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 * Added several checks to X509_verify_cert() according to requirements in
835 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
836 (which may be done by using the CLI option `-x509_strict`):
837 * The basicConstraints of CA certificates must be marked critical.
838 * CA certificates must explicitly include the keyUsage extension.
839 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
840 * The issuer name of any certificate must not be empty.
841 * The subject name of CA certs, certs with keyUsage crlSign,
842 and certs without subjectAlternativeName must not be empty.
843 * If a subjectAlternativeName extension is given it must not be empty.
844 * The signatureAlgorithm field and the cert signature must be consistent.
845 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
846 must not be marked critical.
847 * The authorityKeyIdentifier must be given for X.509v3 certs
848 unless they are self-signed.
849 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
853 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
854 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
858 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
859 used even when parsing explicit parameters, when loading a encoded key
860 or calling `EC_GROUP_new_from_ecpkparameters()`/
861 `EC_GROUP_new_from_ecparameters()`.
862 This prevents bypass of security hardening and performance gains,
863 especially for curves with specialized EC_METHODs.
864 By default, if a key encoded with explicit parameters is loaded and later
865 encoded, the output is still encoded with explicit parameters, even if
866 internally a "named" EC_GROUP is used for computation.
870 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
871 this change, EC_GROUP_set_generator would accept order and/or cofactor as
872 NULL. After this change, only the cofactor parameter can be NULL. It also
873 does some minimal sanity checks on the passed order.
878 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
879 An attack is simple, if the first CMS_recipientInfo is valid but the
880 second CMS_recipientInfo is chosen ciphertext. If the second
881 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
882 encryption key will be replaced by garbage, and the message cannot be
883 decoded, but if the RSA decryption fails, the correct encryption key is
884 used and the recipient will not notice the attack.
885 As a work around for this potential attack the length of the decrypted
886 key must be equal to the cipher default key length, in case the
887 certifiate is not given and all recipientInfo are tried out.
888 The old behaviour can be re-enabled in the CMS code by setting the
889 CMS_DEBUG_DECRYPT flag.
893 * Early start up entropy quality from the DEVRANDOM seed source has been
894 improved for older Linux systems. The RAND subsystem will wait for
895 /dev/random to be producing output before seeding from /dev/urandom.
896 The seeded state is stored for future library initialisations using
897 a system global shared memory segment. The shared memory identifier
898 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
899 the desired value. The default identifier is 114.
903 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
904 when primes for RSA keys are computed.
905 Since we previously always generated primes == 2 (mod 3) for RSA keys,
906 the 2-prime and 3-prime RSA modules were easy to distinguish, since
907 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
908 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
909 This avoids possible fingerprinting of newly generated RSA modules.
913 * Correct the extended master secret constant on EBCDIC systems. Without this
914 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
915 negotiate EMS will fail. Unfortunately this also means that TLS connections
916 between EBCDIC systems with this fix, and EBCDIC systems without this
917 fix will fail if they negotiate EMS.
921 * Changed the library initialisation so that the config file is now loaded
922 by default. This was already the case for libssl. It now occurs for both
923 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
924 OPENSSL_init_crypto() to suppress automatic loading of a config file.
928 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
929 where the former acts as a replacement for ERR_put_error(), and the
930 latter replaces the combination ERR_put_error()+ERR_add_error_data().
931 ERR_raise_data() adds more flexibility by taking a format string and
932 an arbitrary number of arguments following it, to be processed with
937 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
938 to check if a named provider is loaded and available. When called, it
939 will also activate all fallback providers if such are still present.
943 * Enforce a minimum DH modulus size of 512 bits.
947 * Changed DH parameters to generate the order q subgroup instead of 2q.
948 Previously generated DH parameters are still accepted by DH_check
949 but DH_generate_key works around that by clearing bit 0 of the
950 private key for those. This avoids leaking bit 0 of the private key.
954 * Significantly reduce secure memory usage by the randomness pools.
958 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
963 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
964 algorithms. An implementation of a key exchange algorithm can be obtained
965 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
966 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
967 the older EVP_PKEY_derive_init() function. See the man pages for the new
968 functions for further details.
972 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
976 * Removed the function names from error messages and deprecated the
979 * Removed NextStep support and the macro OPENSSL_UNISTD
983 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
984 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
985 Also removed "export var as function" capability; we do not export
986 variables, only functions.
990 * RC5_32_set_key has been changed to return an int type, with 0 indicating
991 an error and 1 indicating success. In previous versions of OpenSSL this
992 was a void type. If a key was set longer than the maximum possible this
997 * Support SM2 signing and verification schemes with X509 certificate.
1001 * Use SHA256 as the default digest for TS query in the ts app.
1005 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1006 This checks that the salt length is at least 128 bits, the derived key
1007 length is at least 112 bits, and that the iteration count is at least 1000.
1008 For backwards compatibility these checks are disabled by default in the
1009 default provider, but are enabled by default in the fips provider.
1010 To enable or disable these checks use the control
1011 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1015 * Default cipher lists/suites are now available via a function, the
1016 #defines are deprecated.
1020 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1021 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1022 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1026 * Join the directories crypto/x509 and crypto/x509v3
1030 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1031 This changes the size when using the genpkey app when no size is given. It
1032 fixes an omission in earlier changes that changed all RSA, DSA and DH
1033 generation apps to use 2048 bits by default.
1037 * Added command 'openssl kdf' that uses the EVP_KDF API.
1041 * Added command 'openssl mac' that uses the EVP_MAC API.
1045 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1046 as default directories. Also added the command 'openssl info'
1047 for scripting purposes.
1051 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1052 deprecated. These undocumented functions were never integrated into the EVP
1053 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1054 Bi-directional IGE mode. These modes were never formally standardised and
1055 usage of these functions is believed to be very small. In particular
1056 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1057 is ever used. The security implications are believed to be minimal, but
1058 this issue was never fixed for backwards compatibility reasons. New code
1059 should not use these modes.
1063 * Add prediction resistance to the DRBG reseeding process.
1067 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1068 mandated by IEEE Std 1619-2018.
1072 * Added newline escaping functionality to a filename when using openssl dgst.
1073 This output format is to replicate the output format found in the `*sum`
1074 checksum programs. This aims to preserve backward compatibility.
1076 *Matt Eaton, Richard Levitte, and Paul Dale*
1078 * Removed the heartbeat message in DTLS feature, as it has very
1079 little usage and doesn't seem to fulfill a valuable purpose.
1080 The configuration option is now deprecated.
1084 * Changed the output of 'openssl {digestname} < file' to display the
1085 digest name in its output.
1089 * Added a new generic trace API which provides support for enabling
1090 instrumentation through trace output. This feature is mainly intended
1091 as an aid for developers and is disabled by default. To utilize it,
1092 OpenSSL needs to be configured with the `enable-trace` option.
1094 If the tracing API is enabled, the application can activate trace output
1095 by registering BIOs as trace channels for a number of tracing and debugging
1098 The 'openssl' application has been expanded to enable any of the types
1099 available via environment variables defined by the user, and serves as
1100 one possible example on how to use this functionality.
1102 *Richard Levitte & Matthias St. Pierre*
1104 * Added build tests for C++. These are generated files that only do one
1105 thing, to include one public OpenSSL head file each. This tests that
1106 the public header files can be usefully included in a C++ application.
1108 This test isn't enabled by default. It can be enabled with the option
1109 'enable-buildtest-c++'.
1113 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1117 * Add KMAC to EVP_MAC.
1121 * Added property based algorithm implementation selection framework to
1126 * Added SCA hardening for modular field inversion in EC_GROUP through
1127 a new dedicated field_inv() pointer in EC_METHOD.
1128 This also addresses a leakage affecting conversions from projective
1129 to affine coordinates.
1131 *Billy Bob Brumley, Nicola Tuveri*
1133 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1134 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1135 those algorithms that were already supported through the EVP_PKEY API
1136 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1137 and scrypt are now wrappers that call EVP_KDF.
1141 * Build devcrypto engine as a dynamic engine.
1143 *Eneas U de Queiroz*
1145 * Add keyed BLAKE2 to EVP_MAC.
1149 * Fix a bug in the computation of the endpoint-pair shared secret used
1150 by DTLS over SCTP. This breaks interoperability with older versions
1151 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1152 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1153 interoperability with such broken implementations. However, enabling
1154 this switch breaks interoperability with correct implementations.
1156 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1157 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1161 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1165 * Change the license to the Apache License v2.0.
1169 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1171 - Major releases (indicated by incrementing the MAJOR release number)
1172 may introduce incompatible API/ABI changes.
1173 - Minor releases (indicated by incrementing the MINOR release number)
1174 may introduce new features but retain API/ABI compatibility.
1175 - Patch releases (indicated by incrementing the PATCH number)
1176 are intended for bug fixes and other improvements of existing
1177 features only (like improving performance or adding documentation)
1178 and retain API/ABI compatibility.
1182 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1186 * Remove the 'dist' target and add a tarball building script. The
1187 'dist' target has fallen out of use, and it shouldn't be
1188 necessary to configure just to create a source distribution.
1192 * Recreate the OS390-Unix config target. It no longer relies on a
1193 special script like it did for OpenSSL pre-1.1.0.
1197 * Instead of having the source directories listed in Configure, add
1198 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1203 * Add GMAC to EVP_MAC.
1207 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1211 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1212 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1213 to facilitate the continued use of MACs through raw private keys in
1214 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1218 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1219 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1223 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1224 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1225 are retained for backwards compatibility.
1229 * AES-XTS mode now enforces that its two keys are different to mitigate
1230 the attacked described in "Efficient Instantiations of Tweakable
1231 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1232 Details of this attack can be obtained from:
1233 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1237 * Rename the object files, i.e. give them other names than in previous
1238 versions. Their names now include the name of the final product, as
1239 well as its type mnemonic (bin, lib, shlib).
1243 * Added new option for 'openssl list', '-objects', which will display the
1244 list of built in objects, i.e. OIDs with names.
1248 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1249 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1254 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1255 improves application performance by removing data copies and providing
1256 applications with zero-copy system calls such as sendfile and splice.
1260 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1261 option is set, openssl cleanses (zeroize) plaintext bytes from
1262 internal buffers after delivering them to the application. Note,
1263 the application is still responsible for cleansing other copies
1264 (e.g.: data received by SSL_read(3)).
1268 * `PKCS12_parse` now maintains the order of the parsed certificates
1269 when outputting them via `*ca` (rather than reversing it).
1273 * Deprecated pthread fork support methods. These were unused so no
1274 replacement is required.
1276 - OPENSSL_fork_prepare()
1277 - OPENSSL_fork_parent()
1278 - OPENSSL_fork_child()
1285 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1289 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1291 * Certificates with explicit curve parameters are now disallowed in
1292 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1296 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1297 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1298 conversely, silently ignore DTLS protocol version bounds when configuring
1299 TLS-based contexts. The commands can be repeated to set bounds of both
1300 types. The same applies with the corresponding "min_protocol" and
1301 "max_protocol" command-line switches, in case some application uses both TLS
1304 SSL_CTX instances that are created for a fixed protocol version (e.g.
1305 TLSv1_server_method()) also silently ignore version bounds. Previously
1306 attempts to apply bounds to these protocol versions would result in an
1307 error. Now only the "version-flexible" SSL_CTX instances are subject to
1308 limits in configuration files in command-line options.
1312 * Handshake now fails if Extended Master Secret extension is dropped
1317 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1319 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1321 * Fixed segmentation fault in SSL_check_chain()
1322 Server or client applications that call the SSL_check_chain() function
1323 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1324 dereference as a result of incorrect handling of the
1325 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1326 or unrecognised signature algorithm is received from the peer. This could
1327 be exploited by a malicious peer in a Denial of Service attack.
1332 * Added AES consttime code for no-asm configurations
1333 an optional constant time support for AES was added
1334 when building openssl for no-asm.
1335 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1336 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1337 At this time this feature is by default disabled.
1338 It will be enabled by default in 3.0.
1342 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1344 * Revert the change of EOF detection while reading in libssl to avoid
1345 regressions in applications depending on the current way of reporting
1346 the EOF. As the existing method is not fully accurate the change to
1347 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1348 branch and will be present in the 3.0 release.
1352 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1353 when primes for RSA keys are computed.
1354 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1355 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1356 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1357 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1358 This avoids possible fingerprinting of newly generated RSA modules.
1362 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1364 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1365 while reading in libssl then we would report an error back to the
1366 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1367 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1368 therefore give a hint as to what went wrong.
1372 * Check that ed25519 and ed448 are allowed by the security level. Previously
1373 signature algorithms not using an MD were not being checked that they were
1374 allowed by the security level.
1378 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1379 was not quite right. The behaviour was not consistent between resumption
1380 and normal handshakes, and also not quite consistent with historical
1381 behaviour. The behaviour in various scenarios has been clarified and
1382 it has been updated to make it match historical behaviour as closely as
1387 * *[VMS only]* The header files that the VMS compilers include automatically,
1388 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1389 that the C++ compiler doesn't understand. This is a shortcoming in the
1390 compiler, but can be worked around with `__cplusplus` guards.
1392 C++ applications that use OpenSSL libraries must be compiled using the
1393 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1394 functions. Otherwise, only functions with symbols of less than 31
1395 characters can be used, as the linker will not be able to successfully
1396 resolve symbols with longer names.
1400 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1401 The presence of this system service is determined at run-time.
1405 * Added newline escaping functionality to a filename when using openssl dgst.
1406 This output format is to replicate the output format found in the `*sum`
1407 checksum programs. This aims to preserve backward compatibility.
1409 *Matt Eaton, Richard Levitte, and Paul Dale*
1411 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1416 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1418 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1419 number generator (RNG). This was intended to include protection in the
1420 event of a fork() system call in order to ensure that the parent and child
1421 processes did not share the same RNG state. However this protection was not
1422 being used in the default case.
1424 A partial mitigation for this issue is that the output from a high
1425 precision timer is mixed into the RNG state so the likelihood of a parent
1426 and child process sharing state is significantly reduced.
1428 If an application already calls OPENSSL_init_crypto() explicitly using
1429 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1432 *Matthias St. Pierre*
1434 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1435 used even when parsing explicit parameters, when loading a encoded key
1436 or calling `EC_GROUP_new_from_ecpkparameters()`/
1437 `EC_GROUP_new_from_ecparameters()`.
1438 This prevents bypass of security hardening and performance gains,
1439 especially for curves with specialized EC_METHODs.
1440 By default, if a key encoded with explicit parameters is loaded and later
1441 encoded, the output is still encoded with explicit parameters, even if
1442 internally a "named" EC_GROUP is used for computation.
1446 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1447 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1448 NULL. After this change, only the cofactor parameter can be NULL. It also
1449 does some minimal sanity checks on the passed order.
1454 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1455 An attack is simple, if the first CMS_recipientInfo is valid but the
1456 second CMS_recipientInfo is chosen ciphertext. If the second
1457 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1458 encryption key will be replaced by garbage, and the message cannot be
1459 decoded, but if the RSA decryption fails, the correct encryption key is
1460 used and the recipient will not notice the attack.
1461 As a work around for this potential attack the length of the decrypted
1462 key must be equal to the cipher default key length, in case the
1463 certifiate is not given and all recipientInfo are tried out.
1464 The old behaviour can be re-enabled in the CMS code by setting the
1465 CMS_DEBUG_DECRYPT flag.
1470 * Early start up entropy quality from the DEVRANDOM seed source has been
1471 improved for older Linux systems. The RAND subsystem will wait for
1472 /dev/random to be producing output before seeding from /dev/urandom.
1473 The seeded state is stored for future library initialisations using
1474 a system global shared memory segment. The shared memory identifier
1475 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1476 the desired value. The default identifier is 114.
1480 * Correct the extended master secret constant on EBCDIC systems. Without this
1481 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1482 negotiate EMS will fail. Unfortunately this also means that TLS connections
1483 between EBCDIC systems with this fix, and EBCDIC systems without this
1484 fix will fail if they negotiate EMS.
1488 * Use Windows installation paths in the mingw builds
1490 Mingw isn't a POSIX environment per se, which means that Windows
1491 paths should be used for installation.
1496 * Changed DH_check to accept parameters with order q and 2q subgroups.
1497 With order 2q subgroups the bit 0 of the private key is not secret
1498 but DH_generate_key works around that by clearing bit 0 of the
1499 private key for those. This avoids leaking bit 0 of the private key.
1503 * Significantly reduce secure memory usage by the randomness pools.
1507 * Revert the DEVRANDOM_WAIT feature for Linux systems
1509 The DEVRANDOM_WAIT feature added a select() call to wait for the
1510 /dev/random device to become readable before reading from the
1511 /dev/urandom device.
1513 It turned out that this change had negative side effects on
1514 performance which were not acceptable. After some discussion it
1515 was decided to revert this feature and leave it up to the OS
1516 resp. the platform maintainer to ensure a proper initialization
1517 during early boot time.
1519 *Matthias St. Pierre*
1521 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1523 * Add build tests for C++. These are generated files that only do one
1524 thing, to include one public OpenSSL head file each. This tests that
1525 the public header files can be usefully included in a C++ application.
1527 This test isn't enabled by default. It can be enabled with the option
1528 'enable-buildtest-c++'.
1532 * Enable SHA3 pre-hashing for ECDSA and DSA.
1536 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1537 This changes the size when using the genpkey app when no size is given. It
1538 fixes an omission in earlier changes that changed all RSA, DSA and DH
1539 generation apps to use 2048 bits by default.
1543 * Reorganize the manual pages to consistently have RETURN VALUES,
1544 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1545 util/fix-doc-nits accordingly.
1547 *Paul Yang, Joshua Lock*
1549 * Add the missing accessor EVP_PKEY_get0_engine()
1553 * Have apps like 's_client' and 's_server' output the signature scheme
1554 along with other cipher suite parameters when debugging.
1558 * Make OPENSSL_config() error agnostic again.
1562 * Do the error handling in RSA decryption constant time.
1566 * Prevent over long nonces in ChaCha20-Poly1305.
1568 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1569 for every encryption operation. RFC 7539 specifies that the nonce value
1570 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1571 and front pads the nonce with 0 bytes if it is less than 12
1572 bytes. However it also incorrectly allows a nonce to be set of up to 16
1573 bytes. In this case only the last 12 bytes are significant and any
1574 additional leading bytes are ignored.
1576 It is a requirement of using this cipher that nonce values are
1577 unique. Messages encrypted using a reused nonce value are susceptible to
1578 serious confidentiality and integrity attacks. If an application changes
1579 the default nonce length to be longer than 12 bytes and then makes a
1580 change to the leading bytes of the nonce expecting the new value to be a
1581 new unique nonce then such an application could inadvertently encrypt
1582 messages with a reused nonce.
1584 Additionally the ignored bytes in a long nonce are not covered by the
1585 integrity guarantee of this cipher. Any application that relies on the
1586 integrity of these ignored leading bytes of a long nonce may be further
1587 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1588 is safe because no such use sets such a long nonce value. However user
1589 applications that use this cipher directly and set a non-default nonce
1590 length to be longer than 12 bytes may be vulnerable.
1592 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1598 * Add DEVRANDOM_WAIT feature for Linux systems
1600 On older Linux systems where the getrandom() system call is not available,
1601 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1602 Contrary to getrandom(), the /dev/urandom device will not block during
1603 early boot when the kernel CSPRNG has not been seeded yet.
1605 To mitigate this known weakness, use select() to wait for /dev/random to
1606 become readable before reading from /dev/urandom.
1608 * Ensure that SM2 only uses SM3 as digest algorithm
1612 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1614 * Change the info callback signals for the start and end of a post-handshake
1615 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1616 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1617 confused by this and assume that a TLSv1.2 renegotiation has started. This
1618 can break KeyUpdate handling. Instead we no longer signal the start and end
1619 of a post handshake message exchange (although the messages themselves are
1620 still signalled). This could break some applications that were expecting
1621 the old signals. However without this KeyUpdate is not usable for many
1626 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1628 * Timing vulnerability in DSA signature generation
1630 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1631 timing side channel attack. An attacker could use variations in the signing
1632 algorithm to recover the private key.
1634 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1639 * Timing vulnerability in ECDSA signature generation
1641 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1642 timing side channel attack. An attacker could use variations in the signing
1643 algorithm to recover the private key.
1645 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1650 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1651 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1652 of two gigabytes and the error handling improved.
1654 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1655 categorized as a normal bug, not a security issue, because the DRBG reseeds
1656 automatically and is fully functional even without additional randomness
1657 provided by the application.
1659 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1661 * Add a new ClientHello callback. Provides a callback interface that gives
1662 the application the ability to adjust the nascent SSL object at the
1663 earliest stage of ClientHello processing, immediately after extensions have
1664 been collected but before they have been processed. In particular, this
1665 callback can adjust the supported TLS versions in response to the contents
1670 * Add SM2 base algorithm support.
