in and -purpose which sets the
16112 intended purpose of the certificate. If a purpose is set then the
16113 new chain verify code is used to check extension consistency.
16117 * Support for the authority information access extension.
16121 * Modify RSA and DSA PEM read routines to transparently handle
16122 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16123 public keys in a format compatible with certificate
16124 SubjectPublicKeyInfo structures. Unfortunately there were already
16125 functions called *_PublicKey_* which used various odd formats so
16126 these are retained for compatibility: however the DSA variants were
16127 never in a public release so they have been deleted. Changed dsa/rsa
16128 utilities to handle the new format: note no releases ever handled public
16129 keys so we should be OK.
16131 The primary motivation for this change is to avoid the same fiasco
16132 that dogs private keys: there are several incompatible private key
16133 formats some of which are standard and some OpenSSL specific and
16134 require various evil hacks to allow partial transparent handling and
16135 even then it doesn't work with DER formats. Given the option anything
16136 other than PKCS#8 should be dumped: but the other formats have to
16137 stay in the name of compatibility.
16139 With public keys and the benefit of hindsight one standard format
16140 is used which works with EVP_PKEY, RSA or DSA structures: though
16141 it clearly returns an error if you try to read the wrong kind of key.
16143 Added a -pubkey option to the 'x509' utility to output the public key.
16144 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16145 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16146 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16147 that do the same as the `EVP_PKEY_assign_*()` except they up the
16148 reference count of the added key (they don't "swallow" the
16153 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16154 CRLs would fail if the file contained no certificates or no CRLs:
16155 added a new function to read in both types and return the number
16156 read: this means that if none are read it will be an error. The
16157 DER versions of the certificate and CRL reader would always fail
16158 because it isn't possible to mix certificates and CRLs in DER format
16159 without choking one or the other routine. Changed this to just read
16160 a certificate: this is the best we can do. Also modified the code
16161 in apps/verify.c to take notice of return codes: it was previously
16162 attempting to read in certificates from NULL pointers and ignoring
16163 any errors: this is one reason why the cert and CRL reader seemed
16164 to work. It doesn't check return codes from the default certificate
16165 routines: these may well fail if the certificates aren't installed.
16169 * Code to support otherName option in GeneralName.
16173 * First update to verify code. Change the verify utility
16174 so it warns if it is passed a self signed certificate:
16175 for consistency with the normal behaviour. X509_verify
16176 has been modified to it will now verify a self signed
16177 certificate if *exactly* the same certificate appears
16178 in the store: it was previously impossible to trust a
16179 single self signed certificate. This means that:
16180 openssl verify ss.pem
16181 now gives a warning about a self signed certificate but
16182 openssl verify -CAfile ss.pem ss.pem
16187 * For servers, store verify_result in SSL_SESSION data structure
16188 (and add it to external session representation).
16189 This is needed when client certificate verifications fails,
16190 but an application-provided verification callback (set by
16191 SSL_CTX_set_cert_verify_callback) allows accepting the session
16192 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16193 but returns 1): When the session is reused, we have to set
16194 ssl->verify_result to the appropriate error code to avoid
16197 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16199 * Fix a bug in the new PKCS#7 code: it didn't consider the
16200 case in PKCS7_dataInit() where the signed PKCS7 structure
16201 didn't contain any existing data because it was being created.
16203 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16205 * Add a salt to the key derivation routines in enc.c. This
16206 forms the first 8 bytes of the encrypted file. Also add a
16207 -S option to allow a salt to be input on the command line.
16211 * New function X509_cmp(). Oddly enough there wasn't a function
16212 to compare two certificates. We do this by working out the SHA1
16213 hash and comparing that. X509_cmp() will be needed by the trust
16218 * SSL_get1_session() is like SSL_get_session(), but increments
16219 the reference count in the SSL_SESSION returned.
16221 *Geoff Thorpe <geoff@eu.c2.net>*
16223 * Fix for 'req': it was adding a null to request attributes.
16224 Also change the X509_LOOKUP and X509_INFO code to handle
16225 certificate auxiliary information.
16229 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16234 * Add the possibility to add extra information to the memory leak
16235 detecting output, to form tracebacks, showing from where each
16236 allocation was originated: CRYPTO_push_info("constant string") adds
16237 the string plus current file name and line number to a per-thread
16238 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16239 is like calling CYRPTO_pop_info() until the stack is empty.
16240 Also updated memory leak detection code to be multi-thread-safe.
16244 * Add options -text and -noout to pkcs7 utility and delete the
16245 encryption options which never did anything. Update docs.
16249 * Add options to some of the utilities to allow the pass phrase
16250 to be included on either the command line (not recommended on
16251 OSes like Unix) or read from the environment. Update the
16252 manpages and fix a few bugs.
16256 * Add a few manpages for some of the openssl commands.
16260 * Fix the -revoke option in ca. It was freeing up memory twice,
16261 leaking and not finding already revoked certificates.
16265 * Extensive changes to support certificate auxiliary information.
16266 This involves the use of X509_CERT_AUX structure and X509_AUX
16267 functions. An X509_AUX function such as PEM_read_X509_AUX()
16268 can still read in a certificate file in the usual way but it
16269 will also read in any additional "auxiliary information". By
16270 doing things this way a fair degree of compatibility can be
16271 retained: existing certificates can have this information added
16272 using the new 'x509' options.
16274 Current auxiliary information includes an "alias" and some trust
16275 settings. The trust settings will ultimately be used in enhanced
16276 certificate chain verification routines: currently a certificate
16277 can only be trusted if it is self signed and then it is trusted
16282 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16283 The problem was that one of the replacement routines had not been working
16284 since SSLeay releases. For now the offending routine has been replaced
16285 with non-optimised assembler. Even so, this now gives around 95%
16286 performance improvement for 1024 bit RSA signs.
16290 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16291 handling. Most clients have the effective key size in bits equal to
16292 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16293 A few however don't do this and instead use the size of the decrypted key
16294 to determine the RC2 key length and the AlgorithmIdentifier to determine
16295 the effective key length. In this case the effective key length can still
16296 be 40 bits but the key length can be 168 bits for example. This is fixed
16297 by manually forcing an RC2 key into the EVP_PKEY structure because the
16298 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16299 the key length and effective key length are equal.
16303 * Add a bunch of functions that should simplify the creation of
16304 X509_NAME structures. Now you should be able to do:
16305 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16306 and have it automatically work out the correct field type and fill in
16307 the structures. The more adventurous can try:
16308 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16309 and it will (hopefully) work out the correct multibyte encoding.
16313 * Change the 'req' utility to use the new field handling and multibyte
16314 copy routines. Before the DN field creation was handled in an ad hoc
16315 way in req, ca, and x509 which was rather broken and didn't support
16316 BMPStrings or UTF8Strings. Since some software doesn't implement
16317 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16318 using the dirstring_type option. See the new comment in the default
16319 openssl.cnf for more info.
16323 * Make crypto/rand/md_rand.c more robust:
16324 - Assure unique random numbers after fork().
16325 - Make sure that concurrent threads access the global counter and
16326 md serializably so that we never lose entropy in them
16327 or use exactly the same state in multiple threads.
16328 Access to the large state is not always serializable because
16329 the additional locking could be a performance killer, and
16330 md should be large enough anyway.
16334 * New file apps/app_rand.c with commonly needed functionality
16335 for handling the random seed file.
16337 Use the random seed file in some applications that previously did not:
16339 dsaparam -genkey (which also ignored its '-rand' option),
16342 x509 (when signing).
16343 Except on systems with /dev/urandom, it is crucial to have a random
16344 seed file at least for key creation, DSA signing, and for DH exchanges;
16345 for RSA signatures we could do without one.
16347 gendh and gendsa (unlike genrsa) used to read only the first byte
16348 of each file listed in the '-rand' option. The function as previously
16349 found in genrsa is now in app_rand.c and is used by all programs
16350 that support '-rand'.
16354 * In RAND_write_file, use mode 0600 for creating files;
16355 don't just chmod when it may be too late.
16359 * Report an error from X509_STORE_load_locations
16360 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16364 * New function ASN1_mbstring_copy() this copies a string in either
16365 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16366 into an ASN1_STRING type. A mask of permissible types is passed
16367 and it chooses the "minimal" type to use or an error if not type
16372 * Add function equivalents to the various macros in asn1.h. The old
16373 macros are retained with an `M_` prefix. Code inside the library can
16374 use the `M_` macros. External code (including the openssl utility)
16375 should *NOT* in order to be "shared library friendly".
16379 * Add various functions that can check a certificate's extensions
16380 to see if it usable for various purposes such as SSL client,
16381 server or S/MIME and CAs of these types. This is currently
16382 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16383 verification. Also added a -purpose flag to x509 utility to
16384 print out all the purposes.
16388 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16393 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16394 for, obtain and decode and extension and obtain its critical flag.
16395 This allows all the necessary extension code to be handled in a
16396 single function call.
16400 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16401 platforms. See crypto/rc4/rc4_enc.c for further details.
16405 * New -noout option to asn1parse. This causes no output to be produced
16406 its main use is when combined with -strparse and -out to extract data
16407 from a file (which may not be in ASN.1 format).
16411 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16412 when producing the local key id.
16414 *Richard Levitte <levitte@stacken.kth.se>*
16416 * New option -dhparam in s_server. This allows a DH parameter file to be
16417 stated explicitly. If it is not stated then it tries the first server
16418 certificate file. The previous behaviour hard coded the filename
16423 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16424 a public key to be input or output. For example:
16425 openssl rsa -in key.pem -pubout -out pubkey.pem
16426 Also added necessary DSA public key functions to handle this.
16430 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16431 in the message. This was handled by allowing
16432 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16434 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16436 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16437 to the end of the strings whereas this didn't. This would cause problems
16438 if strings read with d2i_ASN1_bytes() were later modified.
16440 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16442 * Fix for base64 decode bug. When a base64 bio reads only one line of
16443 data and it contains EOF it will end up returning an error. This is
16444 caused by input 46 bytes long. The cause is due to the way base64
16445 BIOs find the start of base64 encoded data. They do this by trying a
16446 trial decode on each line until they find one that works. When they
16447 do a flag is set and it starts again knowing it can pass all the
16448 data directly through the decoder. Unfortunately it doesn't reset
16449 the context it uses. This means that if EOF is reached an attempt
16450 is made to pass two EOFs through the context and this causes the
16451 resulting error. This can also cause other problems as well. As is
16452 usual with these problems it takes *ages* to find and the fix is
16453 trivial: move one line.
16455 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16457 * Ugly workaround to get s_client and s_server working under Windows. The
16458 old code wouldn't work because it needed to select() on sockets and the
16459 tty (for keypresses and to see if data could be written). Win32 only
16460 supports select() on sockets so we select() with a 1s timeout on the
16461 sockets and then see if any characters are waiting to be read, if none
16462 are present then we retry, we also assume we can always write data to
16463 the tty. This isn't nice because the code then blocks until we've
16464 received a complete line of data and it is effectively polling the
16465 keyboard at 1s intervals: however it's quite a bit better than not
16466 working at all :-) A dedicated Windows application might handle this
16467 with an event loop for example.
16471 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16472 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16473 will be called when RSA_sign() and RSA_verify() are used. This is useful
16474 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16475 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16476 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16477 This necessitated the support of an extra signature type NID_md5_sha1
16478 for SSL signatures and modifications to the SSL library to use it instead
16479 of calling RSA_public_decrypt() and RSA_private_encrypt().
16483 * Add new -verify -CAfile and -CApath options to the crl program, these
16484 will lookup a CRL issuers certificate and verify the signature in a
16485 similar way to the verify program. Tidy up the crl program so it
16486 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16487 less strict. It will now permit CRL extensions even if it is not
16488 a V2 CRL: this will allow it to tolerate some broken CRLs.
16492 * Initialize all non-automatic variables each time one of the openssl
16493 sub-programs is started (this is necessary as they may be started
16494 multiple times from the "OpenSSL>" prompt).
16496 *Lennart Bang, Bodo Moeller*
16498 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16499 removing all other RSA functionality (this is what NO_RSA does). This
16500 is so (for example) those in the US can disable those operations covered
16501 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16506 * Non-copying interface to BIO pairs.
16507 (still largely untested)
16511 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16512 ASCII string. This was handled independently in various places before.
16516 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16517 UTF8 strings a character at a time.
16521 * Use client_version from client hello to select the protocol
16522 (s23_srvr.c) and for RSA client key exchange verification
16523 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16527 * Add various utility functions to handle SPKACs, these were previously
16528 handled by poking round in the structure internals. Added new function
16529 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16530 print, verify and generate SPKACs. Based on an original idea from
16531 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16535 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16539 * Allow the config file extension section to be overwritten on the
16540 command line. Based on an original idea from Massimiliano Pala
16541 <madwolf@comune.modena.it>. The new option is called -extensions
16542 and can be applied to ca, req and x509. Also -reqexts to override
16543 the request extensions in req and -crlexts to override the crl extensions
16548 * Add new feature to the SPKAC handling in ca. Now you can include
16549 the same field multiple times by preceding it by "XXXX." for example:
16552 this is the same syntax as used in the req config file.
16556 * Allow certificate extensions to be added to certificate requests. These
16557 are specified in a 'req_extensions' option of the req section of the
16558 config file. They can be printed out with the -text option to req but
16559 are otherwise ignored at present.
16563 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16564 data read consists of only the final block it would not decrypted because
16565 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16566 A misplaced 'break' also meant the decrypted final block might not be
16567 copied until the next read.
16571 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16572 a few extra parameters to the DH structure: these will be useful if
16573 for example we want the value of 'q' or implement X9.42 DH.
16577 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16578 provides hooks that allow the default DSA functions or functions on a
16579 "per key" basis to be replaced. This allows hardware acceleration and
16580 hardware key storage to be handled without major modification to the
16581 library. Also added low level modexp hooks and CRYPTO_EX structure and
16582 associated functions.
16586 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16587 as "read only": it can't be written to and the buffer it points to will
16588 not be freed. Reading from a read only BIO is much more efficient than
16589 a normal memory BIO. This was added because there are several times when
16590 an area of memory needs to be read from a BIO. The previous method was
16591 to create a memory BIO and write the data to it, this results in two
16592 copies of the data and an O(n^2) reading algorithm. There is a new
16593 function BIO_new_mem_buf() which creates a read only memory BIO from
16594 an area of memory. Also modified the PKCS#7 routines to use read only
16599 * Bugfix: ssl23_get_client_hello did not work properly when called in
16600 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16601 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16602 but a retry condition occurred while trying to read the rest.