1674 * s390x assembly pack: add (improved) hardware-support for the following
1675 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1676 aes-cfb/cfb8, aes-ecb.
1680 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1681 parameter is no longer accepted, as it leads to a corrupt table. NULL
1682 pem_str is reserved for alias entries only.
1686 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1687 step for prime curves. The new implementation is based on formulae from
1688 differential addition-and-doubling in homogeneous projective coordinates
1689 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1690 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1691 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1692 to work in projective coordinates.
1694 *Billy Bob Brumley, Nicola Tuveri*
1696 * Change generating and checking of primes so that the error rate of not
1697 being prime depends on the intended use based on the size of the input.
1698 For larger primes this will result in more rounds of Miller-Rabin.
1699 The maximal error rate for primes with more than 1080 bits is lowered
1702 *Kurt Roeckx, Annie Yousar*
1704 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1708 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1709 moving between systems, and to avoid confusion when a Windows build is
1710 done with mingw vs with MSVC. For POSIX installs, there's still a
1711 symlink or copy named 'tsget' to avoid that confusion as well.
1715 * Revert blinding in ECDSA sign and instead make problematic addition
1716 length-invariant. Switch even to fixed-length Montgomery multiplication.
1720 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1721 step for binary curves. The new implementation is based on formulae from
1722 differential addition-and-doubling in mixed Lopez-Dahab projective
1723 coordinates, modified to independently blind the operands.
1725 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1727 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1728 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1729 EC_METHODs to implement their own specialized "ladder step", to take
1730 advantage of more favorable coordinate systems or more efficient
1731 differential addition-and-doubling algorithms.
1733 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1735 * Modified the random device based seed sources to keep the relevant
1736 file descriptors open rather than reopening them on each access.
1737 This allows such sources to operate in a chroot() jail without
1738 the associated device nodes being available. This behaviour can be
1739 controlled using RAND_keep_random_devices_open().
1743 * Numerous side-channel attack mitigations have been applied. This may have
1744 performance impacts for some algorithms for the benefit of improved
1745 security. Specific changes are noted in this change log by their respective
1750 * AIX shared library support overhaul. Switch to AIX "natural" way of
1751 handling shared libraries, which means collecting shared objects of
1752 different versions and bitnesses in one common archive. This allows to
1753 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1754 doesn't affect the way 3rd party applications are linked, only how
1755 multi-version installation is managed.
1759 * Make ec_group_do_inverse_ord() more robust and available to other
1760 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1761 mitigations are applied to the fallback BN_mod_inverse().
1762 When using this function rather than BN_mod_inverse() directly, new
1763 EC cryptosystem implementations are then safer-by-default.
1767 * Add coordinate blinding for EC_POINT and implement projective
1768 coordinate blinding for generic prime curves as a countermeasure to
1769 chosen point SCA attacks.
1771 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1773 * Add blinding to ECDSA and DSA signatures to protect against side channel
1774 attacks discovered by Keegan Ryan (NCC Group).
1778 * Enforce checking in the pkeyutl command line app to ensure that the input
1779 length does not exceed the maximum supported digest length when performing
1780 a sign, verify or verifyrecover operation.
1784 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1785 I/O in combination with something like select() or poll() will hang. This
1786 can be turned off again using SSL_CTX_clear_mode().
1787 Many applications do not properly handle non-application data records, and
1788 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1789 around the problems in those applications, but can also break some.
1790 It's recommended to read the manpages about SSL_read(), SSL_write(),
1791 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1792 SSL_CTX_set_read_ahead() again.
1796 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1797 now allow empty (zero character) pass phrases.
1801 * Apply blinding to binary field modular inversion and remove patent
1802 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1806 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1807 binary and prime elliptic curves.
1811 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1812 constant time fixed point multiplication.
1816 * Revise elliptic curve scalar multiplication with timing attack
1817 defenses: ec_wNAF_mul redirects to a constant time implementation
1818 when computing fixed point and variable point multiplication (which
1819 in OpenSSL are mostly used with secret scalars in keygen, sign,
1820 ECDH derive operations).
1821 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1824 * Updated CONTRIBUTING
1828 * Updated DRBG / RAND to request nonce and additional low entropy
1829 randomness from the system.
1831 *Matthias St. Pierre*
1833 * Updated 'openssl rehash' to use OpenSSL consistent default.
1837 * Moved the load of the ssl_conf module to libcrypto, which helps
1838 loading engines that libssl uses before libssl is initialised.
1842 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1846 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1848 *Ingo Schwarze, Rich Salz*
1850 * Added output of accepting IP address and port for 'openssl s_server'
1854 * Added a new API for TLSv1.3 ciphersuites:
1855 SSL_CTX_set_ciphersuites()
1856 SSL_set_ciphersuites()
1860 * Memory allocation failures consistently add an error to the error
1865 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1866 in libcrypto when run as setuid/setgid.
1870 * Load any config file by default when libssl is used.
1874 * Added new public header file <openssl/rand_drbg.h> and documentation
1875 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1877 *Matthias St. Pierre*
1879 * QNX support removed (cannot find contributors to get their approval
1880 for the license change).
1884 * TLSv1.3 replay protection for early data has been implemented. See the
1885 SSL_read_early_data() man page for further details.
1889 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1890 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1891 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1892 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1893 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1894 configuration has been separated out. See the ciphers man page or the
1895 SSL_CTX_set_ciphersuites() man page for more information.
1899 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1900 in responder mode now supports the new "-multi" option, which
1901 spawns the specified number of child processes to handle OCSP
1902 requests. The "-timeout" option now also limits the OCSP
1903 responder's patience to wait to receive the full client request
1904 on a newly accepted connection. Child processes are respawned
1905 as needed, and the CA index file is automatically reloaded
1906 when changed. This makes it possible to run the "ocsp" responder
1907 as a long-running service, making the OpenSSL CA somewhat more
1908 feature-complete. In this mode, most diagnostic messages logged
1909 after entering the event loop are logged via syslog(3) rather than
1914 * Added support for X448 and Ed448. Heavily based on original work by
1919 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1920 objects loaded. This adds the functions OSSL_STORE_expect() and
1921 OSSL_STORE_find() as well as needed tools to construct searches and
1922 get the search data out of them.
1926 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1927 version of OpenSSL should review their configuration settings to ensure
1928 that they are still appropriate for TLSv1.3. For further information see:
1929 <https://wiki.openssl.org/index.php/TLS1.3>
1933 * Grand redesign of the OpenSSL random generator
1935 The default RAND method now utilizes an AES-CTR DRBG according to
1936 NIST standard SP 800-90Ar1. The new random generator is essentially
1937 a port of the default random generator from the OpenSSL FIPS 2.0
1938 object module. It is a hybrid deterministic random bit generator
1939 using an AES-CTR bit stream and which seeds and reseeds itself
1940 automatically using trusted system entropy sources.
1942 Some of its new features are:
1943 - Support for multiple DRBG instances with seed chaining.
1944 - The default RAND method makes use of a DRBG.
1945 - There is a public and private DRBG instance.
1946 - The DRBG instances are fork-safe.
1947 - Keep all global DRBG instances on the secure heap if it is enabled.
1948 - The public and private DRBG instance are per thread for lock free
1951 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1953 * Changed Configure so it only says what it does and doesn't dump
1954 so much data. Instead, ./configdata.pm should be used as a script
1955 to display all sorts of configuration data.
1959 * Added processing of "make variables" to Configure.
1963 * Added SHA512/224 and SHA512/256 algorithm support.
1967 * The last traces of Netware support, first removed in 1.1.0, have
1972 * Get rid of Makefile.shared, and in the process, make the processing
1973 of certain files (rc.obj, or the .def/.map/.opt files produced from
1974 the ordinal files) more visible and hopefully easier to trace and
1975 debug (or make silent).
1979 * Make it possible to have environment variable assignments as
1980 arguments to config / Configure.
1984 * Add multi-prime RSA (RFC 8017) support.
1988 * Add SM3 implemented according to GB/T 32905-2016
1989 *Jack Lloyd <jack.lloyd@ribose.com>,*
1990 *Ronald Tse <ronald.tse@ribose.com>,*
1991 *Erick Borsboom <erick.borsboom@ribose.com>*
1993 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1994 as documented in RFC6066.
1995 Based on a patch from Tomasz Moń
1997 *Filipe Raimundo da Silva*
1999 * Add SM4 implemented according to GB/T 32907-2016.
2000 *Jack Lloyd <jack.lloyd@ribose.com>,*
2001 *Ronald Tse <ronald.tse@ribose.com>,*
2002 *Erick Borsboom <erick.borsboom@ribose.com>*
2004 * Reimplement -newreq-nodes and ERR_error_string_n; the
2005 original author does not agree with the license change.
2009 * Add ARIA AEAD TLS support.
2013 * Some macro definitions to support VS6 have been removed. Visual
2014 Studio 6 has not worked since 1.1.0
2018 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2019 without clearing the errors.
2023 * Add "atfork" functions. If building on a system that without
2024 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2025 requirements. The RAND facility now uses/requires this.
2033 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2034 not possible to disable entirely. However, it's still possible to
2035 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2038 To disable, configure with 'no-ui-console'. 'no-ui' is still
2039 possible to use as an alias. Check at compile time with the
2040 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2041 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2045 * Add a STORE module, which implements a uniform and URI based reader of
2046 stores that can contain keys, certificates, CRLs and numerous other
2047 objects. The main API is loosely based on a few stdio functions,
2048 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2049 OSSL_STORE_error and OSSL_STORE_close.
2050 The implementation uses backends called "loaders" to implement arbitrary
2051 URI schemes. There is one built in "loader" for the 'file' scheme.
2055 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2056 then adjusted to work on FreeBSD 8.4 as well.
2057 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2058 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2062 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2063 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2064 error code calls like this:
2066 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2068 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2069 that can be encoded in C. For the foreseeable future, this will only
2072 *Richard Levitte and Tim Hudson*
2074 * Removed BSD cryptodev engine.
2078 * Add a build target 'build_all_generated', to build all generated files
2079 and only that. This can be used to prepare everything that requires
2080 things like perl for a system that lacks perl and then move everything
2081 to that system and do the rest of the build there.
2085 * In the UI interface, make it possible to duplicate the user data. This
2086 can be used by engines that need to retain the data for a longer time
2087 than just the call where this user data is passed.
2091 * Ignore the '-named_curve auto' value for compatibility of applications
2094 *Tomas Mraz <tmraz@fedoraproject.org>*
2096 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2097 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2098 alerts across multiple records (some of which could be empty). In practice
2099 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2100 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2101 support this at all. Supporting it adds significant complexity to the
2102 record layer, and its removal is unlikely to cause interoperability
2107 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2108 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2109 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2114 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2115 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2117 *Richard Levitte, Andy Polyakov*
2119 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2124 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2125 platform rather than 'mingw'.
2129 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2130 success if they are asked to add an object which already exists
2131 in the store. This change cascades to other functions which load
2132 certificates and CRLs.
2136 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2137 facilitate stack unwinding even from assembly subroutines.
2141 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2142 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2146 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2147 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2148 which is the minimum version we support.
2152 * Certificate time validation (X509_cmp_time) enforces stricter
2153 compliance with RFC 5280. Fractional seconds and timezone offsets
2154 are no longer allowed.
2158 * Add support for ARIA
2162 * s_client will now send the Server Name Indication (SNI) extension by
2163 default unless the new "-noservername" option is used. The server name is
2164 based on the host provided to the "-connect" option unless overridden by
2165 using "-servername".
2169 * Add support for SipHash
2173 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2174 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2175 prevent issues where no progress is being made and the peer continually
2176 sends unrecognised record types, using up resources processing them.
2180 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2181 using the algorithm defined in
2182 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2186 * Heartbeat support has been removed; the ABI is changed for now.
2188 *Richard Levitte, Rich Salz*
2190 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2194 * The RSA "null" method, which was partially supported to avoid patent
2195 issues, has been replaced to always returns NULL.
2202 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2204 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2205 used even when parsing explicit parameters, when loading a encoded key
2206 or calling `EC_GROUP_new_from_ecpkparameters()`/
2207 `EC_GROUP_new_from_ecparameters()`.
2208 This prevents bypass of security hardening and performance gains,
2209 especially for curves with specialized EC_METHODs.
2210 By default, if a key encoded with explicit parameters is loaded and later
2211 encoded, the output is still encoded with explicit parameters, even if
2212 internally a "named" EC_GROUP is used for computation.
2216 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2217 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2218 NULL. After this change, only the cofactor parameter can be NULL. It also
2219 does some minimal sanity checks on the passed order.
2224 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2225 An attack is simple, if the first CMS_recipientInfo is valid but the
2226 second CMS_recipientInfo is chosen ciphertext. If the second
2227 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2228 encryption key will be replaced by garbage, and the message cannot be
2229 decoded, but if the RSA decryption fails, the correct encryption key is
2230 used and the recipient will not notice the attack.
2231 As a work around for this potential attack the length of the decrypted
2232 key must be equal to the cipher default key length, in case the
2233 certifiate is not given and all recipientInfo are tried out.
2234 The old behaviour can be re-enabled in the CMS code by setting the
2235 CMS_DEBUG_DECRYPT flag.
2240 * Use Windows installation paths in the mingw builds
2242 Mingw isn't a POSIX environment per se, which means that Windows
2243 paths should be used for installation.
2248 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2250 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2251 This changes the size when using the genpkey app when no size is given. It
2252 fixes an omission in earlier changes that changed all RSA, DSA and DH
2253 generation apps to use 2048 bits by default.
2257 * Prevent over long nonces in ChaCha20-Poly1305.
2259 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2260 for every encryption operation. RFC 7539 specifies that the nonce value
2261 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2262 and front pads the nonce with 0 bytes if it is less than 12
2263 bytes. However it also incorrectly allows a nonce to be set of up to 16
2264 bytes. In this case only the last 12 bytes are significant and any
2265 additional leading bytes are ignored.
2267 It is a requirement of using this cipher that nonce values are
2268 unique. Messages encrypted using a reused nonce value are susceptible to
2269 serious confidentiality and integrity attacks. If an application changes
2270 the default nonce length to be longer than 12 bytes and then makes a
2271 change to the leading bytes of the nonce expecting the new value to be a
2272 new unique nonce then such an application could inadvertently encrypt
2273 messages with a reused nonce.
2275 Additionally the ignored bytes in a long nonce are not covered by the
2276 integrity guarantee of this cipher. Any application that relies on the
2277 integrity of these ignored leading bytes of a long nonce may be further
2278 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2279 is safe because no such use sets such a long nonce value. However user
2280 applications that use this cipher directly and set a non-default nonce
2281 length to be longer than 12 bytes may be vulnerable.
2283 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2289 * Added SCA hardening for modular field inversion in EC_GROUP through
2290 a new dedicated field_inv() pointer in EC_METHOD.
2291 This also addresses a leakage affecting conversions from projective
2292 to affine coordinates.
2294 *Billy Bob Brumley, Nicola Tuveri*
2296 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2297 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2301 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2305 * Remove the 'dist' target and add a tarball building script. The
2306 'dist' target has fallen out of use, and it shouldn't be
2307 necessary to configure just to create a source distribution.
2311 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2313 * Timing vulnerability in DSA signature generation
2315 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2316 timing side channel attack. An attacker could use variations in the signing
2317 algorithm to recover the private key.
2319 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2324 * Timing vulnerability in ECDSA signature generation
2326 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2327 timing side channel attack. An attacker could use variations in the signing
2328 algorithm to recover the private key.
2330 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2335 * Add coordinate blinding for EC_POINT and implement projective
2336 coordinate blinding for generic prime curves as a countermeasure to
2337 chosen point SCA attacks.
2339 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2341 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2343 * Client DoS due to large DH parameter
2345 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2346 malicious server can send a very large prime value to the client. This will
2347 cause the client to spend an unreasonably long period of time generating a
2348 key for this prime resulting in a hang until the client has finished. This
2349 could be exploited in a Denial Of Service attack.
2351 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2356 * Cache timing vulnerability in RSA Key Generation
2358 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2359 a cache timing side channel attack. An attacker with sufficient access to
2360 mount cache timing attacks during the RSA key generation process could
2361 recover the private key.
2363 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2364 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2369 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2370 parameter is no longer accepted, as it leads to a corrupt table. NULL
2371 pem_str is reserved for alias entries only.
2375 * Revert blinding in ECDSA sign and instead make problematic addition
2376 length-invariant. Switch even to fixed-length Montgomery multiplication.
2380 * Change generating and checking of primes so that the error rate of not
2381 being prime depends on the intended use based on the size of the input.
2382 For larger primes this will result in more rounds of Miller-Rabin.
2383 The maximal error rate for primes with more than 1080 bits is lowered
2386 *Kurt Roeckx, Annie Yousar*
2388 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2392 * Add blinding to ECDSA and DSA signatures to protect against side channel
2393 attacks discovered by Keegan Ryan (NCC Group).
2397 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2398 now allow empty (zero character) pass phrases.
2402 * Certificate time validation (X509_cmp_time) enforces stricter
2403 compliance with RFC 5280. Fractional seconds and timezone offsets
2404 are no longer allowed.
2408 * Fixed a text canonicalisation bug in CMS
2410 Where a CMS detached signature is used with text content the text goes
2411 through a canonicalisation process first prior to signing or verifying a
2412 signature. This process strips trailing space at the end of lines, converts
2413 line terminators to CRLF and removes additional trailing line terminators
2414 at the end of a file. A bug in the canonicalisation process meant that
2415 some characters, such as form-feed, were incorrectly treated as whitespace
2416 and removed. This is contrary to the specification (RFC5485). This fix
2417 could mean that detached text data signed with an earlier version of
2418 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2419 signed with a fixed OpenSSL may fail to verify with an earlier version of
2420 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2421 and use the "-binary" flag (for the "cms" command line application) or set
2422 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2426 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2428 * Constructed ASN.1 types with a recursive definition could exceed the stack
2430 Constructed ASN.1 types with a recursive definition (such as can be found
2431 in PKCS7) could eventually exceed the stack given malicious input with
2432 excessive recursion. This could result in a Denial Of Service attack. There
2433 are no such structures used within SSL/TLS that come from untrusted sources
2434 so this is considered safe.
2436 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2442 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2444 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2445 effectively reduced to only comparing the least significant bit of each
2446 byte. This allows an attacker to forge messages that would be considered as
2447 authenticated in an amount of tries lower than that guaranteed by the
2448 security claims of the scheme. The module can only be compiled by the
2449 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2451 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2457 * Add a build target 'build_all_generated', to build all generated files
2458 and only that. This can be used to prepare everything that requires
2459 things like perl for a system that lacks perl and then move everything
2460 to that system and do the rest of the build there.
2464 * Backport SSL_OP_NO_RENGOTIATION
2466 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2467 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2468 changes this is no longer possible in 1.1.0. Therefore the new
2469 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2470 1.1.0 to provide equivalent functionality.
2472 Note that if an application built against 1.1.0h headers (or above) is run
2473 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2474 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2478 * Removed the OS390-Unix config target. It relied on a script that doesn't
2483 * rsaz_1024_mul_avx2 overflow bug on x86_64
2485 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2486 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2487 Analysis suggests that attacks against RSA and DSA as a result of this
2488 defect would be very difficult to perform and are not believed likely.
2489 Attacks against DH1024 are considered just feasible, because most of the
2490 work necessary to deduce information about a private key may be performed
2491 offline. The amount of resources required for such an attack would be
2492 significant. However, for an attack on TLS to be meaningful, the server
2493 would have to share the DH1024 private key among multiple clients, which is
2494 no longer an option since CVE-2016-0701.
2496 This only affects processors that support the AVX2 but not ADX extensions
2497 like Intel Haswell (4th generation).
2499 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2500 was originally found via the OSS-Fuzz project.
2505 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2507 * bn_sqrx8x_internal carry bug on x86_64
2509 There is a carry propagating bug in the x86_64 Montgomery squaring
2510 procedure. No EC algorithms are affected. Analysis suggests that attacks
2511 against RSA and DSA as a result of this defect would be very difficult to
2512 perform and are not believed likely. Attacks against DH are considered just
2513 feasible (although very difficult) because most of the work necessary to
2514 deduce information about a private key may be performed offline. The amount
2515 of resources required for such an attack would be very significant and
2516 likely only accessible to a limited number of attackers. An attacker would
2517 additionally need online access to an unpatched system using the target
2518 private key in a scenario with persistent DH parameters and a private
2519 key that is shared between multiple clients.
2521 This only affects processors that support the BMI1, BMI2 and ADX extensions
2522 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2524 This issue was reported to OpenSSL by the OSS-Fuzz project.
2529 * Malformed X.509 IPAddressFamily could cause OOB read
2531 If an X.509 certificate has a malformed IPAddressFamily extension,
2532 OpenSSL could do a one-byte buffer overread. The most likely result
2533 would be an erroneous display of the certificate in text format.
2535 This issue was reported to OpenSSL by the OSS-Fuzz project.
2540 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2542 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2543 platform rather than 'mingw'.
2547 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2548 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2549 which is the minimum version we support.
2553 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2555 * Encrypt-Then-Mac renegotiation crash
2557 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2558 negotiated where it was not in the original handshake (or vice-versa) then
2559 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2560 and servers are affected.
2562 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2567 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2569 * Truncated packet could crash via OOB read
2571 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2572 cipher is being used, then a truncated packet can cause that host to
2573 perform an out-of-bounds read, usually resulting in a crash.
2575 This issue was reported to OpenSSL by Robert Święcki of Google.
2580 * Bad (EC)DHE parameters cause a client crash
2582 If a malicious server supplies bad parameters for a DHE or ECDHE key
2583 exchange then this can result in the client attempting to dereference a
2584 NULL pointer leading to a client crash. This could be exploited in a Denial
2587 This issue was reported to OpenSSL by Guido Vranken.
2592 * BN_mod_exp may produce incorrect results on x86_64
2594 There is a carry propagating bug in the x86_64 Montgomery squaring
2595 procedure. No EC algorithms are affected. Analysis suggests that attacks
2596 against RSA and DSA as a result of this defect would be very difficult to
2597 perform and are not believed likely. Attacks against DH are considered just
2598 feasible (although very difficult) because most of the work necessary to
2599 deduce information about a private key may be performed offline. The amount
2600 of resources required for such an attack would be very significant and
2601 likely only accessible to a limited number of attackers. An attacker would
2602 additionally need online access to an unpatched system using the target
2603 private key in a scenario with persistent DH parameters and a private
2604 key that is shared between multiple clients. For example this can occur by
2605 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2606 similar to CVE-2015-3193 but must be treated as a separate problem.
2608 This issue was reported to OpenSSL by the OSS-Fuzz project.
2613 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2615 * ChaCha20/Poly1305 heap-buffer-overflow
2617 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2618 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2619 crash. This issue is not considered to be exploitable beyond a DoS.
2621 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2626 * CMS Null dereference
2628 Applications parsing invalid CMS structures can crash with a NULL pointer
2629 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2630 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2631 structure callback if an attempt is made to free certain invalid encodings.
2632 Only CHOICE structures using a callback which do not handle NULL value are
2635 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2640 * Montgomery multiplication may produce incorrect results
2642 There is a carry propagating bug in the Broadwell-specific Montgomery
2643 multiplication procedure that handles input lengths divisible by, but
2644 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2645 and DH private keys are impossible. This is because the subroutine in
2646 question is not used in operations with the private key itself and an input
2647 of the attacker's direct choice. Otherwise the bug can manifest itself as
2648 transient authentication and key negotiation failures or reproducible
2649 erroneous outcome of public-key operations with specially crafted input.
2650 Among EC algorithms only Brainpool P-512 curves are affected and one
2651 presumably can attack ECDH key negotiation. Impact was not analyzed in
2652 detail, because pre-requisites for attack are considered unlikely. Namely
2653 multiple clients have to choose the curve in question and the server has to
2654 share the private key among them, neither of which is default behaviour.
2655 Even then only clients that chose the curve will be affected.
2657 This issue was publicly reported as transient failures and was not
2658 initially recognized as a security issue. Thanks to Richard Morgan for
2659 providing reproducible case.
2664 * Removed automatic addition of RPATH in shared libraries and executables,
2665 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2669 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2671 * Fix Use After Free for large message sizes
2673 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2674 message larger than approx 16k is received then the underlying buffer to
2675 store the incoming message is reallocated and moved. Unfortunately a
2676 dangling pointer to the old location is left which results in an attempt to
2677 write to the previously freed location. This is likely to result in a
2678 crash, however it could potentially lead to execution of arbitrary code.
2680 This issue only affects OpenSSL 1.1.0a.
2682 This issue was reported to OpenSSL by Robert Święcki.
2687 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2689 * OCSP Status Request extension unbounded memory growth
2691 A malicious client can send an excessively large OCSP Status Request
2692 extension. If that client continually requests renegotiation, sending a
2693 large OCSP Status Request extension each time, then there will be unbounded
2694 memory growth on the server. This will eventually lead to a Denial Of
2695 Service attack through memory exhaustion. Servers with a default
2696 configuration are vulnerable even if they do not support OCSP. Builds using
2697 the "no-ocsp" build time option are not affected.
2699 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2704 * SSL_peek() hang on empty record
2706 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2707 sends an empty record. This could be exploited by a malicious peer in a
2708 Denial Of Service attack.
2710 This issue was reported to OpenSSL by Alex Gaynor.
2715 * Excessive allocation of memory in tls_get_message_header() and
2716 dtls1_preprocess_fragment()
2718 A (D)TLS message includes 3 bytes for its length in the header for the
2719 message. This would allow for messages up to 16Mb in length. Messages of
2720 this length are excessive and OpenSSL includes a check to ensure that a
2721 peer is sending reasonably sized messages in order to avoid too much memory
2722 being consumed to service a connection. A flaw in the logic of version
2723 1.1.0 means that memory for the message is allocated too early, prior to
2724 the excessive message length check. Due to way memory is allocated in
2725 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2726 to service a connection. This could lead to a Denial of Service through
2727 memory exhaustion. However, the excessive message length check still takes
2728 place, and this would cause the connection to immediately fail. Assuming
2729 that the application calls SSL_free() on the failed connection in a timely
2730 manner then the 21Mb of allocated memory will then be immediately freed
2731 again. Therefore the excessive memory allocation will be transitory in
2732 nature. This then means that there is only a security impact if:
2734 1) The application does not call SSL_free() in a timely manner in the event
2735 that the connection fails
2737 2) The application is working in a constrained environment where there is
2738 very little free memory
2740 3) The attacker initiates multiple connection attempts such that there are
2741 multiple connections in a state where memory has been allocated for the
2742 connection; SSL_free() has not yet been called; and there is insufficient
2743 memory to service the multiple requests.
2745 Except in the instance of (1) above any Denial Of Service is likely to be
2746 transitory because as soon as the connection fails the memory is
2747 subsequently freed again in the SSL_free() call. However there is an
2748 increased risk during this period of application crashes due to the lack of
2749 memory - which would then mean a more serious Denial of Service.
2751 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2752 (CVE-2016-6307 and CVE-2016-6308)
2756 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2757 had to be removed. Primary reason is that vendor assembler can't
2758 assemble our modules with -KPIC flag. As result it, assembly
2759 support, was not even available as option. But its lack means
2760 lack of side-channel resistant code, which is incompatible with
2761 security by todays standards. Fortunately gcc is readily available
2762 prepackaged option, which we firmly point at...
2766 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2768 * Windows command-line tool supports UTF-8 opt-in option for arguments
2769 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2770 (to any value) allows Windows user to access PKCS#12 file generated
2771 with Windows CryptoAPI and protected with non-ASCII password, as well
2772 as files generated under UTF-8 locale on Linux also protected with
2777 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2778 have been disabled by default and removed from DEFAULT, just like RC4.
2779 See the RC4 item below to re-enable both.
2783 * The method for finding the storage location for the Windows RAND seed file
2784 has changed. First we check %RANDFILE%. If that is not set then we check
2785 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2786 all else fails we fall back to C:\.
2790 * The EVP_EncryptUpdate() function has had its return type changed from void
2791 to int. A return of 0 indicates and error while a return of 1 indicates
2796 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2797 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2798 off the constant time implementation for RSA, DSA and DH have been made
2799 no-ops and deprecated.
2803 * Windows RAND implementation was simplified to only get entropy by
2804 calling CryptGenRandom(). Various other RAND-related tickets
2807 *Joseph Wylie Yandle, Rich Salz*
2809 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2810 and `OPENSSL_LH_`, respectively. The old names are available
2811 with API compatibility. They new names are now completely documented.
2815 * Unify TYPE_up_ref(obj) methods signature.
2816 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2817 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2818 int (instead of void) like all others TYPE_up_ref() methods.
2819 So now these methods also check the return value of CRYPTO_atomic_add(),
2820 and the validity of object reference counter.
2822 *fdasilvayy@gmail.com*
2824 * With Windows Visual Studio builds, the .pdb files are installed
2825 alongside the installed libraries and executables. For a static
2826 library installation, ossl_static.pdb is the associate compiler
2827 generated .pdb file to be used when linking programs.
2831 * Remove openssl.spec. Packaging files belong with the packagers.
2835 * Automatic Darwin/OSX configuration has had a refresh, it will now
2836 recognise x86_64 architectures automatically. You can still decide
2837 to build for a different bitness with the environment variable
2838 KERNEL_BITS (can be 32 or 64), for example:
2840 KERNEL_BITS=32 ./config
2844 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2845 256 bit AES and HMAC with SHA256.
2849 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2853 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2857 * To enable users to have their own config files and build file templates,
2858 Configure looks in the directory indicated by the environment variable
2859 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2860 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2861 name and is used as is.
2865 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2866 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2867 X509_CERT_FILE_CTX was removed.
2871 * "shared" builds are now the default. To create only static libraries use
2872 the "no-shared" Configure option.
2876 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2877 All of these option have not worked for some while and are fundamental
2882 * Make various cleanup routines no-ops and mark them as deprecated. Most
2883 global cleanup functions are no longer required because they are handled
2884 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2885 Explicitly de-initing can cause problems (e.g. where a library that uses
2886 OpenSSL de-inits, but an application is still using it). The affected
2887 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2888 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2889 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2890 COMP_zlib_cleanup().
2894 * --strict-warnings no longer enables runtime debugging options
2895 such as REF_DEBUG. Instead, debug options are automatically
2896 enabled with '--debug' builds.
2898 *Andy Polyakov, Emilia Käsper*
2900 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2901 have been moved out of the public header files. New functions for managing
2902 these have been added.
2906 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2907 objects have been moved out of the public header files. New
2908 functions for managing these have been added.
2912 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2913 have been moved out of the public header files. New functions for managing
2914 these have been added.
2918 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2919 moved out of the public header files. New functions for managing these
2924 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2928 * Removed the mk1mf build scripts.
2932 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2933 it is always safe to #include a header now.
2937 * Removed the aged BC-32 config and all its supporting scripts
2941 * Removed support for Ultrix, Netware, and OS/2.
2945 * Add support for HKDF.
2947 *Alessandro Ghedini*
2949 * Add support for blake2b and blake2s
2953 * Added support for "pipelining". Ciphers that have the
2954 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2955 encryptions/decryptions simultaneously. There are currently no built-in
2956 ciphers with this property but the expectation is that engines will be able
2957 to offer it to significantly improve throughput. Support has been extended
2958 into libssl so that multiple records for a single connection can be
2959 processed in one go (for >=TLS 1.1).
2963 * Added the AFALG engine. This is an async capable engine which is able to
2964 offload work to the Linux kernel. In this initial version it only supports
2965 AES128-CBC. The kernel must be version 4.1.0 or greater.
2969 * OpenSSL now uses a new threading API. It is no longer necessary to
2970 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2971 are two supported threading models: pthreads and windows threads. It is
2972 also possible to configure OpenSSL at compile time for "no-threads". The
2973 old threading API should no longer be used. The functions have been
2974 replaced with "no-op" compatibility macros.
2976 *Alessandro Ghedini, Matt Caswell*
2978 * Modify behavior of ALPN to invoke callback after SNI/servername
2979 callback, such that updates to the SSL_CTX affect ALPN.
2983 * Add SSL_CIPHER queries for authentication and key-exchange.
2987 * Changes to the DEFAULT cipherlist:
2988 - Prefer (EC)DHE handshakes over plain RSA.
2989 - Prefer AEAD ciphers over legacy ciphers.
2990 - Prefer ECDSA over RSA when both certificates are available.
2991 - Prefer TLSv1.2 ciphers/PRF.
2992 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2997 * Change the ECC default curve list to be this, in order: x25519,
2998 secp256r1, secp521r1, secp384r1.
3002 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3003 disabled by default. They can be re-enabled using the
3004 enable-weak-ssl-ciphers option to Configure.
3008 * If the server has ALPN configured, but supports no protocols that the
3009 client advertises, send a fatal "no_application_protocol" alert.
3010 This behaviour is SHALL in RFC 7301, though it isn't universally
3011 implemented by other servers.
3015 * Add X25519 support.
3016 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3017 for public and private key encoding using the format documented in
3018 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3019 key generation and key derivation.
3021 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3026 * Deprecate SRP_VBASE_get_by_user.
3027 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3028 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3029 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3030 seed, even if the seed is configured.
3032 Users should use SRP_VBASE_get1_by_user instead. Note that in
3033 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3034 also that even though configuring the SRP seed attempts to hide
3035 invalid usernames by continuing the handshake with fake
3036 credentials, this behaviour is not constant time and no strong
3037 guarantees are made that the handshake is indistinguishable from
3038 that of a valid user.
3042 * Configuration change; it's now possible to build dynamic engines
3043 without having to build shared libraries and vice versa. This
3044 only applies to the engines in engines/, those in crypto/engine/
3045 will always be built into libcrypto (i.e. "static").
3047 Building dynamic engines is enabled by default; to disable, use
3048 the configuration option "disable-dynamic-engine".
3050 The only requirements for building dynamic engines are the
3051 presence of the DSO module and building with position independent
3052 code, so they will also automatically be disabled if configuring
3053 with "disable-dso" or "disable-pic".
3055 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3056 are also taken away from openssl/opensslconf.h, as they are
3061 * Configuration change; if there is a known flag to compile
3062 position independent code, it will always be applied on the
3063 libcrypto and libssl object files, and never on the application
3064 object files. This means other libraries that use routines from
3065 libcrypto / libssl can be made into shared libraries regardless
3066 of how OpenSSL was configured.
3068 If this isn't desirable, the configuration options "disable-pic"
3069 or "no-pic" can be used to disable the use of PIC. This will
3070 also disable building shared libraries and dynamic engines.
3074 * Removed JPAKE code. It was experimental and has no wide use.
3078 * The INSTALL_PREFIX Makefile variable has been renamed to
3079 DESTDIR. That makes for less confusion on what this variable
3080 is for. Also, the configuration option --install_prefix is
3085 * Heartbeat for TLS has been removed and is disabled by default
3086 for DTLS; configure with enable-heartbeats. Code that uses the
3087 old #define's might need to be updated.
3089 *Emilia Käsper, Rich Salz*
3091 * Rename REF_CHECK to REF_DEBUG.
3095 * New "unified" build system
3097 The "unified" build system is aimed to be a common system for all
3098 platforms we support. With it comes new support for VMS.
3100 This system builds supports building in a different directory tree
3101 than the source tree. It produces one Makefile (for unix family
3102 or lookalikes), or one descrip.mms (for VMS).
3104 The source of information to make the Makefile / descrip.mms is
3105 small files called 'build.info', holding the necessary
3106 information for each directory with source to compile, and a
3107 template in Configurations, like unix-Makefile.tmpl or
3110 With this change, the library names were also renamed on Windows
3111 and on VMS. They now have names that are closer to the standard
3112 on Unix, and include the major version number, and in certain
3113 cases, the architecture they are built for. See "Notes on shared
3114 libraries" in INSTALL.
3116 We rely heavily on the perl module Text::Template.
3120 * Added support for auto-initialisation and de-initialisation of the library.
3121 OpenSSL no longer requires explicit init or deinit routines to be called,
3122 except in certain circumstances. See the OPENSSL_init_crypto() and
3123 OPENSSL_init_ssl() man pages for further information.
3127 * The arguments to the DTLSv1_listen function have changed. Specifically the
3128 "peer" argument is now expected to be a BIO_ADDR object.
3130 * Rewrite of BIO networking library. The BIO library lacked consistent
3131 support of IPv6, and adding it required some more extensive
3132 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3133 which hold all types of addresses and chains of address information.
3134 It also introduces a new API, with functions like BIO_socket,
3135 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3136 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3137 have been adapted accordingly.
3141 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3146 * CRIME protection: disable compression by default, even if OpenSSL is
3147 compiled with zlib enabled. Applications can still enable compression
3148 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3149 using the SSL_CONF library to configure compression.
3153 * The signature of the session callback configured with
3154 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3155 was explicitly marked as `const unsigned char*` instead of
3160 * Always DPURIFY. Remove the use of uninitialized memory in the
3161 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3165 * Removed many obsolete configuration items, including
3166 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3167 MD2_CHAR, MD2_INT, MD2_LONG
3169 IDEA_SHORT, IDEA_LONG
3170 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3172 *Rich Salz, with advice from Andy Polyakov*
3174 * Many BN internals have been moved to an internal header file.
3176 *Rich Salz with help from Andy Polyakov*
3178 * Configuration and writing out the results from it has changed.
3179 Files such as Makefile include/openssl/opensslconf.h and are now
3180 produced through general templates, such as Makefile.in and
3181 crypto/opensslconf.h.in and some help from the perl module
3184 Also, the center of configuration information is no longer
3185 Makefile. Instead, Configure produces a perl module in
3186 configdata.pm which holds most of the config data (in the hash
3187 table %config), the target data that comes from the target
3188 configuration in one of the `Configurations/*.conf` files (in
3193 * To clarify their intended purposes, the Configure options
3194 --prefix and --openssldir change their semantics, and become more
3195 straightforward and less interdependent.
3197 --prefix shall be used exclusively to give the location INSTALLTOP
3198 where programs, scripts, libraries, include files and manuals are
3199 going to be installed. The default is now /usr/local.
3201 --openssldir shall be used exclusively to give the default
3202 location OPENSSLDIR where certificates, private keys, CRLs are
3203 managed. This is also where the default openssl.cnf gets
3205 If the directory given with this option is a relative path, the
3206 values of both the --prefix value and the --openssldir value will
3207 be combined to become OPENSSLDIR.
3208 The default for --openssldir is INSTALLTOP/ssl.
3210 Anyone who uses --openssldir to specify where OpenSSL is to be
3211 installed MUST change to use --prefix instead.
3215 * The GOST engine was out of date and therefore it has been removed. An up
3216 to date GOST engine is now being maintained in an external repository.
3217 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3218 support for GOST ciphersuites (these are only activated if a GOST engine
3223 * EGD is no longer supported by default; use enable-egd when
3226 *Ben Kaduk and Rich Salz*
3228 * The distribution now has Makefile.in files, which are used to
3229 create Makefile's when Configure is run. *Configure must be run
3230 before trying to build now.*
3234 * The return value for SSL_CIPHER_description() for error conditions
3239 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3241 Obtaining and performing DNSSEC validation of TLSA records is
3242 the application's responsibility. The application provides
3243 the TLSA records of its choice to OpenSSL, and these are then
3244 used to authenticate the peer.
3246 The TLSA records need not even come from DNS. They can, for
3247 example, be used to implement local end-entity certificate or
3248 trust-anchor "pinning", where the "pin" data takes the form
3249 of TLSA records, which can augment or replace verification
3250 based on the usual WebPKI public certification authorities.
3254 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3255 continues to support deprecated interfaces in default builds.
3256 However, applications are strongly advised to compile their
3257 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3258 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3259 or the 1.1.0 releases.
3261 In environments in which all applications have been ported to
3262 not use any deprecated interfaces OpenSSL's Configure script
3263 should be used with the --api=1.1.0 option to entirely remove
3264 support for the deprecated features from the library and
3265 unconditionally disable them in the installed headers.
3266 Essentially the same effect can be achieved with the "no-deprecated"
3267 argument to Configure, except that this will always restrict
3268 the build to just the latest API, rather than a fixed API
3271 As applications are ported to future revisions of the API,
3272 they should update their compile-time OPENSSL_API_COMPAT define
3273 accordingly, but in most cases should be able to continue to
3274 compile with later releases.
3276 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3277 0x10000000L and 0x00908000L, respectively. However those
3278 versions did not support the OPENSSL_API_COMPAT feature, and
3279 so applications are not typically tested for explicit support
3280 of just the undeprecated features of either release.
3284 * Add support for setting the minimum and maximum supported protocol.
3285 It can bet set via the SSL_set_min_proto_version() and
3286 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3287 MaxProtocol. It's recommended to use the new APIs to disable
3288 protocols instead of disabling individual protocols using
3289 SSL_set_options() or SSL_CONF's Protocol. This change also
3290 removes support for disabling TLS 1.2 in the OpenSSL TLS
3291 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3295 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3299 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3300 and integrates ECDSA and ECDH functionality into EC. Implementations can
3301 now redirect key generation and no longer need to convert to or from
3304 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3305 include the ec.h header file instead.
3309 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3310 ciphers who are no longer supported and drops support the ephemeral RSA key
3311 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3315 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3316 opaque. For HMAC_CTX, the following constructors and destructors
3319 HMAC_CTX *HMAC_CTX_new(void);
3320 void HMAC_CTX_free(HMAC_CTX *ctx);
3322 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3323 destroy such methods has been added. See EVP_MD_meth_new(3) and
3324 EVP_CIPHER_meth_new(3) for documentation.
3327 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3328 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3329 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3330 an already created structure.