16606 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16607 NID_pkcs7_encrypted by default: this was wrong since this should almost
16608 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16609 the encrypted data type: this is a more sensible place to put it and it
16610 allows the PKCS#12 code to be tidied up that duplicated this
16615 * Changed obj_dat.pl script so it takes its input and output files on
16616 the command line. This should avoid shell escape redirection problems
16621 * Initial support for certificate extension requests, these are included
16622 in things like Xenroll certificate requests. Included functions to allow
16623 extensions to be obtained and added.
16627 * -crlf option to s_client and s_server for sending newlines as
16628 CRLF (as required by many protocols).
16632 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16634 * Install libRSAglue.a when OpenSSL is built with RSAref.
16636 *Ralf S. Engelschall*
16638 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16640 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16642 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16647 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16648 DH parameters/keys (q is lost during that conversion, but the resulting
16649 DH parameters contain its length).
16651 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16652 much faster than DH_generate_parameters (which creates parameters
16653 where `p = 2*q + 1`), and also the smaller q makes DH computations
16654 much more efficient (160-bit exponentiation instead of 1024-bit
16655 exponentiation); so this provides a convenient way to support DHE
16656 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16657 utter importance to use
16658 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16660 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16661 when such DH parameters are used, because otherwise small subgroup
16662 attacks may become possible!
16666 * Avoid memory leak in i2d_DHparams.
16670 * Allow the -k option to be used more than once in the enc program:
16671 this allows the same encrypted message to be read by multiple recipients.
16675 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16676 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16677 it will always use the numerical form of the OID, even if it has a short
16682 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16683 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16684 otherwise bn_mod_exp was called. In the case of hardware keys for example
16685 no private key components need be present and it might store extra data
16686 in the RSA structure, which cannot be accessed from bn_mod_exp.
16687 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16688 private key operations.
16692 * Added support for SPARC Linux.
16696 * pem_password_cb function type incompatibly changed from
16697 typedef int pem_password_cb(char *buf, int size, int rwflag);
16699 ....(char *buf, int size, int rwflag, void *userdata);
16700 so that applications can pass data to their callbacks:
16701 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16702 additional void * argument, which is just handed through whenever
16703 the password callback is called.
16705 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16707 New function SSL_CTX_set_default_passwd_cb_userdata.
16709 Compatibility note: As many C implementations push function arguments
16710 onto the stack in reverse order, the new library version is likely to
16711 interoperate with programs that have been compiled with the old
16712 pem_password_cb definition (PEM_whatever takes some data that
16713 happens to be on the stack as its last argument, and the callback
16714 just ignores this garbage); but there is no guarantee whatsoever that
16717 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16718 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16719 problems not only on Windows, but also on some Unix platforms.
16720 To avoid problematic command lines, these definitions are now in an
16721 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16722 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16726 * MIPS III/IV assembler module is reimplemented.
16730 * More DES library cleanups: remove references to srand/rand and
16731 delete an unused file.
16735 * Add support for the free Netwide assembler (NASM) under Win32,
16736 since not many people have MASM (ml) and it can be hard to obtain.
16737 This is currently experimental but it seems to work OK and pass all
16738 the tests. Check out INSTALL.W32 for info.
16742 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16743 without temporary keys kept an extra copy of the server key,
16744 and connections with temporary keys did not free everything in case
16749 * New function RSA_check_key and new openssl rsa option -check
16750 for verifying the consistency of RSA keys.
16752 *Ulf Moeller, Bodo Moeller*
16754 * Various changes to make Win32 compile work:
16755 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16756 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16757 comparison" warnings.
16758 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16762 * Add a debugging option to PKCS#5 v2 key generation function: when
16763 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16764 derived keys are printed to stderr.
16768 * Copy the flags in ASN1_STRING_dup().
16770 *Roman E. Pavlov <pre@mo.msk.ru>*
16772 * The x509 application mishandled signing requests containing DSA
16773 keys when the signing key was also DSA and the parameters didn't match.
16775 It was supposed to omit the parameters when they matched the signing key:
16776 the verifying software was then supposed to automatically use the CA's
16777 parameters if they were absent from the end user certificate.
16779 Omitting parameters is no longer recommended. The test was also
16780 the wrong way round! This was probably due to unusual behaviour in
16781 EVP_cmp_parameters() which returns 1 if the parameters match.
16782 This meant that parameters were omitted when they *didn't* match and
16783 the certificate was useless. Certificates signed with 'ca' didn't have
16786 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16788 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16789 The interface is as follows:
16790 Applications can use
16791 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16792 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16793 "off" is now the default.
16794 The library internally uses
16795 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16796 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16797 to disable memory-checking temporarily.
16799 Some inconsistent states that previously were possible (and were
16800 even the default) are now avoided.
16802 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16803 with each memory chunk allocated; this is occasionally more helpful
16804 than just having a counter.
16806 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16808 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16813 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16814 which largely parallels "options", but is for changing API behaviour,
16815 whereas "options" are about protocol behaviour.
16816 Initial "mode" flags are:
16818 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16819 a single record has been written.
16820 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16821 retries use the same buffer location.
16822 (But all of the contents must be
16827 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16830 * Fix problems with no-hmac etc.
16832 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16834 * New functions RSA_get_default_method(), RSA_set_method() and
16835 RSA_get_method(). These allows replacement of RSA_METHODs without having
16836 to mess around with the internals of an RSA structure.
16840 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16841 Also really enable memory leak checks in openssl.c and in some
16844 *Chad C. Mulligan, Bodo Moeller*
16846 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16847 up the length of negative integers. This has now been simplified to just
16848 store the length when it is first determined and use it later, rather
16849 than trying to keep track of where data is copied and updating it to
16851 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16853 * Add a new function PKCS7_signatureVerify. This allows the verification
16854 of a PKCS#7 signature but with the signing certificate passed to the
16855 function itself. This contrasts with PKCS7_dataVerify which assumes the
16856 certificate is present in the PKCS#7 structure. This isn't always the
16857 case: certificates can be omitted from a PKCS#7 structure and be
16858 distributed by "out of band" means (such as a certificate database).
16862 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16863 function prototypes in pem.h, also change util/mkdef.pl to add the
16864 necessary function names.
16868 * mk1mf.pl (used by Windows builds) did not properly read the
16869 options set by Configure in the top level Makefile, and Configure
16870 was not even able to write more than one option correctly.
16871 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16875 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16876 file to be loaded from a BIO or FILE pointer. The BIO version will
16877 for example allow memory BIOs to contain config info.
16881 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16882 Whoever hopes to achieve shared-library compatibility across versions
16883 must use this, not the compile-time macro.
16884 (Exercise 0.9.4: Which is the minimum library version required by
16886 Note: All this applies only to multi-threaded programs, others don't
16891 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16892 through a BIO pair triggered the default case, i.e.
16893 SSLerr(...,SSL_R_UNKNOWN_STATE).
16897 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16898 can use the SSL library even if none of the specific BIOs is
16903 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16904 for the encoded length.
16906 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16908 * Add initial documentation of the X509V3 functions.
16912 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16913 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16914 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16915 secure PKCS#8 private key format with a high iteration count.
16919 * Fix determination of Perl interpreter: A perl or perl5
16920 *directory* in $PATH was also accepted as the interpreter.
16922 *Ralf S. Engelschall*
16924 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16925 wrong with it but it was very old and did things like calling
16926 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16927 unusual formatting.
16931 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16932 to use the new extension code.