3331 2) For consistency with the majority of our object creators and
3332 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3333 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3334 for deprecated builds.
3338 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3339 cryptographic operations to be performed asynchronously as long as an
3340 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3341 further details. Libssl has also had this capability integrated with the
3342 introduction of the new mode SSL_MODE_ASYNC and associated error
3343 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3344 pages. This work was developed in partnership with Intel Corp.
3348 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3349 always enabled now. If you want to disable the support you should
3350 exclude it using the list of supported ciphers. This also means that the
3351 "-no_ecdhe" option has been removed from s_server.
3355 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3356 SSL_{CTX_}set1_curves() which can set a list.
3360 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3361 curve you want to support using SSL_{CTX_}set1_curves().
3365 * State machine rewrite. The state machine code has been significantly
3366 refactored in order to remove much duplication of code and solve issues
3367 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3368 further details). This change does have some associated API changes.
3369 Notably the SSL_state() function has been removed and replaced by
3370 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3371 SSL_set_state() has been removed altogether. The previous handshake states
3372 defined in ssl.h and ssl3.h have also been removed.
3376 * All instances of the string "ssleay" in the public API were replaced
3377 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3378 Some error codes related to internal RSA_eay API's were renamed.
3382 * The demo files in crypto/threads were moved to demo/threads.
3386 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3389 *Matt Caswell, Rich Salz*
3391 * New ASN.1 embed macro.
3393 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3394 structure is not allocated: it is part of the parent. That is instead of
3402 This reduces memory fragmentation and make it impossible to accidentally
3403 set a mandatory field to NULL.
3405 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3406 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3407 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3412 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3416 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3417 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3418 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3419 DES and RC4 ciphersuites.
3423 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3424 This changes the decoding behaviour for some invalid messages,
3425 though the change is mostly in the more lenient direction, and
3426 legacy behaviour is preserved as much as possible.
3430 * Fix no-stdio build.
3431 *David Woodhouse <David.Woodhouse@intel.com> and also*
3432 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3434 * New testing framework
3435 The testing framework has been largely rewritten and is now using
3436 perl and the perl modules Test::Harness and an extended variant of
3437 Test::More called OpenSSL::Test to do its work. All test scripts in
3438 test/ have been rewritten into test recipes, and all direct calls to
3439 executables in test/Makefile have become individual recipes using the
3440 simplified testing OpenSSL::Test::Simple.
3442 For documentation on our testing modules, do:
3444 perldoc test/testlib/OpenSSL/Test/Simple.pm
3445 perldoc test/testlib/OpenSSL/Test.pm
3449 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3450 are used; the latter aborts on memory leaks (usually checked on exit).
3451 Some undocumented "set malloc, etc., hooks" functions were removed
3452 and others were changed. All are now documented.
3456 * In DSA_generate_parameters_ex, if the provided seed is too short,
3459 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3461 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3462 from RFC4279, RFC4785, RFC5487, RFC5489.
3464 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3465 original RSA_PSK patch.
3469 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3470 era flag was never set throughout the codebase (only read). Also removed
3471 SSL3_FLAGS_POP_BUFFER which was only used if
3472 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3476 * Changed the default name options in the "ca", "crl", "req" and "x509"
3477 to be "oneline" instead of "compat".
3481 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3482 not aware of clients that still exhibit this bug, and the workaround
3483 hasn't been working properly for a while.
3487 * The return type of BIO_number_read() and BIO_number_written() as well as
3488 the corresponding num_read and num_write members in the BIO structure has
3489 changed from unsigned long to uint64_t. On platforms where an unsigned
3490 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3495 * Given the pervasive nature of TLS extensions it is inadvisable to run
3496 OpenSSL without support for them. It also means that maintaining
3497 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3498 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3502 * Removed support for the two export grade static DH ciphersuites
3503 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3504 were newly added (along with a number of other static DH ciphersuites) to
3505 1.0.2. However the two export ones have *never* worked since they were
3506 introduced. It seems strange in any case to be adding new export
3507 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3511 * Version negotiation has been rewritten. In particular SSLv23_method(),
3512 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3513 and turned into macros which simply call the new preferred function names
3514 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3515 should use the new names instead. Also as part of this change the ssl23.h
3516 header file has been removed.
3520 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3521 code and the associated standard is no longer considered fit-for-purpose.
3525 * RT2547 was closed. When generating a private key, try to make the
3526 output file readable only by the owner. This behavior change might
3527 be noticeable when interacting with other software.
3529 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3534 * Added HTTP GET support to the ocsp command.
3538 * Changed default digest for the dgst and enc commands from MD5 to
3543 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3547 * Added support for TLS extended master secret from
3548 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3549 initial patch which was a great help during development.
3553 * All libssl internal structures have been removed from the public header
3554 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3555 now redundant). Users should not attempt to access internal structures
3556 directly. Instead they should use the provided API functions.
3560 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3561 Access to deprecated functions can be re-enabled by running config with
3562 "enable-deprecated". In addition applications wishing to use deprecated
3563 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3564 will, by default, disable some transitive includes that previously existed
3565 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3569 * Added support for OCB mode. OpenSSL has been granted a patent license
3570 compatible with the OpenSSL license for use of OCB. Details are available
3571 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3572 for OCB can be removed by calling config with no-ocb.
3576 * SSLv2 support has been removed. It still supports receiving a SSLv2
3577 compatible client hello.
3581 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3582 done while fixing the error code for the key-too-small case.
3584 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3586 * CA.sh has been removed; use CA.pl instead.
3590 * Removed old DES API.
3594 * Remove various unsupported platforms:
3600 Sinix/ReliantUNIX RM400
3605 16-bit platforms such as WIN16
3609 * Clean up OPENSSL_NO_xxx #define's
3610 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3611 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3612 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3613 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3614 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3615 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3616 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3617 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3618 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3619 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3623 * Cleaned up dead code
3624 Remove all but one '#ifdef undef' which is to be looked at.
3628 * Clean up calling of xxx_free routines.
3629 Just like free(), fix most of the xxx_free routines to accept
3630 NULL. Remove the non-null checks from callers. Save much code.
3634 * Add secure heap for storage of private keys (when possible).
3635 Add BIO_s_secmem(), CBIGNUM, etc.
3636 Contributed by Akamai Technologies under our Corporate CLA.
3640 * Experimental support for a new, fast, unbiased prime candidate generator,
3641 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3643 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3645 * New output format NSS in the sess_id command line tool. This allows
3646 exporting the session id and the master key in NSS keylog format.
3648 *Martin Kaiser <martin@kaiser.cx>*
3650 * Harmonize version and its documentation. -f flag is used to display
3653 *mancha <mancha1@zoho.com>*
3655 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3656 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3658 *mancha <mancha1@zoho.com>*
3660 * Fix some double frees. These are not thought to be exploitable.
3662 *mancha <mancha1@zoho.com>*
3664 * A missing bounds check in the handling of the TLS heartbeat extension
3665 can be used to reveal up to 64k of memory to a connected client or
3668 Thanks for Neel Mehta of Google Security for discovering this bug and to
3669 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3670 preparing the fix ([CVE-2014-0160])
3672 *Adam Langley, Bodo Moeller*
3674 * Fix for the attack described in the paper "Recovering OpenSSL
3675 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3676 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3677 <http://eprint.iacr.org/2014/140>
3679 Thanks to Yuval Yarom and Naomi Benger for discovering this
3680 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3682 *Yuval Yarom and Naomi Benger*
3684 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3685 this fixes a limitation in previous versions of OpenSSL.
3689 * Experimental encrypt-then-mac support.
3691 Experimental support for encrypt then mac from
3692 draft-gutmann-tls-encrypt-then-mac-02.txt
3694 To enable it set the appropriate extension number (0x42 for the test
3695 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3697 For non-compliant peers (i.e. just about everything) this should have no
3700 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3704 * Add EVP support for key wrapping algorithms, to avoid problems with
3705 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3706 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3707 algorithms and include tests cases.
3711 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3716 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3717 MGF1 digest and OAEP label.
3721 * Make openssl verify return errors.
3723 *Chris Palmer <palmer@google.com> and Ben Laurie*
3725 * New function ASN1_TIME_diff to calculate the difference between two
3726 ASN1_TIME structures or one structure and the current time.
3730 * Update fips_test_suite to support multiple command line options. New
3731 test to induce all self test errors in sequence and check expected
3736 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3737 sign or verify all in one operation.
3741 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3742 test programs and fips_test_suite. Includes functionality to parse
3743 the minimal script output of fipsalgest.pl directly.
3747 * Add authorisation parameter to FIPS_module_mode_set().
3751 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3755 * Use separate DRBG fields for internal and external flags. New function
3756 FIPS_drbg_health_check() to perform on demand health checking. Add
3757 generation tests to fips_test_suite with reduced health check interval to
3758 demonstrate periodic health checking. Add "nodh" option to
3759 fips_test_suite to skip very slow DH test.
3763 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3768 * More extensive health check for DRBG checking many more failure modes.
3769 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3770 combination: call this in fips_test_suite.
3774 * Add support for canonical generation of DSA parameter 'g'. See
3777 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3778 POST to handle HMAC cases.
3782 * Add functions FIPS_module_version() and FIPS_module_version_text()
3783 to return numerical and string versions of the FIPS module number.
3787 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3788 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3789 outside the validated module in the FIPS capable OpenSSL.
3793 * Minor change to DRBG entropy callback semantics. In some cases
3794 there is no multiple of the block length between min_len and
3795 max_len. Allow the callback to return more than max_len bytes
3796 of entropy but discard any extra: it is the callback's responsibility
3797 to ensure that the extra data discarded does not impact the
3798 requested amount of entropy.
3802 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3803 information in FIPS186-3, SP800-57 and SP800-131A.
3807 * CCM support via EVP. Interface is very similar to GCM case except we
3808 must supply all data in one chunk (i.e. no update, final) and the
3809 message length must be supplied if AAD is used. Add algorithm test
3814 * Initial version of POST overhaul. Add POST callback to allow the status
3815 of POST to be monitored and/or failures induced. Modify fips_test_suite
3816 to use callback. Always run all selftests even if one fails.
3820 * XTS support including algorithm test driver in the fips_gcmtest program.
3821 Note: this does increase the maximum key length from 32 to 64 bytes but
3822 there should be no binary compatibility issues as existing applications
3823 will never use XTS mode.
3827 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3828 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3829 performs algorithm blocking for unapproved PRNG types. Also do not
3830 set PRNG type in FIPS_mode_set(): leave this to the application.
3831 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3832 the standard OpenSSL PRNG: set additional data to a date time vector.
3836 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3837 This shouldn't present any incompatibility problems because applications
3838 shouldn't be using these directly and any that are will need to rethink
3839 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3843 * Extensive self tests and health checking required by SP800-90 DRBG.
3844 Remove strength parameter from FIPS_drbg_instantiate and always
3845 instantiate at maximum supported strength.
3849 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3853 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3857 * New function DH_compute_key_padded() to compute a DH key and pad with
3858 leading zeroes if needed: this complies with SP800-56A et al.
3862 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3863 anything, incomplete, subject to change and largely untested at present.
3867 * Modify fipscanisteronly build option to only build the necessary object
3868 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3872 * Add experimental option FIPSSYMS to give all symbols in
3873 fipscanister.o and FIPS or fips prefix. This will avoid
3874 conflicts with future versions of OpenSSL. Add perl script
3875 util/fipsas.pl to preprocess assembly language source files
3876 and rename any affected symbols.
3880 * Add selftest checks and algorithm block of non-fips algorithms in
3881 FIPS mode. Remove DES2 from selftests.
3885 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3886 return internal method without any ENGINE dependencies. Add new
3887 tiny fips sign and verify functions.
3891 * New build option no-ec2m to disable characteristic 2 code.
3895 * New build option "fipscanisteronly". This only builds fipscanister.o
3896 and (currently) associated fips utilities. Uses the file Makefile.fips
3897 instead of Makefile.org as the prototype.
3901 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3902 Update fips_gcmtest to use IV generator.
3906 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3907 setting output buffer to NULL. The `*Final` function must be
3908 called although it will not retrieve any additional data. The tag
3909 can be set or retrieved with a ctrl. The IV length is by default 12
3910 bytes (96 bits) but can be set to an alternative value. If the IV
3911 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3916 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3917 underlying do_cipher function handles all cipher semantics itself
3918 including padding and finalisation. This is useful if (for example)
3919 an ENGINE cipher handles block padding itself. The behaviour of
3920 do_cipher is subtly changed if this flag is set: the return value
3921 is the number of characters written to the output buffer (zero is
3922 no longer an error code) or a negative error code. Also if the
3923 input buffer is NULL and length 0 finalisation should be performed.
3927 * If a candidate issuer certificate is already part of the constructed
3928 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3932 * Improve forward-security support: add functions
3934 void SSL_CTX_set_not_resumable_session_callback(
3935 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3936 void SSL_set_not_resumable_session_callback(
3937 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3939 for use by SSL/TLS servers; the callback function will be called whenever a
3940 new session is created, and gets to decide whether the session may be
3941 cached to make it resumable (return 0) or not (return 1). (As by the
3942 SSL/TLS protocol specifications, the session_id sent by the server will be
3943 empty to indicate that the session is not resumable; also, the server will
3944 not generate RFC 4507 (RFC 5077) session tickets.)
3946 A simple reasonable callback implementation is to return is_forward_secure.
3947 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3948 by the SSL/TLS server library, indicating whether it can provide forward
3951 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3953 * New -verify_name option in command line utilities to set verification
3958 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3959 Add CMAC pkey methods.
3963 * Experimental renegotiation in s_server -www mode. If the client
3964 browses /reneg connection is renegotiated. If /renegcert it is
3965 renegotiated requesting a certificate.
3969 * Add an "external" session cache for debugging purposes to s_server. This
3970 should help trace issues which normally are only apparent in deployed
3971 multi-process servers.
3975 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3976 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3977 BIO_set_cipher() and some obscure PEM functions were changed so they
3978 can now return an error. The RAND changes required a change to the
3979 RAND_METHOD structure.
3983 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3984 a gcc attribute to warn if the result of a function is ignored. This
3985 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3986 whose return value is often ignored.
3990 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3991 These allow SCTs (signed certificate timestamps) to be requested and
3992 validated when establishing a connection.
3994 *Rob Percival <robpercival@google.com>*
3999 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4001 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4002 used even when parsing explicit parameters, when loading a encoded key
4003 or calling `EC_GROUP_new_from_ecpkparameters()`/
4004 `EC_GROUP_new_from_ecparameters()`.
4005 This prevents bypass of security hardening and performance gains,
4006 especially for curves with specialized EC_METHODs.
4007 By default, if a key encoded with explicit parameters is loaded and later
4008 encoded, the output is still encoded with explicit parameters, even if
4009 internally a "named" EC_GROUP is used for computation.
4013 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4014 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4015 NULL. After this change, only the cofactor parameter can be NULL. It also
4016 does some minimal sanity checks on the passed order.
4021 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4022 An attack is simple, if the first CMS_recipientInfo is valid but the
4023 second CMS_recipientInfo is chosen ciphertext. If the second
4024 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4025 encryption key will be replaced by garbage, and the message cannot be
4026 decoded, but if the RSA decryption fails, the correct encryption key is
4027 used and the recipient will not notice the attack.
4028 As a work around for this potential attack the length of the decrypted
4029 key must be equal to the cipher default key length, in case the
4030 certifiate is not given and all recipientInfo are tried out.
4031 The old behaviour can be re-enabled in the CMS code by setting the
4032 CMS_DEBUG_DECRYPT flag.
4037 * Document issue with installation paths in diverse Windows builds
4039 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4040 binaries and run-time config file.
4045 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4047 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4048 This changes the size when using the genpkey app when no size is given. It
4049 fixes an omission in earlier changes that changed all RSA, DSA and DH
4050 generation apps to use 2048 bits by default.
4054 * Add FIPS support for Android Arm 64-bit
4056 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4057 Module in Version 2.0.10. For some reason, the corresponding target
4058 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4059 built with FIPS support on Android Arm 64-bit. This omission has been
4062 *Matthias St. Pierre*
4064 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4066 * 0-byte record padding oracle
4068 If an application encounters a fatal protocol error and then calls
4069 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4070 then OpenSSL can respond differently to the calling application if a 0 byte
4071 record is received with invalid padding compared to if a 0 byte record is
4072 received with an invalid MAC. If the application then behaves differently
4073 based on that in a way that is detectable to the remote peer, then this
4074 amounts to a padding oracle that could be used to decrypt data.
4076 In order for this to be exploitable "non-stitched" ciphersuites must be in
4077 use. Stitched ciphersuites are optimised implementations of certain
4078 commonly used ciphersuites. Also the application must call SSL_shutdown()
4079 twice even if a protocol error has occurred (applications should not do
4080 this but some do anyway).
4082 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4083 Aviram, with additional investigation by Steven Collison and Andrew
4084 Hourselt. It was reported to OpenSSL on 10th December 2018.
4089 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4093 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4095 * Microarchitecture timing vulnerability in ECC scalar multiplication
4097 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4098 shown to be vulnerable to a microarchitecture timing side channel attack.
4099 An attacker with sufficient access to mount local timing attacks during
4100 ECDSA signature generation could recover the private key.
4102 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4103 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4109 * Timing vulnerability in DSA signature generation
4111 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4112 timing side channel attack. An attacker could use variations in the signing
4113 algorithm to recover the private key.
4115 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4120 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4121 Module, accidentally introduced while backporting security fixes from the
4122 development branch and hindering the use of ECC in FIPS mode.
4126 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4128 * Client DoS due to large DH parameter
4130 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4131 malicious server can send a very large prime value to the client. This will
4132 cause the client to spend an unreasonably long period of time generating a
4133 key for this prime resulting in a hang until the client has finished. This
4134 could be exploited in a Denial Of Service attack.
4136 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4141 * Cache timing vulnerability in RSA Key Generation
4143 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4144 a cache timing side channel attack. An attacker with sufficient access to
4145 mount cache timing attacks during the RSA key generation process could
4146 recover the private key.
4148 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4149 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4154 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4155 parameter is no longer accepted, as it leads to a corrupt table. NULL
4156 pem_str is reserved for alias entries only.
4160 * Revert blinding in ECDSA sign and instead make problematic addition
4161 length-invariant. Switch even to fixed-length Montgomery multiplication.
4165 * Change generating and checking of primes so that the error rate of not
4166 being prime depends on the intended use based on the size of the input.
4167 For larger primes this will result in more rounds of Miller-Rabin.
4168 The maximal error rate for primes with more than 1080 bits is lowered
4171 *Kurt Roeckx, Annie Yousar*
4173 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4177 * Add blinding to ECDSA and DSA signatures to protect against side channel
4178 attacks discovered by Keegan Ryan (NCC Group).
4182 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4183 now allow empty (zero character) pass phrases.
4187 * Certificate time validation (X509_cmp_time) enforces stricter
4188 compliance with RFC 5280. Fractional seconds and timezone offsets
4189 are no longer allowed.
4193 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4195 * Constructed ASN.1 types with a recursive definition could exceed the stack
4197 Constructed ASN.1 types with a recursive definition (such as can be found
4198 in PKCS7) could eventually exceed the stack given malicious input with
4199 excessive recursion. This could result in a Denial Of Service attack. There
4200 are no such structures used within SSL/TLS that come from untrusted sources
4201 so this is considered safe.
4203 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4209 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4211 * Read/write after SSL object in error state
4213 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4214 mechanism. The intent was that if a fatal error occurred during a handshake
4215 then OpenSSL would move into the error state and would immediately fail if
4216 you attempted to continue the handshake. This works as designed for the
4217 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4218 SSL_connect()), however due to a bug it does not work correctly if
4219 SSL_read() or SSL_write() is called directly. In that scenario, if the
4220 handshake fails then a fatal error will be returned in the initial function
4221 call. If SSL_read()/SSL_write() is subsequently called by the application
4222 for the same SSL object then it will succeed and the data is passed without
4223 being decrypted/encrypted directly from the SSL/TLS record layer.
4225 In order to exploit this issue an application bug would have to be present
4226 that resulted in a call to SSL_read()/SSL_write() being issued after having
4227 already received a fatal error.
4229 This issue was reported to OpenSSL by David Benjamin (Google).
4234 * rsaz_1024_mul_avx2 overflow bug on x86_64
4236 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4237 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4238 Analysis suggests that attacks against RSA and DSA as a result of this
4239 defect would be very difficult to perform and are not believed likely.
4240 Attacks against DH1024 are considered just feasible, because most of the
4241 work necessary to deduce information about a private key may be performed
4242 offline. The amount of resources required for such an attack would be
4243 significant. However, for an attack on TLS to be meaningful, the server
4244 would have to share the DH1024 private key among multiple clients, which is
4245 no longer an option since CVE-2016-0701.
4247 This only affects processors that support the AVX2 but not ADX extensions
4248 like Intel Haswell (4th generation).
4250 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4251 was originally found via the OSS-Fuzz project.
4256 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4258 * bn_sqrx8x_internal carry bug on x86_64
4260 There is a carry propagating bug in the x86_64 Montgomery squaring
4261 procedure. No EC algorithms are affected. Analysis suggests that attacks
4262 against RSA and DSA as a result of this defect would be very difficult to
4263 perform and are not believed likely. Attacks against DH are considered just
4264 feasible (although very difficult) because most of the work necessary to
4265 deduce information about a private key may be performed offline. The amount
4266 of resources required for such an attack would be very significant and
4267 likely only accessible to a limited number of attackers. An attacker would
4268 additionally need online access to an unpatched system using the target
4269 private key in a scenario with persistent DH parameters and a private
4270 key that is shared between multiple clients.
4272 This only affects processors that support the BMI1, BMI2 and ADX extensions
4273 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4275 This issue was reported to OpenSSL by the OSS-Fuzz project.
4280 * Malformed X.509 IPAddressFamily could cause OOB read
4282 If an X.509 certificate has a malformed IPAddressFamily extension,
4283 OpenSSL could do a one-byte buffer overread. The most likely result
4284 would be an erroneous display of the certificate in text format.
4286 This issue was reported to OpenSSL by the OSS-Fuzz project.
4291 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4293 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4294 platform rather than 'mingw'.
4298 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4300 * Truncated packet could crash via OOB read
4302 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4303 cipher is being used, then a truncated packet can cause that host to
4304 perform an out-of-bounds read, usually resulting in a crash.
4306 This issue was reported to OpenSSL by Robert Święcki of Google.
4311 * BN_mod_exp may produce incorrect results on x86_64
4313 There is a carry propagating bug in the x86_64 Montgomery squaring
4314 procedure. No EC algorithms are affected. Analysis suggests that attacks
4315 against RSA and DSA as a result of this defect would be very difficult to
4316 perform and are not believed likely. Attacks against DH are considered just
4317 feasible (although very difficult) because most of the work necessary to
4318 deduce information about a private key may be performed offline. The amount
4319 of resources required for such an attack would be very significant and
4320 likely only accessible to a limited number of attackers. An attacker would
4321 additionally need online access to an unpatched system using the target
4322 private key in a scenario with persistent DH parameters and a private
4323 key that is shared between multiple clients. For example this can occur by
4324 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4325 similar to CVE-2015-3193 but must be treated as a separate problem.
4327 This issue was reported to OpenSSL by the OSS-Fuzz project.
4332 * Montgomery multiplication may produce incorrect results
4334 There is a carry propagating bug in the Broadwell-specific Montgomery
4335 multiplication procedure that handles input lengths divisible by, but
4336 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4337 and DH private keys are impossible. This is because the subroutine in
4338 question is not used in operations with the private key itself and an input
4339 of the attacker's direct choice. Otherwise the bug can manifest itself as
4340 transient authentication and key negotiation failures or reproducible
4341 erroneous outcome of public-key operations with specially crafted input.
4342 Among EC algorithms only Brainpool P-512 curves are affected and one
4343 presumably can attack ECDH key negotiation. Impact was not analyzed in
4344 detail, because pre-requisites for attack are considered unlikely. Namely
4345 multiple clients have to choose the curve in question and the server has to
4346 share the private key among them, neither of which is default behaviour.
4347 Even then only clients that chose the curve will be affected.
4349 This issue was publicly reported as transient failures and was not
4350 initially recognized as a security issue. Thanks to Richard Morgan for
4351 providing reproducible case.
4356 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4357 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4358 prevent issues where no progress is being made and the peer continually
4359 sends unrecognised record types, using up resources processing them.
4363 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4365 * Missing CRL sanity check
4367 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4368 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4369 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4371 This issue only affects the OpenSSL 1.0.2i
4376 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4378 * OCSP Status Request extension unbounded memory growth
4380 A malicious client can send an excessively large OCSP Status Request
4381 extension. If that client continually requests renegotiation, sending a
4382 large OCSP Status Request extension each time, then there will be unbounded
4383 memory growth on the server. This will eventually lead to a Denial Of
4384 Service attack through memory exhaustion. Servers with a default
4385 configuration are vulnerable even if they do not support OCSP. Builds using
4386 the "no-ocsp" build time option are not affected.
4388 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4393 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4396 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4402 * OOB write in MDC2_Update()
4404 An overflow can occur in MDC2_Update() either if called directly or
4405 through the EVP_DigestUpdate() function using MDC2. If an attacker
4406 is able to supply very large amounts of input data after a previous
4407 call to EVP_EncryptUpdate() with a partial block then a length check
4408 can overflow resulting in a heap corruption.
4410 The amount of data needed is comparable to SIZE_MAX which is impractical
4413 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4418 * Malformed SHA512 ticket DoS
4420 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4421 DoS attack where a malformed ticket will result in an OOB read which will
4424 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4425 a custom server callback and ticket lookup mechanism.
4427 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4432 * OOB write in BN_bn2dec()
4434 The function BN_bn2dec() does not check the return value of BN_div_word().
4435 This can cause an OOB write if an application uses this function with an
4436 overly large BIGNUM. This could be a problem if an overly large certificate
4437 or CRL is printed out from an untrusted source. TLS is not affected because
4438 record limits will reject an oversized certificate before it is parsed.
4440 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4445 * OOB read in TS_OBJ_print_bio()
4447 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4448 the total length the OID text representation would use and not the amount
4449 of data written. This will result in OOB reads when large OIDs are
4452 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4457 * Pointer arithmetic undefined behaviour
4459 Avoid some undefined pointer arithmetic
4461 A common idiom in the codebase is to check limits in the following manner:
4464 Where "p" points to some malloc'd data of SIZE bytes and
4467 "len" here could be from some externally supplied data (e.g. from a TLS
4470 The rules of C pointer arithmetic are such that "p + len" is only well
4471 defined where len <= SIZE. Therefore the above idiom is actually
4472 undefined behaviour.
4474 For example this could cause problems if some malloc implementation
4475 provides an address for "p" such that "p + len" actually overflows for
4476 values of len that are too big and therefore p + len < limit.
4478 This issue was reported to OpenSSL by Guido Vranken
4483 * Constant time flag not preserved in DSA signing
4485 Operations in the DSA signing algorithm should run in constant time in
4486 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4487 implementation means that a non-constant time codepath is followed for
4488 certain operations. This has been demonstrated through a cache-timing
4489 attack to be sufficient for an attacker to recover the private DSA key.
4491 This issue was reported by César Pereida (Aalto University), Billy Brumley
4492 (Tampere University of Technology), and Yuval Yarom (The University of
4493 Adelaide and NICTA).
4498 * DTLS buffered message DoS
4500 In a DTLS connection where handshake messages are delivered out-of-order
4501 those messages that OpenSSL is not yet ready to process will be buffered
4502 for later use. Under certain circumstances, a flaw in the logic means that
4503 those messages do not get removed from the buffer even though the handshake
4504 has been completed. An attacker could force up to approx. 15 messages to
4505 remain in the buffer when they are no longer required. These messages will
4506 be cleared when the DTLS connection is closed. The default maximum size for
4507 a message is 100k. Therefore the attacker could force an additional 1500k
4508 to be consumed per connection. By opening many simulataneous connections an
4509 attacker could cause a DoS attack through memory exhaustion.
4511 This issue was reported to OpenSSL by Quan Luo.
4516 * DTLS replay protection DoS
4518 A flaw in the DTLS replay attack protection mechanism means that records
4519 that arrive for future epochs update the replay protection "window" before
4520 the MAC for the record has been validated. This could be exploited by an
4521 attacker by sending a record for the next epoch (which does not have to
4522 decrypt or have a valid MAC), with a very large sequence number. This means
4523 that all subsequent legitimate packets are dropped causing a denial of
4524 service for a specific DTLS connection.
4526 This issue was reported to OpenSSL by the OCAP audit team.
4531 * Certificate message OOB reads
4533 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4534 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4535 theoretical DoS risk but this has not been observed in practice on common
4538 The messages affected are client certificate, client certificate request
4539 and server certificate. As a result the attack can only be performed
4540 against a client or a server which enables client authentication.
4542 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4547 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4549 * Prevent padding oracle in AES-NI CBC MAC check
4551 A MITM attacker can use a padding oracle attack to decrypt traffic
4552 when the connection uses an AES CBC cipher and the server support
4555 This issue was introduced as part of the fix for Lucky 13 padding
4556 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4557 constant time by making sure that always the same bytes are read and
4558 compared against either the MAC or padding bytes. But it no longer
4559 checked that there was enough data to have both the MAC and padding
4562 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4567 * Fix EVP_EncodeUpdate overflow
4569 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4570 Base64 encoding of binary data. If an attacker is able to supply very large
4571 amounts of input data then a length check can overflow resulting in a heap
4574 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4575 the `PEM_write_bio*` family of functions. These are mainly used within the
4576 OpenSSL command line applications, so any application which processes data
4577 from an untrusted source and outputs it as a PEM file should be considered
4578 vulnerable to this issue. User applications that call these APIs directly
4579 with large amounts of untrusted data may also be vulnerable.
4581 This issue was reported by Guido Vranken.
4586 * Fix EVP_EncryptUpdate overflow
4588 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4589 is able to supply very large amounts of input data after a previous call to
4590 EVP_EncryptUpdate() with a partial block then a length check can overflow
4591 resulting in a heap corruption. Following an analysis of all OpenSSL
4592 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4593 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4594 the first called function after an EVP_EncryptInit(), and therefore that
4595 specific call must be safe. The second form is where the length passed to
4596 EVP_EncryptUpdate() can be seen from the code to be some small value and
4597 therefore there is no possibility of an overflow. Since all instances are
4598 one of these two forms, it is believed that there can be no overflows in
4599 internal code due to this problem. It should be noted that
4600 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4601 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4602 of these calls have also been analysed too and it is believed there are no
4603 instances in internal usage where an overflow could occur.
4605 This issue was reported by Guido Vranken.
4610 * Prevent ASN.1 BIO excessive memory allocation
4612 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4613 a short invalid encoding can cause allocation of large amounts of memory
4614 potentially consuming excessive resources or exhausting memory.
4616 Any application parsing untrusted data through d2i BIO functions is
4617 affected. The memory based functions such as d2i_X509() are *not* affected.
4618 Since the memory based functions are used by the TLS library, TLS
4619 applications are not affected.
4621 This issue was reported by Brian Carpenter.
4628 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4629 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4630 in arbitrary stack data being returned in the buffer.
4632 This issue was reported by Guido Vranken.
4637 * Modify behavior of ALPN to invoke callback after SNI/servername
4638 callback, such that updates to the SSL_CTX affect ALPN.
4642 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4647 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4648 methods are enabled and ssl2 is disabled the methods return NULL.
4652 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4654 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4655 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4656 provide any "EXPORT" or "LOW" strength ciphers.
4660 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4661 is by default disabled at build-time. Builds that are not configured with
4662 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4663 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4664 will need to explicitly call either of:
4666 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4668 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4670 as appropriate. Even if either of those is used, or the application
4671 explicitly uses the version-specific SSLv2_method() or its client and
4672 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4673 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4674 ciphers, and SSLv2 56-bit DES are no longer available.
4679 * Fix a double-free in DSA code
4681 A double free bug was discovered when OpenSSL parses malformed DSA private
4682 keys and could lead to a DoS attack or memory corruption for applications
4683 that receive DSA private keys from untrusted sources. This scenario is
4686 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4692 * Disable SRP fake user seed to address a server memory leak.
4694 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4696 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4697 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4698 was changed to ignore the "fake user" SRP seed, even if the seed
4701 Users should use SRP_VBASE_get1_by_user instead. Note that in
4702 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4703 also that even though configuring the SRP seed attempts to hide
4704 invalid usernames by continuing the handshake with fake
4705 credentials, this behaviour is not constant time and no strong
4706 guarantees are made that the handshake is indistinguishable from
4707 that of a valid user.
4712 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4714 In the BN_hex2bn function the number of hex digits is calculated using an
4715 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4716 large values of `i` this can result in `bn_expand` not allocating any
4717 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4718 field as NULL leading to a subsequent NULL ptr deref. For very large values
4719 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4720 In this case memory is allocated to the internal BIGNUM data field, but it
4721 is insufficiently sized leading to heap corruption. A similar issue exists
4722 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4723 is ever called by user applications with very large untrusted hex/dec data.
4724 This is anticipated to be a rare occurrence.
4726 All OpenSSL internal usage of these functions use data that is not expected
4727 to be untrusted, e.g. config file data or application command line
4728 arguments. If user developed applications generate config file data based
4729 on untrusted data then it is possible that this could also lead to security
4730 consequences. This is also anticipated to be rare.
4732 This issue was reported to OpenSSL by Guido Vranken.
4737 * Fix memory issues in `BIO_*printf` functions
4739 The internal `fmtstr` function used in processing a "%s" format string in
4740 the `BIO_*printf` functions could overflow while calculating the length of a
4741 string and cause an OOB read when printing very long strings.
4743 Additionally the internal `doapr_outch` function can attempt to write to an
4744 OOB memory location (at an offset from the NULL pointer) in the event of a
4745 memory allocation failure. In 1.0.2 and below this could be caused where
4746 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4747 could be in processing a very long "%s" format string. Memory leaks can
4750 The first issue may mask the second issue dependent on compiler behaviour.
4751 These problems could enable attacks where large amounts of untrusted data
4752 is passed to the `BIO_*printf` functions. If applications use these functions
4753 in this way then they could be vulnerable. OpenSSL itself uses these
4754 functions when printing out human-readable dumps of ASN.1 data. Therefore
4755 applications that print this data could be vulnerable if the data is from
4756 untrusted sources. OpenSSL command line applications could also be
4757 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4758 as command line arguments.
4760 Libssl is not considered directly vulnerable. Additionally certificates etc
4761 received via remote connections via libssl are also unlikely to be able to
4762 trigger these issues because of message size limits enforced within libssl.
4764 This issue was reported to OpenSSL Guido Vranken.
4769 * Side channel attack on modular exponentiation
4771 A side-channel attack was found which makes use of cache-bank conflicts on
4772 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4773 of RSA keys. The ability to exploit this issue is limited as it relies on
4774 an attacker who has control of code in a thread running on the same
4775 hyper-threaded core as the victim thread which is performing decryptions.
4777 This issue was reported to OpenSSL by Yuval Yarom, The University of
4778 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4779 Nadia Heninger, University of Pennsylvania with more information at
4780 <http://cachebleed.info>.
4785 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4786 if no keysize is specified with default_bits. This fixes an
4787 omission in an earlier change that changed all RSA/DSA key generation
4788 apps to use 2048 bits by default.
4792 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4794 * DH small subgroups
4796 Historically OpenSSL only ever generated DH parameters based on "safe"
4797 primes. More recently (in version 1.0.2) support was provided for
4798 generating X9.42 style parameter files such as those required for RFC 5114
4799 support. The primes used in such files may not be "safe". Where an
4800 application is using DH configured with parameters based on primes that are
4801 not "safe" then an attacker could use this fact to find a peer's private
4802 DH exponent. This attack requires that the attacker complete multiple
4803 handshakes in which the peer uses the same private DH exponent. For example
4804 this could be used to discover a TLS server's private DH exponent if it's
4805 reusing the private DH exponent or it's using a static DH ciphersuite.
4807 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4808 TLS. It is not on by default. If the option is not set then the server
4809 reuses the same private DH exponent for the life of the server process and
4810 would be vulnerable to this attack. It is believed that many popular
4811 applications do set this option and would therefore not be at risk.
4813 The fix for this issue adds an additional check where a "q" parameter is
4814 available (as is the case in X9.42 based parameters). This detects the
4815 only known attack, and is the only possible defense for static DH
4816 ciphersuites. This could have some performance impact.
4818 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4819 default and cannot be disabled. This could have some performance impact.
4821 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4826 * SSLv2 doesn't block disabled ciphers
4828 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4829 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4830 been disabled, provided that the SSLv2 protocol was not also disabled via
4833 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4834 and Sebastian Schinzel.
4839 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4841 * BN_mod_exp may produce incorrect results on x86_64
4843 There is a carry propagating bug in the x86_64 Montgomery squaring
4844 procedure. No EC algorithms are affected. Analysis suggests that attacks
4845 against RSA and DSA as a result of this defect would be very difficult to
4846 perform and are not believed likely. Attacks against DH are considered just
4847 feasible (although very difficult) because most of the work necessary to
4848 deduce information about a private key may be performed offline. The amount
4849 of resources required for such an attack would be very significant and
4850 likely only accessible to a limited number of attackers. An attacker would
4851 additionally need online access to an unpatched system using the target
4852 private key in a scenario with persistent DH parameters and a private
4853 key that is shared between multiple clients. For example this can occur by
4854 default in OpenSSL DHE based SSL/TLS ciphersuites.
4856 This issue was reported to OpenSSL by Hanno Böck.
4861 * Certificate verify crash with missing PSS parameter
4863 The signature verification routines will crash with a NULL pointer
4864 dereference if presented with an ASN.1 signature using the RSA PSS
4865 algorithm and absent mask generation function parameter. Since these
4866 routines are used to verify certificate signature algorithms this can be
4867 used to crash any certificate verification operation and exploited in a
4868 DoS attack. Any application which performs certificate verification is
4869 vulnerable including OpenSSL clients and servers which enable client
4872 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4877 * X509_ATTRIBUTE memory leak
4879 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4880 memory. This structure is used by the PKCS#7 and CMS routines so any
4881 application which reads PKCS#7 or CMS data from untrusted sources is
4882 affected. SSL/TLS is not affected.
4884 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4890 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4891 This changes the decoding behaviour for some invalid messages,
4892 though the change is mostly in the more lenient direction, and
4893 legacy behaviour is preserved as much as possible.
4897 * In DSA_generate_parameters_ex, if the provided seed is too short,
4900 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4902 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4904 * Alternate chains certificate forgery
4906 During certificate verification, OpenSSL will attempt to find an
4907 alternative certificate chain if the first attempt to build such a chain
4908 fails. An error in the implementation of this logic can mean that an
4909 attacker could cause certain checks on untrusted certificates to be
4910 bypassed, such as the CA flag, enabling them to use a valid leaf
4911 certificate to act as a CA and "issue" an invalid certificate.
4913 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4918 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4920 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4921 incompatibility in the handling of HMAC. The previous ABI has now been
4926 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4928 * Malformed ECParameters causes infinite loop
4930 When processing an ECParameters structure OpenSSL enters an infinite loop
4931 if the curve specified is over a specially malformed binary polynomial
4934 This can be used to perform denial of service against any
4935 system which processes public keys, certificate requests or
4936 certificates. This includes TLS clients and TLS servers with
4937 client authentication enabled.
4939 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4944 * Exploitable out-of-bounds read in X509_cmp_time
4946 X509_cmp_time does not properly check the length of the ASN1_TIME
4947 string and can read a few bytes out of bounds. In addition,
4948 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4951 An attacker can use this to craft malformed certificates and CRLs of
4952 various sizes and potentially cause a segmentation fault, resulting in
4953 a DoS on applications that verify certificates or CRLs. TLS clients
4954 that verify CRLs are affected. TLS clients and servers with client
4955 authentication enabled may be affected if they use custom verification
4958 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4959 independently by Hanno Böck.
4964 * PKCS7 crash with missing EnvelopedContent
4966 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4967 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4968 with missing content and trigger a NULL pointer dereference on parsing.
4970 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4971 structures from untrusted sources are affected. OpenSSL clients and
4972 servers are not affected.
4974 This issue was reported to OpenSSL by Michal Zalewski (Google).
4979 * CMS verify infinite loop with unknown hash function
4981 When verifying a signedData message the CMS code can enter an infinite loop
4982 if presented with an unknown hash function OID. This can be used to perform
4983 denial of service against any system which verifies signedData messages using
4985 This issue was reported to OpenSSL by Johannes Bauer.
4990 * Race condition handling NewSessionTicket
4992 If a NewSessionTicket is received by a multi-threaded client when attempting to
4993 reuse a previous ticket then a race condition can occur potentially leading to
4994 a double free of the ticket data.
4999 * Only support 256-bit or stronger elliptic curves with the
5000 'ecdh_auto' setting (server) or by default (client). Of supported
5001 curves, prefer P-256 (both).
5005 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5007 * ClientHello sigalgs DoS fix
5009 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5010 invalid signature algorithms extension a NULL pointer dereference will
5011 occur. This can be exploited in a DoS attack against the server.
5013 This issue was was reported to OpenSSL by David Ramos of Stanford
5017 *Stephen Henson and Matt Caswell*
5019 * Multiblock corrupted pointer fix
5021 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5022 feature only applies on 64 bit x86 architecture platforms that support AES
5023 NI instructions. A defect in the implementation of "multiblock" can cause
5024 OpenSSL's internal write buffer to become incorrectly set to NULL when
5025 using non-blocking IO. Typically, when the user application is using a
5026 socket BIO for writing, this will only result in a failed connection.