16936 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16937 with macros. This should make it easier to change their form, add extra
16938 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16943 * Add to configuration table a new entry that can specify an alternative
16944 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16945 according to Mark Crispin <MRC@Panda.COM>.
16949 * DES CBC did not update the IV. Weird.
16953 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16954 Changing the behaviour of the former might break existing programs --
16955 where IV updating is needed, des_ncbc_encrypt can be used.
16958 * When bntest is run from "make test" it drives bc to check its
16959 calculations, as well as internally checking them. If an internal check
16960 fails, it needs to cause bc to give a non-zero result or make test carries
16961 on without noticing the failure. Fixed.
16965 * DES library cleanups.
16969 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16970 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16971 ciphers. NOTE: although the key derivation function has been verified
16972 against some published test vectors it has not been extensively tested
16973 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16978 * Instead of "mkdir -p", which is not fully portable, use new
16979 Perl script "util/mkdir-p.pl".
16983 * Rewrite the way password based encryption (PBE) is handled. It used to
16984 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16985 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16986 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16987 the 'parameter' field of the AlgorithmIdentifier is passed to the
16988 underlying key generation function so it must do its own ASN1 parsing.
16989 This has also changed the EVP_PBE_CipherInit() function which now has a
16990 'parameter' argument instead of literal salt and iteration count values
16991 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16995 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16996 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16997 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16998 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16999 value was just used as a "magic string" and not used directly its
17000 value doesn't matter.
17004 * Introduce some semblance of const correctness to BN. Shame C doesn't
17009 * "linux-sparc64" configuration (ultrapenguin).
17011 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17012 "linux-sparc" configuration.
17014 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17016 * config now generates no-xxx options for missing ciphers.
17020 * Support the EBCDIC character set (work in progress).
17021 File ebcdic.c not yet included because it has a different license.
17023 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17025 * Support BS2000/OSD-POSIX.
17027 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17029 * Make callbacks for key generation use `void *` instead of `char *`.
17033 * Make S/MIME samples compile (not yet tested).
17037 * Additional typesafe stacks.
17041 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17045 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17047 * New configuration variant "sco5-gcc".
17049 * Updated some demos.
17051 *Sean O Riordain, Wade Scholine*
17053 * Add missing BIO_free at exit of pkcs12 application.
17057 * Fix memory leak in conf.c.
17061 * Updates for Win32 to assembler version of MD5.
17065 * Set #! path to perl in apps/der_chop to where we found it
17066 instead of using a fixed path.
17070 * SHA library changes for irix64-mips4-cc.
17074 * Improvements for VMS support.
17078 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17080 * Bignum library bug fix. IRIX 6 passes "make test" now!
17081 This also avoids the problems with SC4.2 and unpatched SC5.
17083 *Andy Polyakov <appro@fy.chalmers.se>*
17085 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17086 These are required because of the typesafe stack would otherwise break
17087 existing code. If old code used a structure member which used to be STACK
17088 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17089 sk_num or sk_value it would produce an error because the num, data members
17090 are not present in STACK_OF. Now it just produces a warning. sk_set
17091 replaces the old method of assigning a value to sk_value
17092 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17093 that does this will no longer work (and should use sk_set instead) but
17094 this could be regarded as a "questionable" behaviour anyway.
17098 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17099 correctly handle encrypted S/MIME data.
17103 * Change type of various DES function arguments from des_cblock
17104 (which means, in function argument declarations, pointer to char)
17105 to des_cblock * (meaning pointer to array with 8 char elements),
17106 which allows the compiler to do more typechecking; it was like
17107 that back in SSLeay, but with lots of ugly casts.
17109 Introduce new type const_des_cblock.
17113 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17114 problems: find RecipientInfo structure that matches recipient certificate
17115 and initialise the ASN1 structures properly based on passed cipher.
17119 * Belatedly make the BN tests actually check the results.
17123 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17124 to and from BNs: it was completely broken. New compilation option
17125 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17126 key elements as negative integers.
17130 * Reorganize and speed up MD5.
17132 *Andy Polyakov <appro@fy.chalmers.se>*
17136 *Richard Levitte <richard@levitte.org>*
17138 * New option -out to asn1parse to allow the parsed structure to be
17139 output to a file. This is most useful when combined with the -strparse
17140 option to examine the output of things like OCTET STRINGS.
17144 * Make SSL library a little more fool-proof by not requiring any longer
17145 that `SSL_set_{accept,connect}_state` be called before
17146 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17147 in many applications because usually everything *appeared* to work as
17148 intended anyway -- now it really works as intended).
17152 * Move openssl.cnf out of lib/.
17156 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17157 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17158 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17160 *Ralf S. Engelschall*
17162 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17163 handle PKCS#7 enveloped data properly.
17165 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17167 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17168 copying pointers. The cert_st handling is changed by this in
17169 various ways (and thus what used to be known as ctx->default_cert
17170 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17171 any longer when s->cert does not give us what we need).
17172 ssl_cert_instantiate becomes obsolete by this change.
17173 As soon as we've got the new code right (possibly it already is?),
17174 we have solved a couple of bugs of the earlier code where s->cert
17175 was used as if it could not have been shared with other SSL structures.
17177 Note that using the SSL API in certain dirty ways now will result
17178 in different behaviour than observed with earlier library versions:
17179 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17180 does not influence s as it used to.
17182 In order to clean up things more thoroughly, inside SSL_SESSION
17183 we don't use CERT any longer, but a new structure SESS_CERT
17184 that holds per-session data (if available); currently, this is
17185 the peer's certificate chain and, for clients, the server's certificate
17186 and temporary key. CERT holds only those values that can have
17187 meaningful defaults in an SSL_CTX.
17191 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17192 from the internal representation. Various PKCS#7 fixes: remove some
17193 evil casts and set the enc_dig_alg field properly based on the signing
17198 * Allow PKCS#12 password to be set from the command line or the
17199 environment. Let 'ca' get its config file name from the environment
17200 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17205 * Allow certificate policies extension to use an IA5STRING for the
17206 organization field. This is contrary to the PKIX definition but
17207 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17212 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17213 without disallowing inline assembler and the like for non-pedantic builds.
17217 * Support Borland C++ builder.
17219 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17225 * SHA-1 cleanups and performance enhancements.
17227 *Andy Polyakov <appro@fy.chalmers.se>*
17229 * Sparc v8plus assembler for the bignum library.
17231 *Andy Polyakov <appro@fy.chalmers.se>*
17233 * Accept any -xxx and +xxx compiler options in Configure.
17237 * Update HPUX configuration.
17241 * Add missing `sk_<type>_unshift()` function to safestack.h
17243 *Ralf S. Engelschall*
17245 * New function SSL_CTX_use_certificate_chain_file that sets the
17246 "extra_cert"s in addition to the certificate. (This makes sense
17247 only for "PEM" format files, as chains as a whole are not
17252 * Support verify_depth from the SSL API.
17253 x509_vfy.c had what can be considered an off-by-one-error:
17254 Its depth (which was not part of the external interface)
17255 was actually counting the number of certificates in a chain;
17256 now it really counts the depth.
17260 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17261 instead of X509err, which often resulted in confusing error
17262 messages since the error codes are not globally unique
17263 (e.g. an alleged error in ssl3_accept when a certificate
17264 didn't match the private key).
17266 * New function SSL_CTX_set_session_id_context that allows to set a default
17267 value (so that you don't need SSL_set_session_id_context for each
17268 connection using the SSL_CTX).