5027 However if some other BIO is used then it is likely that a segmentation
5028 fault will be triggered, thus enabling a potential DoS attack.
5030 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5035 * Segmentation fault in DTLSv1_listen fix
5037 The DTLSv1_listen function is intended to be stateless and processes the
5038 initial ClientHello from many peers. It is common for user code to loop
5039 over the call to DTLSv1_listen until a valid ClientHello is received with
5040 an associated cookie. A defect in the implementation of DTLSv1_listen means
5041 that state is preserved in the SSL object from one invocation to the next
5042 that can lead to a segmentation fault. Errors processing the initial
5043 ClientHello can trigger this scenario. An example of such an error could be
5044 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5047 This issue was reported to OpenSSL by Per Allansson.
5052 * Segmentation fault in ASN1_TYPE_cmp fix
5054 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5055 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5056 certificate signature algorithm consistency this can be used to crash any
5057 certificate verification operation and exploited in a DoS attack. Any
5058 application which performs certificate verification is vulnerable including
5059 OpenSSL clients and servers which enable client authentication.
5064 * Segmentation fault for invalid PSS parameters fix
5066 The signature verification routines will crash with a NULL pointer
5067 dereference if presented with an ASN.1 signature using the RSA PSS
5068 algorithm and invalid parameters. Since these routines are used to verify
5069 certificate signature algorithms this can be used to crash any
5070 certificate verification operation and exploited in a DoS attack. Any
5071 application which performs certificate verification is vulnerable including
5072 OpenSSL clients and servers which enable client authentication.
5074 This issue was was reported to OpenSSL by Brian Carpenter.
5079 * ASN.1 structure reuse memory corruption fix
5081 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5082 memory corruption via an invalid write. Such reuse is and has been
5083 strongly discouraged and is believed to be rare.
5085 Applications that parse structures containing CHOICE or ANY DEFINED BY
5086 components may be affected. Certificate parsing (d2i_X509 and related
5087 functions) are however not affected. OpenSSL clients and servers are
5093 * PKCS7 NULL pointer dereferences fix
5095 The PKCS#7 parsing code does not handle missing outer ContentInfo
5096 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5097 missing content and trigger a NULL pointer dereference on parsing.
5099 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5100 otherwise parse PKCS#7 structures from untrusted sources are
5101 affected. OpenSSL clients and servers are not affected.
5103 This issue was reported to OpenSSL by Michal Zalewski (Google).
5108 * DoS via reachable assert in SSLv2 servers fix
5110 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5111 servers that both support SSLv2 and enable export cipher suites by sending
5112 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5114 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5115 (OpenSSL development team).
5120 * Empty CKE with client auth and DHE fix
5122 If client auth is used then a server can seg fault in the event of a DHE
5123 ciphersuite being selected and a zero length ClientKeyExchange message
5124 being sent by the client. This could be exploited in a DoS attack.
5129 * Handshake with unseeded PRNG fix
5131 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5132 with an unseeded PRNG. The conditions are:
5133 - The client is on a platform where the PRNG has not been seeded
5134 automatically, and the user has not seeded manually
5135 - A protocol specific client method version has been used (i.e. not
5136 SSL_client_methodv23)
5137 - A ciphersuite is used that does not require additional random data from
5138 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5140 If the handshake succeeds then the client random that has been used will
5141 have been generated from a PRNG with insufficient entropy and therefore the
5142 output may be predictable.
5144 For example using the following command with an unseeded openssl will
5145 succeed on an unpatched platform:
5147 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5152 * Use After Free following d2i_ECPrivatekey error fix
5154 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5155 could cause a use after free condition. This, in turn, could cause a double
5156 free in several private key parsing functions (such as d2i_PrivateKey
5157 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5158 for applications that receive EC private keys from untrusted
5159 sources. This scenario is considered rare.
5161 This issue was discovered by the BoringSSL project and fixed in their
5167 * X509_to_X509_REQ NULL pointer deref fix
5169 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5170 the certificate key is invalid. This function is rarely used in practice.
5172 This issue was discovered by Brian Carpenter.
5177 * Removed the export ciphers from the DEFAULT ciphers
5181 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5183 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5184 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5185 So far those who have to target multiple platforms would compromise
5186 and argue that binary targeting say ARMv5 would still execute on
5187 ARMv8. "Universal" build resolves this compromise by providing
5188 near-optimal performance even on newer platforms.
5192 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5193 (other platforms pending).
5195 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5197 * Add support for the SignedCertificateTimestampList certificate and
5198 OCSP response extensions from RFC6962.
5202 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5203 for corner cases. (Certain input points at infinity could lead to
5204 bogus results, with non-infinity inputs mapped to infinity too.)
5208 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5209 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5210 common cases are optimized and there still is room for further
5211 improvements. Vector Permutation AES for Altivec is also added.
5215 * Add support for little-endian ppc64 Linux target.
5217 *Marcelo Cerri (IBM)*
5219 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5220 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5221 are optimized and there still is room for further improvements.
5222 Both 32- and 64-bit modes are supported.
5224 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5226 * Improved ARMv7 NEON support.
5230 * Support for SPARC Architecture 2011 crypto extensions, first
5231 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5232 SHA256/512, MD5, GHASH and modular exponentiation.
5234 *Andy Polyakov, David Miller*
5236 * Accelerated modular exponentiation for Intel processors, a.k.a.
5239 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5241 * Support for new and upcoming Intel processors, including AVX2,
5242 BMI and SHA ISA extensions. This includes additional "stitched"
5243 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5246 This work was sponsored by Intel Corp.
5250 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5251 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5252 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5256 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5257 this fixes a limitation in previous versions of OpenSSL.
5261 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5262 MGF1 digest and OAEP label.
5266 * Add EVP support for key wrapping algorithms, to avoid problems with
5267 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5268 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5269 algorithms and include tests cases.
5273 * Add functions to allocate and set the fields of an ECDSA_METHOD
5276 *Douglas E. Engert, Steve Henson*
5278 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5279 difference in days and seconds between two tm or ASN1_TIME structures.
5283 * Add -rev test option to s_server to just reverse order of characters
5284 received by client and send back to server. Also prints an abbreviated
5285 summary of the connection parameters.
5289 * New option -brief for s_client and s_server to print out a brief summary
5290 of connection parameters.
5294 * Add callbacks for arbitrary TLS extensions.
5296 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5298 * New option -crl_download in several openssl utilities to download CRLs
5299 from CRLDP extension in certificates.
5303 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5307 * New function X509_CRL_diff to generate a delta CRL from the difference
5308 of two full CRLs. Add support to "crl" utility.
5312 * New functions to set lookup_crls function and to retrieve
5313 X509_STORE from X509_STORE_CTX.
5317 * Print out deprecated issuer and subject unique ID fields in
5322 * Extend OCSP I/O functions so they can be used for simple general purpose
5323 HTTP as well as OCSP. New wrapper function which can be used to download
5324 CRLs using the OCSP API.
5328 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5332 * `SSL_CONF*` functions. These provide a common framework for application
5333 configuration using configuration files or command lines.
5337 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5338 message callback and prints the results. Needs compile time option
5339 "enable-ssl-trace". New options to s_client and s_server to enable
5344 * New ctrl and macro to retrieve supported points extensions.
5345 Print out extension in s_server and s_client.
5349 * New functions to retrieve certificate signature and signature
5354 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5359 * New Suite B modes for TLS code. These use and enforce the requirements
5360 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5361 only use Suite B curves. The Suite B modes can be set by using the
5362 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5366 * New chain verification flags for Suite B levels of security. Check
5367 algorithms are acceptable when flags are set in X509_verify_cert.
5371 * Make tls1_check_chain return a set of flags indicating checks passed
5372 by a certificate chain. Add additional tests to handle client
5373 certificates: checks for matching certificate type and issuer name
5378 * If an attempt is made to use a signature algorithm not in the peer
5379 preference list abort the handshake. If client has no suitable
5380 signature algorithms in response to a certificate request do not
5381 use the certificate.
5385 * If server EC tmp key is not in client preference list abort handshake.
5389 * Add support for certificate stores in CERT structure. This makes it
5390 possible to have different stores per SSL structure or one store in
5391 the parent SSL_CTX. Include distinct stores for certificate chain
5392 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5393 to build and store a certificate chain in CERT structure: returning
5394 an error if the chain cannot be built: this will allow applications
5395 to test if a chain is correctly configured.
5397 Note: if the CERT based stores are not set then the parent SSL_CTX
5398 store is used to retain compatibility with existing behaviour.
5402 * New function ssl_set_client_disabled to set a ciphersuite disabled
5403 mask based on the current session, check mask when sending client
5404 hello and checking the requested ciphersuite.
5408 * New ctrls to retrieve and set certificate types in a certificate
5409 request message. Print out received values in s_client. If certificate
5410 types is not set with custom values set sensible values based on
5411 supported signature algorithms.
5415 * Support for distinct client and server supported signature algorithms.
5419 * Add certificate callback. If set this is called whenever a certificate
5420 is required by client or server. An application can decide which
5421 certificate chain to present based on arbitrary criteria: for example
5422 supported signature algorithms. Add very simple example to s_server.
5423 This fixes many of the problems and restrictions of the existing client
5424 certificate callback: for example you can now clear an existing
5425 certificate and specify the whole chain.
5429 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5430 the certificate can be used for (if anything). Set valid_flags field
5431 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5432 to have similar checks in it.
5434 Add new "cert_flags" field to CERT structure and include a "strict mode".
5435 This enforces some TLS certificate requirements (such as only permitting
5436 certificate signature algorithms contained in the supported algorithms
5437 extension) which some implementations ignore: this option should be used
5438 with caution as it could cause interoperability issues.
5442 * Update and tidy signature algorithm extension processing. Work out
5443 shared signature algorithms based on preferences and peer algorithms
5444 and print them out in s_client and s_server. Abort handshake if no
5445 shared signature algorithms.
5449 * Add new functions to allow customised supported signature algorithms
5450 for SSL and SSL_CTX structures. Add options to s_client and s_server
5455 * New function SSL_certs_clear() to delete all references to certificates
5456 from an SSL structure. Before this once a certificate had been added
5457 it couldn't be removed.
5461 * Integrate hostname, email address and IP address checking with certificate
5462 verification. New verify options supporting checking in openssl utility.
5466 * Fixes and wildcard matching support to hostname and email checking
5467 functions. Add manual page.
5469 *Florian Weimer (Red Hat Product Security Team)*
5471 * New functions to check a hostname email or IP address against a
5472 certificate. Add options x509 utility to print results of checks against
5477 * Fix OCSP checking.
5479 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5481 * Initial experimental support for explicitly trusted non-root CAs.
5482 OpenSSL still tries to build a complete chain to a root but if an
5483 intermediate CA has a trust setting included that is used. The first
5484 setting is used: whether to trust (e.g., -addtrust option to the x509
5489 * Add -trusted_first option which attempts to find certificates in the
5490 trusted store even if an untrusted chain is also supplied.
5494 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5495 platform support for Linux and Android.
5499 * Support for linux-x32, ILP32 environment in x86_64 framework.
5503 * Experimental multi-implementation support for FIPS capable OpenSSL.
5504 When in FIPS mode the approved implementations are used as normal,
5505 when not in FIPS mode the internal unapproved versions are used instead.
5506 This means that the FIPS capable OpenSSL isn't forced to use the
5507 (often lower performance) FIPS implementations outside FIPS mode.
5511 * Transparently support X9.42 DH parameters when calling
5512 PEM_read_bio_DHparameters. This means existing applications can handle
5513 the new parameter format automatically.
5517 * Initial experimental support for X9.42 DH parameter format: mainly
5518 to support use of 'q' parameter for RFC5114 parameters.
5522 * Add DH parameters from RFC5114 including test data to dhtest.
5526 * Support for automatic EC temporary key parameter selection. If enabled
5527 the most preferred EC parameters are automatically used instead of
5528 hardcoded fixed parameters. Now a server just has to call:
5529 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5530 support ECDH and use the most appropriate parameters.
5534 * Enhance and tidy EC curve and point format TLS extension code. Use
5535 static structures instead of allocation if default values are used.
5536 New ctrls to set curves we wish to support and to retrieve shared curves.
5537 Print out shared curves in s_server. New options to s_server and s_client
5538 to set list of supported curves.
5542 * New ctrls to retrieve supported signature algorithms and
5543 supported curve values as an array of NIDs. Extend openssl utility
5544 to print out received values.
5548 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5549 between NIDs and the more common NIST names such as "P-256". Enhance
5550 ecparam utility and ECC method to recognise the NIST names for curves.
5554 * Enhance SSL/TLS certificate chain handling to support different
5555 chains for each certificate instead of one chain in the parent SSL_CTX.
5559 * Support for fixed DH ciphersuite client authentication: where both
5560 server and client use DH certificates with common parameters.
5564 * Support for fixed DH ciphersuites: those requiring DH server
5569 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5571 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5572 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5573 X509_CINF_get_signature were reverted post internal team review.
5578 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5580 * OCSP Status Request extension unbounded memory growth
5582 A malicious client can send an excessively large OCSP Status Request
5583 extension. If that client continually requests renegotiation, sending a
5584 large OCSP Status Request extension each time, then there will be unbounded
5585 memory growth on the server. This will eventually lead to a Denial Of
5586 Service attack through memory exhaustion. Servers with a default
5587 configuration are vulnerable even if they do not support OCSP. Builds using
5588 the "no-ocsp" build time option are not affected.
5590 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5595 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5598 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5604 * OOB write in MDC2_Update()
5606 An overflow can occur in MDC2_Update() either if called directly or
5607 through the EVP_DigestUpdate() function using MDC2. If an attacker
5608 is able to supply very large amounts of input data after a previous
5609 call to EVP_EncryptUpdate() with a partial block then a length check
5610 can overflow resulting in a heap corruption.
5612 The amount of data needed is comparable to SIZE_MAX which is impractical
5615 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5620 * Malformed SHA512 ticket DoS
5622 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5623 DoS attack where a malformed ticket will result in an OOB read which will
5626 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5627 a custom server callback and ticket lookup mechanism.
5629 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5634 * OOB write in BN_bn2dec()
5636 The function BN_bn2dec() does not check the return value of BN_div_word().
5637 This can cause an OOB write if an application uses this function with an
5638 overly large BIGNUM. This could be a problem if an overly large certificate
5639 or CRL is printed out from an untrusted source. TLS is not affected because
5640 record limits will reject an oversized certificate before it is parsed.
5642 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5647 * OOB read in TS_OBJ_print_bio()
5649 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5650 the total length the OID text representation would use and not the amount
5651 of data written. This will result in OOB reads when large OIDs are
5654 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5659 * Pointer arithmetic undefined behaviour
5661 Avoid some undefined pointer arithmetic
5663 A common idiom in the codebase is to check limits in the following manner:
5666 Where "p" points to some malloc'd data of SIZE bytes and
5669 "len" here could be from some externally supplied data (e.g. from a TLS
5672 The rules of C pointer arithmetic are such that "p + len" is only well
5673 defined where len <= SIZE. Therefore the above idiom is actually
5674 undefined behaviour.
5676 For example this could cause problems if some malloc implementation
5677 provides an address for "p" such that "p + len" actually overflows for
5678 values of len that are too big and therefore p + len < limit.
5680 This issue was reported to OpenSSL by Guido Vranken
5685 * Constant time flag not preserved in DSA signing
5687 Operations in the DSA signing algorithm should run in constant time in
5688 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5689 implementation means that a non-constant time codepath is followed for
5690 certain operations. This has been demonstrated through a cache-timing
5691 attack to be sufficient for an attacker to recover the private DSA key.
5693 This issue was reported by César Pereida (Aalto University), Billy Brumley
5694 (Tampere University of Technology), and Yuval Yarom (The University of
5695 Adelaide and NICTA).
5700 * DTLS buffered message DoS
5702 In a DTLS connection where handshake messages are delivered out-of-order
5703 those messages that OpenSSL is not yet ready to process will be buffered
5704 for later use. Under certain circumstances, a flaw in the logic means that
5705 those messages do not get removed from the buffer even though the handshake
5706 has been completed. An attacker could force up to approx. 15 messages to
5707 remain in the buffer when they are no longer required. These messages will
5708 be cleared when the DTLS connection is closed. The default maximum size for
5709 a message is 100k. Therefore the attacker could force an additional 1500k
5710 to be consumed per connection. By opening many simulataneous connections an
5711 attacker could cause a DoS attack through memory exhaustion.
5713 This issue was reported to OpenSSL by Quan Luo.
5718 * DTLS replay protection DoS
5720 A flaw in the DTLS replay attack protection mechanism means that records
5721 that arrive for future epochs update the replay protection "window" before
5722 the MAC for the record has been validated. This could be exploited by an
5723 attacker by sending a record for the next epoch (which does not have to
5724 decrypt or have a valid MAC), with a very large sequence number. This means
5725 that all subsequent legitimate packets are dropped causing a denial of
5726 service for a specific DTLS connection.
5728 This issue was reported to OpenSSL by the OCAP audit team.
5733 * Certificate message OOB reads
5735 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5736 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5737 theoretical DoS risk but this has not been observed in practice on common
5740 The messages affected are client certificate, client certificate request
5741 and server certificate. As a result the attack can only be performed
5742 against a client or a server which enables client authentication.
5744 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5749 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5751 * Prevent padding oracle in AES-NI CBC MAC check
5753 A MITM attacker can use a padding oracle attack to decrypt traffic
5754 when the connection uses an AES CBC cipher and the server support
5757 This issue was introduced as part of the fix for Lucky 13 padding
5758 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5759 constant time by making sure that always the same bytes are read and
5760 compared against either the MAC or padding bytes. But it no longer
5761 checked that there was enough data to have both the MAC and padding
5764 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5769 * Fix EVP_EncodeUpdate overflow
5771 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5772 Base64 encoding of binary data. If an attacker is able to supply very large
5773 amounts of input data then a length check can overflow resulting in a heap
5776 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5777 the `PEM_write_bio*` family of functions. These are mainly used within the
5778 OpenSSL command line applications, so any application which processes data
5779 from an untrusted source and outputs it as a PEM file should be considered
5780 vulnerable to this issue. User applications that call these APIs directly
5781 with large amounts of untrusted data may also be vulnerable.
5783 This issue was reported by Guido Vranken.
5788 * Fix EVP_EncryptUpdate overflow
5790 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5791 is able to supply very large amounts of input data after a previous call to
5792 EVP_EncryptUpdate() with a partial block then a length check can overflow
5793 resulting in a heap corruption. Following an analysis of all OpenSSL
5794 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5795 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5796 the first called function after an EVP_EncryptInit(), and therefore that
5797 specific call must be safe. The second form is where the length passed to
5798 EVP_EncryptUpdate() can be seen from the code to be some small value and
5799 therefore there is no possibility of an overflow. Since all instances are
5800 one of these two forms, it is believed that there can be no overflows in
5801 internal code due to this problem. It should be noted that
5802 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5803 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5804 of these calls have also been analysed too and it is believed there are no
5805 instances in internal usage where an overflow could occur.
5807 This issue was reported by Guido Vranken.
5812 * Prevent ASN.1 BIO excessive memory allocation
5814 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5815 a short invalid encoding can casuse allocation of large amounts of memory
5816 potentially consuming excessive resources or exhausting memory.
5818 Any application parsing untrusted data through d2i BIO functions is
5819 affected. The memory based functions such as d2i_X509() are *not* affected.
5820 Since the memory based functions are used by the TLS library, TLS
5821 applications are not affected.
5823 This issue was reported by Brian Carpenter.
5830 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5831 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5832 in arbitrary stack data being returned in the buffer.
5834 This issue was reported by Guido Vranken.
5839 * Modify behavior of ALPN to invoke callback after SNI/servername
5840 callback, such that updates to the SSL_CTX affect ALPN.
5844 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5849 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5850 methods are enabled and ssl2 is disabled the methods return NULL.
5854 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5856 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5857 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5858 provide any "EXPORT" or "LOW" strength ciphers.
5862 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5863 is by default disabled at build-time. Builds that are not configured with
5864 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5865 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5866 will need to explicitly call either of:
5868 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5870 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5872 as appropriate. Even if either of those is used, or the application
5873 explicitly uses the version-specific SSLv2_method() or its client and
5874 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5875 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5876 ciphers, and SSLv2 56-bit DES are no longer available.
5881 * Fix a double-free in DSA code
5883 A double free bug was discovered when OpenSSL parses malformed DSA private
5884 keys and could lead to a DoS attack or memory corruption for applications
5885 that receive DSA private keys from untrusted sources. This scenario is
5888 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5894 * Disable SRP fake user seed to address a server memory leak.