17272 * OAEP decoding bug fix.
17276 * Support INSTALL_PREFIX for package builders, as proposed by
17281 * New Configure options "threads" and "no-threads". For systems
17282 where the proper compiler options are known (currently Solaris
17283 and Linux), "threads" is the default.
17287 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17291 * Install various scripts to $(OPENSSLDIR)/misc, not to
17292 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17293 such as /usr/local/bin.
17297 * "make linux-shared" to build shared libraries.
17299 *Niels Poppe <niels@netbox.org>*
17301 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17305 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17306 extension adding in x509 utility.
17310 * Remove NOPROTO sections and error code comments.
17314 * Partial rewrite of the DEF file generator to now parse the ANSI
17319 * New Configure options --prefix=DIR and --openssldir=DIR.
17323 * Complete rewrite of the error code script(s). It is all now handled
17324 by one script at the top level which handles error code gathering,
17325 header rewriting and C source file generation. It should be much better
17326 than the old method: it now uses a modified version of Ulf's parser to
17327 read the ANSI prototypes in all header files (thus the old K&R definitions
17328 aren't needed for error creation any more) and do a better job of
17329 translating function codes into names. The old 'ASN1 error code embedded
17330 in a comment' is no longer necessary and it doesn't use .err files which
17331 have now been deleted. Also the error code call doesn't have to appear all
17332 on one line (which resulted in some large lines...).
17336 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17340 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17341 0 (which usually indicates a closed connection), but continue reading.
17345 * Fix some race conditions.
17349 * Add support for CRL distribution points extension. Add Certificate
17350 Policies and CRL distribution points documentation.
17354 * Move the autogenerated header file parts to crypto/opensslconf.h.
17358 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17359 8 of keying material. Merlin has also confirmed interop with this fix
17360 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17362 *Merlin Hughes <merlin@baltimore.ie>*
17364 * Fix lots of warnings.
17366 *Richard Levitte <levitte@stacken.kth.se>*
17368 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17369 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17371 *Richard Levitte <levitte@stacken.kth.se>*
17373 * Fix problems with sizeof(long) == 8.
17375 *Andy Polyakov <appro@fy.chalmers.se>*
17377 * Change functions to ANSI C.
17381 * Fix typos in error codes.
17383 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17385 * Remove defunct assembler files from Configure.
17389 * SPARC v8 assembler BIGNUM implementation.
17391 *Andy Polyakov <appro@fy.chalmers.se>*
17393 * Support for Certificate Policies extension: both print and set.
17394 Various additions to support the r2i method this uses.
17398 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17399 return a const string when you are expecting an allocated buffer.
17403 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17404 types DirectoryString and DisplayText.
17408 * Add code to allow r2i extensions to access the configuration database,
17409 add an LHASH database driver and add several ctx helper functions.
17413 * Fix an evil bug in bn_expand2() which caused various BN functions to
17414 fail when they extended the size of a BIGNUM.
17418 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17419 support typesafe stack.
17423 * Fix typo in SSL_[gs]et_options().
17425 *Nils Frostberg <nils@medcom.se>*
17427 * Delete various functions and files that belonged to the (now obsolete)
17428 old X509V3 handling code.
17432 * New Configure option "rsaref".
17436 * Don't auto-generate pem.h.
17440 * Introduce type-safe ASN.1 SETs.
17444 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17446 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17448 * Introduce type-safe STACKs. This will almost certainly break lots of code
17449 that links with OpenSSL (well at least cause lots of warnings), but fear
17450 not: the conversion is trivial, and it eliminates loads of evil casts. A
17451 few STACKed things have been converted already. Feel free to convert more.
17452 In the fullness of time, I'll do away with the STACK type altogether.
17456 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17457 specified in `<certfile>` by updating the entry in the index.txt file.
17458 This way one no longer has to edit the index.txt file manually for
17459 revoking a certificate. The -revoke option does the gory details now.
17461 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17463 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17464 `-text` option at all and this way the `-noout -text` combination was
17465 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17467 *Ralf S. Engelschall*
17469 * Make sure a corresponding plain text error message exists for the
17470 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17471 verify callback function determined that a certificate was revoked.
17473 *Ralf S. Engelschall*
17475 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17476 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17477 all available ciphers including rc5, which was forgotten until now.
17478 In order to let the testing shell script know which algorithms
17479 are available, a new (up to now undocumented) command
17480 `openssl list-cipher-commands` is used.
17484 * Bugfix: s_client occasionally would sleep in select() when
17485 it should have checked SSL_pending() first.
17489 * New functions DSA_do_sign and DSA_do_verify to provide access to
17490 the raw DSA values prior to ASN.1 encoding.
17494 * Tweaks to Configure
17496 *Niels Poppe <niels@netbox.org>*
17498 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17503 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17507 * New config option to avoid instructions that are illegal on the 80386.
17508 The default code is faster, but requires at least a 486.
17512 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17513 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17514 same as SSL2_VERSION anyway.
17518 * New "-showcerts" option for s_client.
17522 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17523 application. Various cleanups and fixes.
17527 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17528 modify error routines to work internally. Add error codes and PBE init
17529 to library startup routines.
17533 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17534 packing functions to asn1 and evp. Changed function names and error
17535 codes along the way.
17539 * PKCS12 integration: and so it begins... First of several patches to
17540 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17541 objects to objects.h
17545 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17546 and display support for Thawte strong extranet extension.
17550 * Add LinuxPPC support.
17552 *Jeff Dubrule <igor@pobox.org>*
17554 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17555 bn_div_words in alpha.s.
17557 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17559 * Make sure the RSA OAEP test is skipped under -DRSAref because
17560 OAEP isn't supported when OpenSSL is built with RSAref.
17562 *Ulf Moeller <ulf@fitug.de>*
17564 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17565 so they no longer are missing under -DNOPROTO.
17567 *Soren S. Jorvang <soren@t.dk>*
17569 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17571 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17572 doesn't work when the session is reused. Coming soon!
17576 * Fix a security hole, that allows sessions to be reused in the wrong
17577 context thus bypassing client cert protection! All software that uses
17578 client certs and session caches in multiple contexts NEEDS PATCHING to
17579 allow session reuse! A fuller solution is in the works.
17581 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17583 * Some more source tree cleanups (removed obsolete files
17584 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17585 permission on "config" script to be executable) and a fix for the INSTALL
17588 *Ulf Moeller <ulf@fitug.de>*
17590 * Remove some legacy and erroneous uses of malloc, free instead of
17593 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17595 * Make rsa_oaep_test return non-zero on error.
17597 *Ulf Moeller <ulf@fitug.de>*
17599 * Add support for native Solaris shared libraries. Configure
17600 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17601 if someone would make that last step automatic.
17603 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17605 * ctx_size was not built with the right compiler during "make links". Fixed.
17609 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17610 except NULL ciphers". This means the default cipher list will no longer
17611 enable NULL ciphers. They need to be specifically enabled e.g. with
17612 the string "DEFAULT:eNULL".
17616 * Fix to RSA private encryption routines: if p < q then it would
17617 occasionally produce an invalid result. This will only happen with
17618 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17622 * Be less restrictive and allow also `perl util/perlpath.pl
17623 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17624 because this way one can also use an interpreter named `perl5` (which is
17625 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17626 installed as `perl`).
17628 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17630 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17632 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17634 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17635 advapi32.lib to Win32 build and change the pem test comparison
17636 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17637 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17638 and crypto/des/ede_cbcm_enc.c.
17642 * DES quad checksum was broken on big-endian architectures. Fixed.
17646 * Comment out two functions in bio.h that aren't implemented. Fix up the
17647 Win32 test batch file so it (might) work again. The Win32 test batch file
17648 is horrible: I feel ill....
17652 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17653 in e_os.h. Audit of header files to check ANSI and non ANSI
17654 sections: 10 functions were absent from non ANSI section and not exported
17655 from Windows DLLs. Fixed up libeay.num for new functions.
17659 * Make `openssl version` output lines consistent.
17661 *Ralf S. Engelschall*
17663 * Fix Win32 symbol export lists for BIO functions: Added
17664 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17665 to ms/libeay{16,32}.def.
17667 *Ralf S. Engelschall*
17669 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17670 fine under Unix and passes some trivial tests I've now added. But the
17671 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17672 added to make sure no one expects that this stuff really works in the
17673 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17674 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17677 *Ralf S. Engelschall*
17679 * Fix the generation of two part addresses in perl.
17681 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17683 * Add config entry for Linux on MIPS.
17685 *John Tobey <jtobey@channel1.com>*
17687 * Make links whenever Configure is run, unless we are on Windoze.
17691 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17692 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17697 * Add a useful kludge to allow package maintainers to specify compiler and
17698 other platforms details on the command line without having to patch the
17699 Configure script every time: One now can use
17700 `perl Configure <id>:<details>`,
17701 i.e. platform ids are allowed to have details appended
17702 to them (separated by colons). This is treated as there would be a static
17703 pre-configured entry in Configure's %table under key `<id>` with value
17704 `<details>` and `perl Configure <id>` is called. So, when you want to
17705 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17706 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17707 now, which overrides the FreeBSD-elf entry on-the-fly.
17709 *Ralf S. Engelschall*
17711 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17715 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17716 on the `perl Configure ...` command line. This way one can compile
17717 OpenSSL libraries with Position Independent Code (PIC) which is needed
17718 for linking it into DSOs.
17720 *Ralf S. Engelschall*
17722 * Remarkably, export ciphers were totally broken and no-one had noticed!
17727 * Cleaned up the LICENSE document: The official contact for any license
17728 questions now is the OpenSSL core team under openssl-core@openssl.org.
17729 And add a paragraph about the dual-license situation to make sure people
17730 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17731 to the OpenSSL toolkit.
17733 *Ralf S. Engelschall*
17735 * General source tree makefile cleanups: Made `making xxx in yyy...`
17736 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17737 Additionally cleaned up the `make links` target: Remove unnecessary
17738 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17739 to speed processing and no longer clutter the display with confusing
17740 stuff. Instead only the actually done links are displayed.
17742 *Ralf S. Engelschall*
17744 * Permit null encryption ciphersuites, used for authentication only. It used
17745 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17746 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17751 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17752 signed attributes when verifying signatures (this would break them),
17753 the detached data encoding was wrong and public keys obtained using
17754 X509_get_pubkey() weren't freed.
17758 * Add text documentation for the BUFFER functions. Also added a work around
17759 to a Win95 console bug. This was triggered by the password read stuff: the
17760 last character typed gets carried over to the next fread(). If you were
17761 generating a new cert request using 'req' for example then the last
17762 character of the passphrase would be CR which would then enter the first
17767 * Added the new 'Includes OpenSSL Cryptography Software' button as
17768 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17769 button and can be used by applications based on OpenSSL to show the
17770 relationship to the OpenSSL project.
17772 *Ralf S. Engelschall*
17774 * Remove confusing variables in function signatures in files
17775 ssl/ssl_lib.c and ssl/ssl.h.
17777 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17779 * Don't install bss_file.c under PREFIX/include/
17781 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17783 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17784 functions that return function pointers and has support for NT specific
17785 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17786 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17787 unsigned to signed types: this was killing the Win32 compile.
17791 * Add new certificate file to stack functions,
17792 SSL_add_dir_cert_subjects_to_stack() and
17793 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17794 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17795 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17796 This means that Apache-SSL and similar packages don't have to mess around
17797 to add as many CAs as they want to the preferred list.
17801 * Experiment with doxygen documentation. Currently only partially applied to
17803 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17804 openssl.doxy as the configuration file.
17808 * Get rid of remaining C++-style comments which strict C compilers hate.
17810 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17812 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17813 compiled in by default: it has problems with large keys.
17817 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17818 DH private keys and/or callback functions which directly correspond to
17819 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17820 is needed for applications which have to configure certificates on a
17821 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17823 For the RSA certificate situation is makes no difference, but
17824 for the DSA certificate situation this fixes the "no shared cipher"
17825 problem where the OpenSSL cipher selection procedure failed because the
17826 temporary keys were not overtaken from the context and the API provided
17827 no way to reconfigure them.
17828 The new functions now let applications reconfigure the stuff and they
17829 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17830 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17831 non-public-API function ssl_cert_instantiate() is used as a helper
17832 function and also to reduce code redundancy inside ssl_rsa.c.
17834 *Ralf S. Engelschall*
17836 * Move s_server -dcert and -dkey options out of the undocumented feature
17837 area because they are useful for the DSA situation and should be
17838 recognized by the users.
17840 *Ralf S. Engelschall*
17842 * Fix the cipher decision scheme for export ciphers: the export bits are
17843 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17844 SSL_EXP_MASK. So, the original variable has to be used instead of the
17845 already masked variable.
17847 *Richard Levitte <levitte@stacken.kth.se>*
17849 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17851 *Richard Levitte <levitte@stacken.kth.se>*
17853 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17854 from `int` to `unsigned int` because it is a length and initialized by
17855 EVP_DigestFinal() which expects an `unsigned int *`.
17857 *Richard Levitte <levitte@stacken.kth.se>*
17859 * Don't hard-code path to Perl interpreter on shebang line of Configure
17860 script. Instead use the usual Shell->Perl transition trick.
17862 *Ralf S. Engelschall*
17864 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17865 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17866 -noout -modulus` as it's already the case for `openssl rsa -noout
17867 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17868 currently the public key is printed (a decision which was already done by
17869 `openssl dsa -modulus` in the past) which serves a similar purpose.
17870 Additionally the NO_RSA no longer completely removes the whole -modulus
17871 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17874 *Ralf S. Engelschall*
17876 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17877 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17879 *Arne Ansper <arne@ats.cyber.ee>*
17881 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17882 to be added. Now both 'req' and 'ca' can use new objects defined in the
17887 * Add cool BIO that does syslog (or event log on NT).
17889 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17891 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17892 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17893 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17894 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17898 * Add preliminary config info for new extension code.
17902 * Make RSA_NO_PADDING really use no padding.
17904 *Ulf Moeller <ulf@fitug.de>*
17906 * Generate errors when private/public key check is done.
17910 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17911 for some CRL extensions and new objects added.
17915 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17916 key usage extension and fuller support for authority key id.
17920 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17921 padding method for RSA, which is recommended for new applications in PKCS
17922 #1 v2.0 (RFC 2437, October 1998).
17923 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17924 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17925 against Bleichbacher's attack on RSA.