5896 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5898 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5899 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5900 was changed to ignore the "fake user" SRP seed, even if the seed
5903 Users should use SRP_VBASE_get1_by_user instead. Note that in
5904 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5905 also that even though configuring the SRP seed attempts to hide
5906 invalid usernames by continuing the handshake with fake
5907 credentials, this behaviour is not constant time and no strong
5908 guarantees are made that the handshake is indistinguishable from
5909 that of a valid user.
5914 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5916 In the BN_hex2bn function the number of hex digits is calculated using an
5917 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5918 large values of `i` this can result in `bn_expand` not allocating any
5919 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5920 field as NULL leading to a subsequent NULL ptr deref. For very large values
5921 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5922 In this case memory is allocated to the internal BIGNUM data field, but it
5923 is insufficiently sized leading to heap corruption. A similar issue exists
5924 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5925 is ever called by user applications with very large untrusted hex/dec data.
5926 This is anticipated to be a rare occurrence.
5928 All OpenSSL internal usage of these functions use data that is not expected
5929 to be untrusted, e.g. config file data or application command line
5930 arguments. If user developed applications generate config file data based
5931 on untrusted data then it is possible that this could also lead to security
5932 consequences. This is also anticipated to be rare.
5934 This issue was reported to OpenSSL by Guido Vranken.
5939 * Fix memory issues in `BIO_*printf` functions
5941 The internal `fmtstr` function used in processing a "%s" format string in
5942 the `BIO_*printf` functions could overflow while calculating the length of a
5943 string and cause an OOB read when printing very long strings.
5945 Additionally the internal `doapr_outch` function can attempt to write to an
5946 OOB memory location (at an offset from the NULL pointer) in the event of a
5947 memory allocation failure. In 1.0.2 and below this could be caused where
5948 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5949 could be in processing a very long "%s" format string. Memory leaks can
5952 The first issue may mask the second issue dependent on compiler behaviour.
5953 These problems could enable attacks where large amounts of untrusted data
5954 is passed to the `BIO_*printf` functions. If applications use these functions
5955 in this way then they could be vulnerable. OpenSSL itself uses these
5956 functions when printing out human-readable dumps of ASN.1 data. Therefore
5957 applications that print this data could be vulnerable if the data is from
5958 untrusted sources. OpenSSL command line applications could also be
5959 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5960 as command line arguments.
5962 Libssl is not considered directly vulnerable. Additionally certificates etc
5963 received via remote connections via libssl are also unlikely to be able to
5964 trigger these issues because of message size limits enforced within libssl.
5966 This issue was reported to OpenSSL Guido Vranken.
5971 * Side channel attack on modular exponentiation
5973 A side-channel attack was found which makes use of cache-bank conflicts on
5974 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5975 of RSA keys. The ability to exploit this issue is limited as it relies on
5976 an attacker who has control of code in a thread running on the same
5977 hyper-threaded core as the victim thread which is performing decryptions.
5979 This issue was reported to OpenSSL by Yuval Yarom, The University of
5980 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5981 Nadia Heninger, University of Pennsylvania with more information at
5982 <http://cachebleed.info>.
5987 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5988 if no keysize is specified with default_bits. This fixes an
5989 omission in an earlier change that changed all RSA/DSA key generation
5990 apps to use 2048 bits by default.
5994 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5996 * Protection for DH small subgroup attacks
5998 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5999 switched on by default and cannot be disabled. This could have some
6004 * SSLv2 doesn't block disabled ciphers
6006 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6007 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6008 been disabled, provided that the SSLv2 protocol was not also disabled via
6011 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6012 and Sebastian Schinzel.
6017 * Reject DH handshakes with parameters shorter than 1024 bits.
6021 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6023 * Certificate verify crash with missing PSS parameter
6025 The signature verification routines will crash with a NULL pointer
6026 dereference if presented with an ASN.1 signature using the RSA PSS
6027 algorithm and absent mask generation function parameter. Since these
6028 routines are used to verify certificate signature algorithms this can be
6029 used to crash any certificate verification operation and exploited in a
6030 DoS attack. Any application which performs certificate verification is
6031 vulnerable including OpenSSL clients and servers which enable client
6034 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6039 * X509_ATTRIBUTE memory leak
6041 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6042 memory. This structure is used by the PKCS#7 and CMS routines so any
6043 application which reads PKCS#7 or CMS data from untrusted sources is
6044 affected. SSL/TLS is not affected.
6046 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6052 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6053 This changes the decoding behaviour for some invalid messages,
6054 though the change is mostly in the more lenient direction, and
6055 legacy behaviour is preserved as much as possible.
6059 * In DSA_generate_parameters_ex, if the provided seed is too short,
6060 use a random seed, as already documented.
6062 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6064 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6066 * Alternate chains certificate forgery
6068 During certificate verfification, OpenSSL will attempt to find an
6069 alternative certificate chain if the first attempt to build such a chain
6070 fails. An error in the implementation of this logic can mean that an
6071 attacker could cause certain checks on untrusted certificates to be
6072 bypassed, such as the CA flag, enabling them to use a valid leaf
6073 certificate to act as a CA and "issue" an invalid certificate.
6075 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6081 * Race condition handling PSK identify hint
6083 If PSK identity hints are received by a multi-threaded client then
6084 the values are wrongly updated in the parent SSL_CTX structure. This can
6085 result in a race condition potentially leading to a double free of the
6091 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6093 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6094 incompatibility in the handling of HMAC. The previous ABI has now been
6097 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6099 * Malformed ECParameters causes infinite loop
6101 When processing an ECParameters structure OpenSSL enters an infinite loop
6102 if the curve specified is over a specially malformed binary polynomial
6105 This can be used to perform denial of service against any
6106 system which processes public keys, certificate requests or
6107 certificates. This includes TLS clients and TLS servers with
6108 client authentication enabled.
6110 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6115 * Exploitable out-of-bounds read in X509_cmp_time
6117 X509_cmp_time does not properly check the length of the ASN1_TIME
6118 string and can read a few bytes out of bounds. In addition,
6119 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6122 An attacker can use this to craft malformed certificates and CRLs of
6123 various sizes and potentially cause a segmentation fault, resulting in
6124 a DoS on applications that verify certificates or CRLs. TLS clients
6125 that verify CRLs are affected. TLS clients and servers with client
6126 authentication enabled may be affected if they use custom verification
6129 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6130 independently by Hanno Böck.
6135 * PKCS7 crash with missing EnvelopedContent
6137 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6138 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6139 with missing content and trigger a NULL pointer dereference on parsing.
6141 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6142 structures from untrusted sources are affected. OpenSSL clients and
6143 servers are not affected.
6145 This issue was reported to OpenSSL by Michal Zalewski (Google).
6150 * CMS verify infinite loop with unknown hash function
6152 When verifying a signedData message the CMS code can enter an infinite loop
6153 if presented with an unknown hash function OID. This can be used to perform
6154 denial of service against any system which verifies signedData messages using
6156 This issue was reported to OpenSSL by Johannes Bauer.
6161 * Race condition handling NewSessionTicket
6163 If a NewSessionTicket is received by a multi-threaded client when attempting to
6164 reuse a previous ticket then a race condition can occur potentially leading to
6165 a double free of the ticket data.
6170 * Reject DH handshakes with parameters shorter than 768 bits.
6172 *Kurt Roeckx and Emilia Kasper*
6174 * dhparam: generate 2048-bit parameters by default.
6176 *Kurt Roeckx and Emilia Kasper*
6178 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6180 * Segmentation fault in ASN1_TYPE_cmp fix
6182 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6183 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6184 certificate signature algorithm consistency this can be used to crash any
6185 certificate verification operation and exploited in a DoS attack. Any
6186 application which performs certificate verification is vulnerable including
6187 OpenSSL clients and servers which enable client authentication.
6192 * ASN.1 structure reuse memory corruption fix
6194 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6195 memory corruption via an invalid write. Such reuse is and has been
6196 strongly discouraged and is believed to be rare.
6198 Applications that parse structures containing CHOICE or ANY DEFINED BY
6199 components may be affected. Certificate parsing (d2i_X509 and related
6200 functions) are however not affected. OpenSSL clients and servers are
6206 * PKCS7 NULL pointer dereferences fix
6208 The PKCS#7 parsing code does not handle missing outer ContentInfo
6209 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6210 missing content and trigger a NULL pointer dereference on parsing.
6212 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6213 otherwise parse PKCS#7 structures from untrusted sources are
6214 affected. OpenSSL clients and servers are not affected.
6216 This issue was reported to OpenSSL by Michal Zalewski (Google).
6221 * DoS via reachable assert in SSLv2 servers fix
6223 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6224 servers that both support SSLv2 and enable export cipher suites by sending
6225 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6227 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6228 (OpenSSL development team).
6233 * Use After Free following d2i_ECPrivatekey error fix
6235 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6236 could cause a use after free condition. This, in turn, could cause a double
6237 free in several private key parsing functions (such as d2i_PrivateKey
6238 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6239 for applications that receive EC private keys from untrusted
6240 sources. This scenario is considered rare.
6242 This issue was discovered by the BoringSSL project and fixed in their
6248 * X509_to_X509_REQ NULL pointer deref fix
6250 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6251 the certificate key is invalid. This function is rarely used in practice.
6253 This issue was discovered by Brian Carpenter.
6258 * Removed the export ciphers from the DEFAULT ciphers
6262 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6264 * Build fixes for the Windows and OpenVMS platforms
6266 *Matt Caswell and Richard Levitte*
6268 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6270 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6271 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6272 dereference. This could lead to a Denial Of Service attack. Thanks to
6273 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6278 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6279 dtls1_buffer_record function under certain conditions. In particular this
6280 could occur if an attacker sent repeated DTLS records with the same
6281 sequence number but for the next epoch. The memory leak could be exploited
6282 by an attacker in a Denial of Service attack through memory exhaustion.
6283 Thanks to Chris Mueller for reporting this issue.
6288 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6289 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6290 method would be set to NULL which could later result in a NULL pointer
6291 dereference. Thanks to Frank Schmirler for reporting this issue.
6296 * Abort handshake if server key exchange message is omitted for ephemeral
6299 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6300 reporting this issue.
6305 * Remove non-export ephemeral RSA code on client and server. This code
6306 violated the TLS standard by allowing the use of temporary RSA keys in
6307 non-export ciphersuites and could be used by a server to effectively
6308 downgrade the RSA key length used to a value smaller than the server
6309 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6310 INRIA or reporting this issue.
6315 * Fixed issue where DH client certificates are accepted without verification.
6316 An OpenSSL server will accept a DH certificate for client authentication
6317 without the certificate verify message. This effectively allows a client to
6318 authenticate without the use of a private key. This only affects servers
6319 which trust a client certificate authority which issues certificates
6320 containing DH keys: these are extremely rare and hardly ever encountered.
6321 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6327 * Ensure that the session ID context of an SSL is updated when its
6328 SSL_CTX is updated via SSL_set_SSL_CTX.
6330 The session ID context is typically set from the parent SSL_CTX,
6331 and can vary with the CTX.
6335 * Fix various certificate fingerprint issues.
6337 By using non-DER or invalid encodings outside the signed portion of a
6338 certificate the fingerprint can be changed without breaking the signature.
6339 Although no details of the signed portion of the certificate can be changed
6340 this can cause problems with some applications: e.g. those using the
6341 certificate fingerprint for blacklists.
6343 1. Reject signatures with non zero unused bits.
6345 If the BIT STRING containing the signature has non zero unused bits reject
6346 the signature. All current signature algorithms require zero unused bits.
6348 2. Check certificate algorithm consistency.
6350 Check the AlgorithmIdentifier inside TBS matches the one in the
6351 certificate signature. NB: this will result in signature failure
6352 errors for some broken certificates.
6354 Thanks to Konrad Kraszewski from Google for reporting this issue.
6356 3. Check DSA/ECDSA signatures use DER.
6358 Re-encode DSA/ECDSA signatures and compare with the original received
6359 signature. Return an error if there is a mismatch.
6361 This will reject various cases including garbage after signature
6362 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6363 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6364 (negative or with leading zeroes).
6366 Further analysis was conducted and fixes were developed by Stephen Henson
6367 of the OpenSSL core team.
6373 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6374 results on some platforms, including x86_64. This bug occurs at random
6375 with a very low probability, and is not known to be exploitable in any
6376 way, though its exact impact is difficult to determine. Thanks to Pieter
6377 Wuille (Blockstream) who reported this issue and also suggested an initial
6378 fix. Further analysis was conducted by the OpenSSL development team and
6379 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6380 the OpenSSL core team.
6385 * Do not resume sessions on the server if the negotiated protocol
6386 version does not match the session's version. Resuming with a different
6387 version, while not strictly forbidden by the RFC, is of questionable
6388 sanity and breaks all known clients.
6390 *David Benjamin, Emilia Käsper*
6392 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6393 early CCS messages during renegotiation. (Note that because
6394 renegotiation is encrypted, this early CCS was not exploitable.)
6398 * Tighten client-side session ticket handling during renegotiation:
6399 ensure that the client only accepts a session ticket if the server sends
6400 the extension anew in the ServerHello. Previously, a TLS client would
6401 reuse the old extension state and thus accept a session ticket if one was
6402 announced in the initial ServerHello.
6404 Similarly, ensure that the client requires a session ticket if one
6405 was advertised in the ServerHello. Previously, a TLS client would
6406 ignore a missing NewSessionTicket message.
6410 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6414 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6415 sends a carefully crafted handshake message, to cause OpenSSL to fail
6416 to free up to 64k of memory causing a memory leak. This could be
6417 exploited in a Denial Of Service attack. This issue affects OpenSSL
6418 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6419 whether SRTP is used or configured. Implementations of OpenSSL that
6420 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6422 The fix was developed by the OpenSSL team.
6427 * Session Ticket Memory Leak.
6429 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6430 integrity of that ticket is first verified. In the event of a session
6431 ticket integrity check failing, OpenSSL will fail to free memory
6432 causing a memory leak. By sending a large number of invalid session
6433 tickets an attacker could exploit this issue in a Denial Of Service
6439 * Build option no-ssl3 is incomplete.
6441 When OpenSSL is configured with "no-ssl3" as a build option, servers
6442 could accept and complete a SSL 3.0 handshake, and clients could be
6443 configured to send them.
6446 *Akamai and the OpenSSL team*
6448 * Add support for TLS_FALLBACK_SCSV.
6449 Client applications doing fallback retries should call
6450 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6453 *Adam Langley, Bodo Moeller*
6455 * Add additional DigestInfo checks.
6457 Re-encode DigestInto in DER and check against the original when
6458 verifying RSA signature: this will reject any improperly encoded
6459 DigestInfo structures.
6461 Note: this is a precautionary measure and no attacks are currently known.
6465 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6467 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6468 SRP code can be overrun an internal buffer. Add sanity check that
6469 g, A, B < N to SRP code.
6471 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6472 Group for discovering this issue.
6477 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6478 TLS 1.0 instead of higher protocol versions when the ClientHello message
6479 is badly fragmented. This allows a man-in-the-middle attacker to force a
6480 downgrade to TLS 1.0 even if both the server and the client support a
6481 higher protocol version, by modifying the client's TLS records.
6483 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6484 researching this issue.
6489 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6490 to a denial of service attack. A malicious server can crash the client
6491 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6492 ciphersuite and sending carefully crafted handshake messages.
6494 Thanks to Felix Gröbert (Google) for discovering and researching this
6500 * By sending carefully crafted DTLS packets an attacker could cause openssl
6501 to leak memory. This can be exploited through a Denial of Service attack.
6502 Thanks to Adam Langley for discovering and researching this issue.
6507 * An attacker can force openssl to consume large amounts of memory whilst
6508 processing DTLS handshake messages. This can be exploited through a
6509 Denial of Service attack.
6510 Thanks to Adam Langley for discovering and researching this issue.
6515 * An attacker can force an error condition which causes openssl to crash
6516 whilst processing DTLS packets due to memory being freed twice. This
6517 can be exploited through a Denial of Service attack.
6518 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6524 * If a multithreaded client connects to a malicious server using a resumed
6525 session and the server sends an ec point format extension it could write
6526 up to 255 bytes to freed memory.
6528 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6534 * A malicious server can crash an OpenSSL client with a null pointer
6535 dereference (read) by specifying an SRP ciphersuite even though it was not
6536 properly negotiated with the client. This can be exploited through a
6537 Denial of Service attack.
6539 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6540 discovering and researching this issue.
6545 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6546 X509_name_oneline, X509_name_print_ex et al. to leak some information
6547 from the stack. Applications may be affected if they echo pretty printing
6548 output to the attacker.
6550 Thanks to Ivan Fratric (Google) for discovering this issue.
6553 *Emilia Käsper, and Steve Henson*
6555 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6556 for corner cases. (Certain input points at infinity could lead to
6557 bogus results, with non-infinity inputs mapped to infinity too.)
6561 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6563 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6564 handshake can force the use of weak keying material in OpenSSL
6565 SSL/TLS clients and servers.
6567 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6568 researching this issue. ([CVE-2014-0224])
6570 *KIKUCHI Masashi, Steve Henson*
6572 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6573 OpenSSL DTLS client the code can be made to recurse eventually crashing
6576 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6579 *Imre Rad, Steve Henson*
6581 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6582 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6583 client or server. This is potentially exploitable to run arbitrary
6584 code on a vulnerable client or server.
6586 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6588 *Jüri Aedla, Steve Henson*
6590 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6591 are subject to a denial of service attack.
6593 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6594 this issue. ([CVE-2014-3470])
6596 *Felix Gröbert, Ivan Fratric, Steve Henson*
6598 * Harmonize version and its documentation. -f flag is used to display
6601 *mancha <mancha1@zoho.com>*
6603 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6604 in i2d_ECPrivateKey.
6606 *mancha <mancha1@zoho.com>*
6608 * Fix some double frees. These are not thought to be exploitable.
6610 *mancha <mancha1@zoho.com>*
6612 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6614 * A missing bounds check in the handling of the TLS heartbeat extension
6615 can be used to reveal up to 64k of memory to a connected client or
6618 Thanks for Neel Mehta of Google Security for discovering this bug and to
6619 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6620 preparing the fix ([CVE-2014-0160])
6622 *Adam Langley, Bodo Moeller*
6624 * Fix for the attack described in the paper "Recovering OpenSSL
6625 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6626 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6627 <http://eprint.iacr.org/2014/140>
6629 Thanks to Yuval Yarom and Naomi Benger for discovering this
6630 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6632 *Yuval Yarom and Naomi Benger*
6634 * TLS pad extension: draft-agl-tls-padding-03
6636 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6637 TLS client Hello record length value would otherwise be > 255 and
6638 less that 512 pad with a dummy extension containing zeroes so it
6639 is at least 512 bytes long.
6641 *Adam Langley, Steve Henson*
6643 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6645 * Fix for TLS record tampering bug. A carefully crafted invalid
6646 handshake could crash OpenSSL with a NULL pointer exception.
6647 Thanks to Anton Johansson for reporting this issues.
6650 * Keep original DTLS digest and encryption contexts in retransmission
6651 structures so we can use the previous session parameters if they need
6652 to be resent. ([CVE-2013-6450])
6656 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6657 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6658 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6659 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6660 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6661 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6663 *Rob Stradling, Adam Langley*
6665 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6667 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6668 supporting platforms or when small records were transferred.
6670 *Andy Polyakov, Steve Henson*
6672 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6674 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6676 This addresses the flaw in CBC record processing discovered by
6677 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6678 at: <http://www.isg.rhul.ac.uk/tls/>
6680 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6681 Security Group at Royal Holloway, University of London
6682 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6683 Emilia Käsper for the initial patch.
6686 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6688 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6689 ciphersuites which can be exploited in a denial of service attack.
6690 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6691 and detecting this bug and to Wolfgang Ettlinger
6692 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6697 * Return an error when checking OCSP signatures when key is NULL.
6698 This fixes a DoS attack. ([CVE-2013-0166])
6702 * Make openssl verify return errors.
6704 *Chris Palmer <palmer@google.com> and Ben Laurie*
6706 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6707 the right response is stapled. Also change SSL_get_certificate()
6708 so it returns the certificate actually sent.
6709 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6711 *Rob Stradling <rob.stradling@comodo.com>*
6713 * Fix possible deadlock when decoding public keys.
6717 * Don't use TLS 1.0 record version number in initial client hello
6722 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6724 * Sanity check record length before skipping explicit IV in TLS
6725 1.2, 1.1 and DTLS to fix DoS attack.
6727 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6728 fuzzing as a service testing platform.
6733 * Initialise tkeylen properly when encrypting CMS messages.
6734 Thanks to Solar Designer of Openwall for reporting this issue.
6738 * In FIPS mode don't try to use composite ciphers as they are not
6743 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6745 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6746 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6747 mean any application compiled against OpenSSL 1.0.0 headers setting
6748 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6749 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6750 0x10000000L Any application which was previously compiled against
6751 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6752 will need to be recompiled as a result. Letting be results in
6753 inability to disable specifically TLS 1.1 and in client context,
6754 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6758 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6759 disable just protocol X, but all protocols above X *if* there are
6760 protocols *below* X still enabled. In more practical terms it means
6761 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6762 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6763 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6768 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6770 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6771 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6772 in CRYPTO_realloc_clean.