17926 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17929 * Updates to the new SSL compression code
17931 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17933 * Fix so that the version number in the master secret, when passed
17934 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17935 (because the server will not accept higher), that the version number
17936 is 0x03,0x01, not 0x03,0x00
17938 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17940 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17941 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17942 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17946 * Support for RAW extensions where an arbitrary extension can be
17947 created by including its DER encoding. See apps/openssl.cnf for
17952 * Make sure latest Perl versions don't interpret some generated C array
17953 code as Perl array code in the crypto/err/err_genc.pl script.
17955 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17957 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17958 not many people have the assembler. Various Win32 compilation fixes and
17959 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17960 build instructions.
17964 * Modify configure script 'Configure' to automatically create crypto/date.h
17965 file under Win32 and also build pem.h from pem.org. New script
17966 util/mkfiles.pl to create the MINFO file on environments that can't do a
17967 'make files': perl util/mkfiles.pl >MINFO should work.
17971 * Major rework of DES function declarations, in the pursuit of correctness
17972 and purity. As a result, many evil casts evaporated, and some weirdness,
17973 too. You may find this causes warnings in your code. Zapping your evil
17974 casts will probably fix them. Mostly.
17978 * Fix for a typo in asn1.h. Bug fix to object creation script
17979 obj_dat.pl. It considered a zero in an object definition to mean
17980 "end of object": none of the objects in objects.h have any zeros
17981 so it wasn't spotted.
17983 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17985 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17986 Masking (CBCM). In the absence of test vectors, the best I have been able
17987 to do is check that the decrypt undoes the encrypt, so far. Send me test
17988 vectors if you have them.
17992 * Correct calculation of key length for export ciphers (too much space was
17993 allocated for null ciphers). This has not been tested!
17997 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17998 message is now correct (it understands "crypto" and "ssl" on its
17999 command line). There is also now an "update" option. This will update
18000 the util/ssleay.num and util/libeay.num files with any new functions.
18002 perl util/mkdef.pl crypto ssl update
18003 it will update them.
18007 * Overhauled the Perl interface:
18008 - ported BN stuff to OpenSSL's different BN library
18009 - made the perl/ source tree CVS-aware
18010 - renamed the package from SSLeay to OpenSSL (the files still contain
18011 their history because I've copied them in the repository)
18012 - removed obsolete files (the test scripts will be replaced
18013 by better Test::Harness variants in the future)
18015 *Ralf S. Engelschall*
18017 * First cut for a very conservative source tree cleanup:
18018 1. merge various obsolete readme texts into doc/ssleay.txt
18019 where we collect the old documents and readme texts.
18020 2. remove the first part of files where I'm already sure that we no
18021 longer need them because of three reasons: either they are just temporary
18022 files which were left by Eric or they are preserved original files where
18023 I've verified that the diff is also available in the CVS via "cvs diff
18024 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18025 the crypto/md/ stuff).
18027 *Ralf S. Engelschall*
18029 * More extension code. Incomplete support for subject and issuer alt
18030 name, issuer and authority key id. Change the i2v function parameters
18031 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18032 what that's for :-) Fix to ASN1 macro which messed up
18033 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18037 * Preliminary support for ENUMERATED type. This is largely copied from the
18042 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18044 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18046 * Make sure `make rehash` target really finds the `openssl` program.
18048 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18050 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18051 like to hear about it if this slows down other processors.
18055 * Add CygWin32 platform information to Configure script.
18057 *Alan Batie <batie@aahz.jf.intel.com>*
18059 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18061 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18063 * New program nseq to manipulate netscape certificate sequences
18067 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18072 * Fixes to BN code. Previously the default was to define BN_RECURSION
18073 but the BN code had some problems that would cause failures when
18074 doing certificate verification and some other functions.
18076 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18078 * Add ASN1 and PEM code to support netscape certificate sequences.
18082 * Add ASN1 and PEM code to support netscape certificate sequences.
18086 * Add several PKIX and private extended key usage OIDs.
18090 * Modify the 'ca' program to handle the new extension code. Modify
18091 openssl.cnf for new extension format, add comments.
18095 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18096 and add a sample to openssl.cnf so req -x509 now adds appropriate
18101 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18102 error code, add initial support to X509_print() and x509 application.
18106 * Takes a deep breath and start adding X509 V3 extension support code. Add
18107 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18108 stuff is currently isolated and isn't even compiled yet.
18112 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18113 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18114 Removed the versions check from X509 routines when loading extensions:
18115 this allows certain broken certificates that don't set the version
18116 properly to be processed.
18120 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18121 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18122 can still be regenerated with "make depend".
18126 * Spelling mistake in C version of CAST-128.
18128 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18130 * Changes to the error generation code. The perl script err-code.pl
18131 now reads in the old error codes and retains the old numbers, only
18132 adding new ones if necessary. It also only changes the .err files if new
18133 codes are added. The makefiles have been modified to only insert errors
18134 when needed (to avoid needlessly modifying header files). This is done
18135 by only inserting errors if the .err file is newer than the auto generated
18136 C file. To rebuild all the error codes from scratch (the old behaviour)
18137 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18138 or delete all the .err files.
18142 * CAST-128 was incorrectly implemented for short keys. The C version has
18143 been fixed, but is untested. The assembler versions are also fixed, but
18144 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18145 to regenerate it if needed.
18146 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18147 Hagino <itojun@kame.net>*
18149 * File was opened incorrectly in randfile.c.
18151 *Ulf Möller <ulf@fitug.de>*
18153 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18154 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18155 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18156 al: it's just almost always a UTCTime. Note this patch adds new error
18157 codes so do a "make errors" if there are problems.
18161 * Correct Linux 1 recognition in config.
18163 *Ulf Möller <ulf@fitug.de>*
18165 * Remove pointless MD5 hash when using DSA keys in ca.
18167 *Anonymous <nobody@replay.com>*
18169 * Generate an error if given an empty string as a cert directory. Also
18170 generate an error if handed NULL (previously returned 0 to indicate an
18171 error, but didn't set one).
18173 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18175 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18179 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18180 parameters. This was causing a warning which killed off the Win32 compile.
18184 * Remove C++ style comments from crypto/bn/bn_local.h.
18186 *Neil Costigan <neil.costigan@celocom.com>*
18188 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18189 based on a text string, looking up short and long names and finally
18190 "dot" format. The "dot" format stuff didn't work. Added new function
18191 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18192 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18193 OID is not part of the table.
18197 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18198 X509_LOOKUP_by_alias().
18202 * Sort openssl functions by name.
18206 * Get the gendsa program working (hopefully) and add it to app list. Remove
18207 encryption from sample DSA keys (in case anyone is interested the password
18212 * Make *all* `*_free` functions accept a NULL pointer.
18214 *Frans Heymans <fheymans@isaserver.be>*
18216 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18219 *Anonymous <nobody@replay.com>*
18221 * s_server should send the CAfile as acceptable CAs, not its own cert.
18223 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18225 * Don't blow it for numeric -newkey arguments to apps/req.
18227 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18229 * Temp key "for export" tests were wrong in s3_srvr.c.
18231 *Anonymous <nobody@replay.com>*
18233 * Add prototype for temp key callback functions
18234 SSL_CTX_set_tmp_{rsa,dh}_callback().
18238 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18239 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18243 * X509_name_add_entry() freed the wrong thing after an error.
18245 *Arne Ansper <arne@ats.cyber.ee>*
18247 * rsa_eay.c would attempt to free a NULL context.
18249 *Arne Ansper <arne@ats.cyber.ee>*
18251 * BIO_s_socket() had a broken should_retry() on Windoze.
18253 *Arne Ansper <arne@ats.cyber.ee>*
18255 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18257 *Arne Ansper <arne@ats.cyber.ee>*
18259 * Make sure the already existing X509_STORE->depth variable is initialized
18260 in X509_STORE_new(), but document the fact that this variable is still
18261 unused in the certificate verification process.
18263 *Ralf S. Engelschall*
18265 * Fix the various library and apps files to free up pkeys obtained from
18266 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18270 * Fix reference counting in X509_PUBKEY_get(). This makes
18271 demos/maurice/example2.c work, amongst others, probably.
18273 *Steve Henson and Ben Laurie*
18275 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18276 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18277 are no longer created. This way we have a single and consistent command
18278 line interface `openssl <command>`, similar to `cvs <command>`.
18280 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18282 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18283 BIT STRING wrapper always have zero unused bits.
18287 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18291 * Make the top-level INSTALL documentation easier to understand.
18295 * Makefiles updated to exit if an error occurs in a sub-directory
18296 make (including if user presses ^C) [Paul Sutton]
18298 * Make Montgomery context stuff explicit in RSA data structure.
18302 * Fix build order of pem and err to allow for generated pem.h.
18306 * Fix renumbering bug in X509_NAME_delete_entry().
18310 * Enhanced the err-ins.pl script so it makes the error library number
18311 global and can add a library name. This is needed for external ASN1 and
18312 other error libraries.
18316 * Fixed sk_insert which never worked properly.
18320 * Fix ASN1 macros so they can handle indefinite length constructed
18321 EXPLICIT tags. Some non standard certificates use these: they can now
18326 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18327 into a single doc/ssleay.txt bundle. This way the information is still
18328 preserved but no longer messes up this directory. Now it's new room for
18329 the new set of documentation files.
18331 *Ralf S. Engelschall*
18333 * SETs were incorrectly DER encoded. This was a major pain, because they
18334 shared code with SEQUENCEs, which aren't coded the same. This means that
18335 almost everything to do with SETs or SEQUENCEs has either changed name or
18336 number of arguments.
18338 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18340 * Fix test data to work with the above.
18344 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18345 was already fixed by Eric for 0.9.1 it seems.
18347 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18349 * Autodetect FreeBSD3.
18353 * Fix various bugs in Configure. This affects the following platforms:
18357 unixware-2.0-pentium
18362 * Eliminate generated files from CVS. Reorder tests to regenerate files
18363 before they are needed.
18367 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18371 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18373 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18374 changed SSLeay to OpenSSL in version strings.
18376 *Ralf S. Engelschall*
18378 * Some fixups to the top-level documents.
18382 * Fixed the nasty bug where rsaref.h was not found under compile-time
18383 because the symlink to include/ was missing.
18385 *Ralf S. Engelschall*
18387 * Incorporated the popular no-RSA/DSA-only patches
18388 which allow to compile a RSA-free SSLeay.
18390 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18392 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18393 when "ssleay" is still not found.
18395 *Ralf S. Engelschall*
18397 * Added more platforms to Configure: Cray T3E, HPUX 11,
18399 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18401 * Updated the README file.
18403 *Ralf S. Engelschall*
18405 * Added various .cvsignore files in the CVS repository subdirs
18406 to make a "cvs update" really silent.
18408 *Ralf S. Engelschall*
18410 * Recompiled the error-definition header files and added
18411 missing symbols to the Win32 linker tables.
18413 *Ralf S. Engelschall*
18415 * Cleaned up the top-level documents;
18416 o new files: CHANGES and LICENSE
18417 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18418 o merged COPYRIGHT into LICENSE
18419 o removed obsolete TODO file
18420 o renamed MICROSOFT to INSTALL.W32
18422 *Ralf S. Engelschall*
18424 * Removed dummy files from the 0.9.1b source tree:
18425 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18426 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18427 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18428 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18429 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18431 *Ralf S. Engelschall*
18433 * Added various platform portability fixes.
18437 * The Genesis of the OpenSSL rpject:
18438 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18439 Young and Tim J. Hudson created while they were working for C2Net until
18442 *The OpenSSL Project*
18444 ### Changes between 0.9.0b and 0.9.1b [not released]
18446 * Updated a few CA certificates under certs/
18450 * Changed some BIGNUM api stuff.
18454 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18455 DGUX x86, Linux Alpha, etc.
18459 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18460 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18465 * Add -strparse option to asn1pars program which parses nested
18468 *Dr Stephen Henson <shenson@bigfoot.com>*
18470 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18474 * DSA fix for "ca" program.
18478 * Added "-genkey" option to "dsaparam" program.
18482 * Added RIPE MD160 (rmd160) message digest.
18486 * Added -a (all) option to "ssleay version" command.
18490 * Added PLATFORM define which is the id given to Configure.
18494 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18498 * Extended the ASN.1 parser routines.
18502 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18506 * Added a BN_CTX to the BN library.
18510 * Fixed the weak key values in DES library
18514 * Changed API in EVP library for cipher aliases.
18518 * Added support for RC2/64bit cipher.
18522 * Converted the lhash library to the crypto/mem.c functions.
18526 * Added more recognized ASN.1 object ids.
18530 * Added more RSA padding checks for SSL/TLS.
18534 * Added BIO proxy/filter functionality.
18538 * Added extra_certs to SSL_CTX which can be used
18539 send extra CA certificates to the client in the CA cert chain sending
18540 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18544 * Now Fortezza is denied in the authentication phase because
18545 this is key exchange mechanism is not supported by SSLeay at all.
18549 * Additional PKCS1 checks.
18553 * Support the string "TLSv1" for all TLS v1 ciphers.
18557 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18558 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18562 * Fixed a few memory leaks.
18566 * Fixed various code and comment typos.
18570 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18571 bytes sent in the client random.
18573 *Edward Bishop <ebishop@spyglass.com>*
18577 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18578 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18579 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18580 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18581 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18582 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18583 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18584 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18585 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18586 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18587 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18588 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18589 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18590 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18591 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18592 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18593 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18594 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18595 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18596 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18597 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18598 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18599 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18600 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18601 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18602 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18603 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18604 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18605 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18606 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18607 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18608 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18609 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18610 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18611 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18612 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18613 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18614 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18615 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18616 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18617 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18618 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18619 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18620 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18621 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18622 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18623 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18624 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18625 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18626 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18627 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18628 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18629 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18630 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18631 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18632 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18633 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18634 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18635 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18636 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18637 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18638 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18639 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18640 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18641 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18642 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18643 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18644 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18645 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18646 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18647 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18648 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18649 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18650 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18651 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18652 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18653 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18654 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18655 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18656 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18657 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18658 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18659 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18660 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18661 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18662 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18663 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18664 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18665 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18666 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18667 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18668 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18669 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18670 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18671 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18672 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18673 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18674 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18675 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18676 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18677 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18678 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18679 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18680 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18681 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18682 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18683 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18684 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18685 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18686 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18687 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18688 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18689 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18690 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18691 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18692 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18693 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18694 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18695 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18696 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18697 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18698 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18699 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18700 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18701 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18702 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18703 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18704 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18705 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18706 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18707 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18708 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18709 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18710 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18711 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18712 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18713 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18714 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18715 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18716 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18717 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18718 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18719 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18720 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18721 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18722 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18723 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18724 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18725 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18726 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18727 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18728 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18729 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18730 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18731 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18732 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18733 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18734 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18735 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18736 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18737 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655