6774 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6775 issue and to Adam Langley <agl@chromium.org> for fixing it.
6778 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6780 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6784 * Workarounds for some broken servers that "hang" if a client hello
6785 record length exceeds 255 bytes.
6787 1. Do not use record version number > TLS 1.0 in initial client
6788 hello: some (but not all) hanging servers will now work.
6789 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6790 the number of ciphers sent in the client hello. This should be
6791 set to an even number, such as 50, for example by passing:
6792 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6793 Most broken servers should now work.
6794 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6795 TLS 1.2 client support entirely.
6799 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6803 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6805 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6806 STRING form instead of a DigestInfo.
6810 * The format used for MDC2 RSA signatures is inconsistent between EVP
6811 and the RSA_sign/RSA_verify functions. This was made more apparent when
6812 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6813 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6814 the correct format in RSA_verify so both forms transparently work.
6818 * Some servers which support TLS 1.0 can choke if we initially indicate
6819 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6820 encrypted premaster secret. As a workaround use the maximum permitted
6821 client version in client hello, this should keep such servers happy
6822 and still work with previous versions of OpenSSL.
6826 * Add support for TLS/DTLS heartbeats.
6828 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6830 * Add support for SCTP.
6832 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6834 * Improved PRNG seeding for VOS.
6836 *Paul Green <Paul.Green@stratus.com>*
6838 * Extensive assembler packs updates, most notably:
6840 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6841 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6842 - x86_64: bit-sliced AES implementation;
6843 - ARM: NEON support, contemporary platforms optimizations;
6844 - s390x: z196 support;
6845 - `*`: GHASH and GF(2^m) multiplication implementations;
6849 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6850 (removal of unnecessary code)
6852 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6854 * Add TLS key material exporter from RFC 5705.
6858 * Add DTLS-SRTP negotiation from RFC 5764.
6862 * Add Next Protocol Negotiation,
6863 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6864 disabled with a no-npn flag to config or Configure. Code donated
6867 *Adam Langley <agl@google.com> and Ben Laurie*
6869 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6870 NIST-P256, NIST-P521, with constant-time single point multiplication on
6871 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6872 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6873 Code made available under Apache License version 2.0.
6875 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6876 line to include this in your build of OpenSSL, and run "make depend" (or
6877 "make update"). This enables the following EC_METHODs:
6879 EC_GFp_nistp224_method()
6880 EC_GFp_nistp256_method()
6881 EC_GFp_nistp521_method()
6883 EC_GROUP_new_by_curve_name() will automatically use these (while
6884 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6887 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6889 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6890 all platforms. Move ssize_t definition from e_os.h to the public
6891 header file e_os2.h as it now appears in public header file cms.h
6895 * New -sigopt option to the ca, req and x509 utilities. Additional
6896 signature parameters can be passed using this option and in
6901 * Add RSA PSS signing function. This will generate and set the
6902 appropriate AlgorithmIdentifiers for PSS based on those in the
6903 corresponding EVP_MD_CTX structure. No application support yet.
6907 * Support for companion algorithm specific ASN1 signing routines.
6908 New function ASN1_item_sign_ctx() signs a pre-initialised
6909 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6910 the appropriate parameters.
6914 * Add new algorithm specific ASN1 verification initialisation function
6915 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6916 handling will be the same no matter what EVP_PKEY_METHOD is used.
6917 Add a PSS handler to support verification of PSS signatures: checked
6918 against a number of sample certificates.
6922 * Add signature printing for PSS. Add PSS OIDs.
6924 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6926 * Add algorithm specific signature printing. An individual ASN1 method
6927 can now print out signatures instead of the standard hex dump.
6929 More complex signatures (e.g. PSS) can print out more meaningful
6930 information. Include DSA version that prints out the signature
6935 * Password based recipient info support for CMS library: implementing
6940 * Split password based encryption into PBES2 and PBKDF2 functions. This
6941 neatly separates the code into cipher and PBE sections and is required
6942 for some algorithms that split PBES2 into separate pieces (such as
6943 password based CMS).
6947 * Session-handling fixes:
6948 - Fix handling of connections that are resuming with a session ID,
6949 but also support Session Tickets.
6950 - Fix a bug that suppressed issuing of a new ticket if the client
6951 presented a ticket with an expired session.
6952 - Try to set the ticket lifetime hint to something reasonable.
6953 - Make tickets shorter by excluding irrelevant information.
6954 - On the client side, don't ignore renewed tickets.
6956 *Adam Langley, Bodo Moeller (Google)*
6958 * Fix PSK session representation.
6962 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6964 This work was sponsored by Intel.
6968 * Add GCM support to TLS library. Some custom code is needed to split
6969 the IV between the fixed (from PRF) and explicit (from TLS record)
6970 portions. This adds all GCM ciphersuites supported by RFC5288 and
6971 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6972 add a special AESGCM string for GCM only.
6976 * Expand range of ctrls for AES GCM. Permit setting invocation
6977 field on decrypt and retrieval of invocation field only on encrypt.
6981 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6982 As required by RFC5289 these ciphersuites cannot be used if for
6983 versions of TLS earlier than 1.2.
6987 * For FIPS capable OpenSSL interpret a NULL default public key method
6988 as unset and return the appropriate default but do *not* set the default.
6989 This means we can return the appropriate method in applications that
6990 switch between FIPS and non-FIPS modes.
6994 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6995 ENGINE is used then we cannot handle that in the FIPS module so we
6996 keep original code iff non-FIPS operations are allowed.
7000 * Add -attime option to openssl utilities.
7002 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7004 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7008 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7009 FIPS EC methods unconditionally for now.
7013 * New build option no-ec2m to disable characteristic 2 code.
7017 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7018 all cases can be covered as some introduce binary incompatibilities.
7022 * Redirect RSA operations to FIPS module including keygen,
7023 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7027 * Add similar low level API blocking to ciphers.
7031 * Low level digest APIs are not approved in FIPS mode: any attempt
7032 to use these will cause a fatal error. Applications that *really* want
7033 to use them can use the `private_*` version instead.
7037 * Redirect cipher operations to FIPS module for FIPS builds.
7041 * Redirect digest operations to FIPS module for FIPS builds.
7045 * Update build system to add "fips" flag which will link in fipscanister.o
7046 for static and shared library builds embedding a signature if needed.
7050 * Output TLS supported curves in preference order instead of numerical
7051 order. This is currently hardcoded for the highest order curves first.
7052 This should be configurable so applications can judge speed vs strength.
7056 * Add TLS v1.2 server support for client authentication.
7060 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7065 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7066 FIPS modules versions.
7070 * Add TLS v1.2 client side support for client authentication. Keep cache
7071 of handshake records longer as we don't know the hash algorithm to use
7072 until after the certificate request message is received.
7076 * Initial TLS v1.2 client support. Add a default signature algorithms
7077 extension including all the algorithms we support. Parse new signature
7078 format in client key exchange. Relax some ECC signing restrictions for
7079 TLS v1.2 as indicated in RFC5246.
7083 * Add server support for TLS v1.2 signature algorithms extension. Switch
7084 to new signature format when needed using client digest preference.
7085 All server ciphersuites should now work correctly in TLS v1.2. No client
7086 support yet and no support for client certificates.
7090 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7091 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7092 ciphersuites. At present only RSA key exchange ciphersuites work with
7093 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7094 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7095 and version checking.
7099 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7100 with this defined it will not be affected by any changes to ssl internal
7101 structures. Add several utility functions to allow openssl application
7102 to work with OPENSSL_NO_SSL_INTERN defined.
7106 * A long standing patch to add support for SRP from EdelWeb (Peter
7107 Sylvester and Christophe Renou) was integrated.
7108 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7109 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7112 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7116 * Permit abbreviated handshakes when renegotiating using the function
7117 SSL_renegotiate_abbreviated().
7119 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7121 * Add call to ENGINE_register_all_complete() to
7122 ENGINE_load_builtin_engines(), so some implementations get used
7123 automatically instead of needing explicit application support.
7127 * Add support for TLS key exporter as described in RFC5705.
7129 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7131 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7132 a few changes are required:
7134 Add SSL_OP_NO_TLSv1_1 flag.
7135 Add TLSv1_1 methods.
7136 Update version checking logic to handle version 1.1.
7137 Add explicit IV handling (ported from DTLS code).
7138 Add command line options to s_client/s_server.
7145 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7147 * X509_ATTRIBUTE memory leak
7149 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7150 memory. This structure is used by the PKCS#7 and CMS routines so any
7151 application which reads PKCS#7 or CMS data from untrusted sources is
7152 affected. SSL/TLS is not affected.
7154 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7160 * Race condition handling PSK identify hint
7162 If PSK identity hints are received by a multi-threaded client then
7163 the values are wrongly updated in the parent SSL_CTX structure. This can
7164 result in a race condition potentially leading to a double free of the
7170 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7172 * Malformed ECParameters causes infinite loop
7174 When processing an ECParameters structure OpenSSL enters an infinite loop
7175 if the curve specified is over a specially malformed binary polynomial
7178 This can be used to perform denial of service against any
7179 system which processes public keys, certificate requests or
7180 certificates. This includes TLS clients and TLS servers with
7181 client authentication enabled.
7183 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7188 * Exploitable out-of-bounds read in X509_cmp_time
7190 X509_cmp_time does not properly check the length of the ASN1_TIME
7191 string and can read a few bytes out of bounds. In addition,
7192 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7195 An attacker can use this to craft malformed certificates and CRLs of
7196 various sizes and potentially cause a segmentation fault, resulting in
7197 a DoS on applications that verify certificates or CRLs. TLS clients
7198 that verify CRLs are affected. TLS clients and servers with client
7199 authentication enabled may be affected if they use custom verification
7202 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7203 independently by Hanno Böck.
7208 * PKCS7 crash with missing EnvelopedContent
7210 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7211 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7212 with missing content and trigger a NULL pointer dereference on parsing.
7214 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7215 structures from untrusted sources are affected. OpenSSL clients and
7216 servers are not affected.
7218 This issue was reported to OpenSSL by Michal Zalewski (Google).
7223 * CMS verify infinite loop with unknown hash function
7225 When verifying a signedData message the CMS code can enter an infinite loop
7226 if presented with an unknown hash function OID. This can be used to perform
7227 denial of service against any system which verifies signedData messages using
7229 This issue was reported to OpenSSL by Johannes Bauer.
7234 * Race condition handling NewSessionTicket
7236 If a NewSessionTicket is received by a multi-threaded client when attempting to
7237 reuse a previous ticket then a race condition can occur potentially leading to
7238 a double free of the ticket data.
7243 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7245 * Segmentation fault in ASN1_TYPE_cmp fix
7247 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7248 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7249 certificate signature algorithm consistency this can be used to crash any
7250 certificate verification operation and exploited in a DoS attack. Any
7251 application which performs certificate verification is vulnerable including
7252 OpenSSL clients and servers which enable client authentication.
7257 * ASN.1 structure reuse memory corruption fix
7259 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7260 memory corruption via an invalid write. Such reuse is and has been
7261 strongly discouraged and is believed to be rare.
7263 Applications that parse structures containing CHOICE or ANY DEFINED BY
7264 components may be affected. Certificate parsing (d2i_X509 and related
7265 functions) are however not affected. OpenSSL clients and servers are
7271 * PKCS7 NULL pointer dereferences fix
7273 The PKCS#7 parsing code does not handle missing outer ContentInfo
7274 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7275 missing content and trigger a NULL pointer dereference on parsing.
7277 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7278 otherwise parse PKCS#7 structures from untrusted sources are
7279 affected. OpenSSL clients and servers are not affected.
7281 This issue was reported to OpenSSL by Michal Zalewski (Google).
7286 * DoS via reachable assert in SSLv2 servers fix
7288 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7289 servers that both support SSLv2 and enable export cipher suites by sending
7290 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7292 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7293 (OpenSSL development team).
7298 * Use After Free following d2i_ECPrivatekey error fix
7300 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7301 could cause a use after free condition. This, in turn, could cause a double
7302 free in several private key parsing functions (such as d2i_PrivateKey
7303 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7304 for applications that receive EC private keys from untrusted
7305 sources. This scenario is considered rare.
7307 This issue was discovered by the BoringSSL project and fixed in their
7313 * X509_to_X509_REQ NULL pointer deref fix
7315 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7316 the certificate key is invalid. This function is rarely used in practice.
7318 This issue was discovered by Brian Carpenter.
7323 * Removed the export ciphers from the DEFAULT ciphers
7327 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7329 * Build fixes for the Windows and OpenVMS platforms
7331 *Matt Caswell and Richard Levitte*
7333 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7335 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7336 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7337 dereference. This could lead to a Denial Of Service attack. Thanks to
7338 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7343 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7344 dtls1_buffer_record function under certain conditions. In particular this
7345 could occur if an attacker sent repeated DTLS records with the same
7346 sequence number but for the next epoch. The memory leak could be exploited
7347 by an attacker in a Denial of Service attack through memory exhaustion.
7348 Thanks to Chris Mueller for reporting this issue.
7353 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7354 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7355 method would be set to NULL which could later result in a NULL pointer
7356 dereference. Thanks to Frank Schmirler for reporting this issue.
7361 * Abort handshake if server key exchange message is omitted for ephemeral
7364 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7365 reporting this issue.
7370 * Remove non-export ephemeral RSA code on client and server. This code
7371 violated the TLS standard by allowing the use of temporary RSA keys in
7372 non-export ciphersuites and could be used by a server to effectively
7373 downgrade the RSA key length used to a value smaller than the server
7374 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7375 INRIA or reporting this issue.
7380 * Fixed issue where DH client certificates are accepted without verification.
7381 An OpenSSL server will accept a DH certificate for client authentication
7382 without the certificate verify message. This effectively allows a client to
7383 authenticate without the use of a private key. This only affects servers
7384 which trust a client certificate authority which issues certificates
7385 containing DH keys: these are extremely rare and hardly ever encountered.
7386 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7392 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7393 results on some platforms, including x86_64. This bug occurs at random
7394 with a very low probability, and is not known to be exploitable in any
7395 way, though its exact impact is difficult to determine. Thanks to Pieter
7396 Wuille (Blockstream) who reported this issue and also suggested an initial
7397 fix. Further analysis was conducted by the OpenSSL development team and
7398 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7399 the OpenSSL core team.
7404 * Fix various certificate fingerprint issues.
7406 By using non-DER or invalid encodings outside the signed portion of a
7407 certificate the fingerprint can be changed without breaking the signature.
7408 Although no details of the signed portion of the certificate can be changed
7409 this can cause problems with some applications: e.g. those using the
7410 certificate fingerprint for blacklists.
7412 1. Reject signatures with non zero unused bits.
7414 If the BIT STRING containing the signature has non zero unused bits reject
7415 the signature. All current signature algorithms require zero unused bits.
7417 2. Check certificate algorithm consistency.
7419 Check the AlgorithmIdentifier inside TBS matches the one in the
7420 certificate signature. NB: this will result in signature failure
7421 errors for some broken certificates.
7423 Thanks to Konrad Kraszewski from Google for reporting this issue.
7425 3. Check DSA/ECDSA signatures use DER.
7427 Reencode DSA/ECDSA signatures and compare with the original received
7428 signature. Return an error if there is a mismatch.
7430 This will reject various cases including garbage after signature
7431 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7432 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7433 (negative or with leading zeroes).
7435 Further analysis was conducted and fixes were developed by Stephen Henson
7436 of the OpenSSL core team.
7442 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7444 * Session Ticket Memory Leak.
7446 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7447 integrity of that ticket is first verified. In the event of a session
7448 ticket integrity check failing, OpenSSL will fail to free memory
7449 causing a memory leak. By sending a large number of invalid session
7450 tickets an attacker could exploit this issue in a Denial Of Service
7456 * Build option no-ssl3 is incomplete.
7458 When OpenSSL is configured with "no-ssl3" as a build option, servers
7459 could accept and complete a SSL 3.0 handshake, and clients could be
7460 configured to send them.
7463 *Akamai and the OpenSSL team*
7465 * Add support for TLS_FALLBACK_SCSV.
7466 Client applications doing fallback retries should call
7467 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7470 *Adam Langley, Bodo Moeller*
7472 * Add additional DigestInfo checks.
7474 Reencode DigestInto in DER and check against the original when
7475 verifying RSA signature: this will reject any improperly encoded
7476 DigestInfo structures.
7478 Note: this is a precautionary measure and no attacks are currently known.
7482 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7484 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7485 to a denial of service attack. A malicious server can crash the client
7486 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7487 ciphersuite and sending carefully crafted handshake messages.
7489 Thanks to Felix Gröbert (Google) for discovering and researching this
7495 * By sending carefully crafted DTLS packets an attacker could cause openssl
7496 to leak memory. This can be exploited through a Denial of Service attack.
7497 Thanks to Adam Langley for discovering and researching this issue.
7502 * An attacker can force openssl to consume large amounts of memory whilst
7503 processing DTLS handshake messages. This can be exploited through a
7504 Denial of Service attack.
7505 Thanks to Adam Langley for discovering and researching this issue.
7510 * An attacker can force an error condition which causes openssl to crash
7511 whilst processing DTLS packets due to memory being freed twice. This
7512 can be exploited through a Denial of Service attack.
7513 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7519 * If a multithreaded client connects to a malicious server using a resumed
7520 session and the server sends an ec point format extension it could write
7521 up to 255 bytes to freed memory.
7523 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7529 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7530 X509_name_oneline, X509_name_print_ex et al. to leak some information
7531 from the stack. Applications may be affected if they echo pretty printing
7532 output to the attacker.
7534 Thanks to Ivan Fratric (Google) for discovering this issue.
7537 *Emilia Käsper, and Steve Henson*
7539 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7540 for corner cases. (Certain input points at infinity could lead to
7541 bogus results, with non-infinity inputs mapped to infinity too.)
7545 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7547 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7548 handshake can force the use of weak keying material in OpenSSL
7549 SSL/TLS clients and servers.
7551 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7552 researching this issue. ([CVE-2014-0224])
7554 *KIKUCHI Masashi, Steve Henson*
7556 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7557 OpenSSL DTLS client the code can be made to recurse eventually crashing
7560 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7563 *Imre Rad, Steve Henson*
7565 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7566 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7567 client or server. This is potentially exploitable to run arbitrary
7568 code on a vulnerable client or server.
7570 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7572 *Jüri Aedla, Steve Henson*
7574 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7575 are subject to a denial of service attack.
7577 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7578 this issue. ([CVE-2014-3470])
7580 *Felix Gröbert, Ivan Fratric, Steve Henson*
7582 * Harmonize version and its documentation. -f flag is used to display
7585 *mancha <mancha1@zoho.com>*
7587 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7588 in i2d_ECPrivateKey.
7590 *mancha <mancha1@zoho.com>*
7592 * Fix some double frees. These are not thought to be exploitable.
7594 *mancha <mancha1@zoho.com>*
7596 * Fix for the attack described in the paper "Recovering OpenSSL
7597 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7598 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7599 <http://eprint.iacr.org/2014/140>
7601 Thanks to Yuval Yarom and Naomi Benger for discovering this
7602 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7604 *Yuval Yarom and Naomi Benger*
7606 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7608 * Keep original DTLS digest and encryption contexts in retransmission
7609 structures so we can use the previous session parameters if they need
7610 to be resent. ([CVE-2013-6450])
7614 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7615 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7616 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7617 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7618 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7619 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7621 *Rob Stradling, Adam Langley*
7623 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7625 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7627 This addresses the flaw in CBC record processing discovered by
7628 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7629 at: <http://www.isg.rhul.ac.uk/tls/>
7631 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7632 Security Group at Royal Holloway, University of London
7633 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7634 Emilia Käsper for the initial patch.
7637 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7639 * Return an error when checking OCSP signatures when key is NULL.
7640 This fixes a DoS attack. ([CVE-2013-0166])
7644 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7645 the right response is stapled. Also change SSL_get_certificate()
7646 so it returns the certificate actually sent.
7647 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7648 (This is a backport)
7650 *Rob Stradling <rob.stradling@comodo.com>*
7652 * Fix possible deadlock when decoding public keys.
7656 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7658 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7661 * Sanity check record length before skipping explicit IV in DTLS
7664 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7665 fuzzing as a service testing platform.
7670 * Initialise tkeylen properly when encrypting CMS messages.
7671 Thanks to Solar Designer of Openwall for reporting this issue.
7675 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7677 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7678 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7679 in CRYPTO_realloc_clean.
7681 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7682 issue and to Adam Langley <agl@chromium.org> for fixing it.
7685 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7687 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7689 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7690 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7691 content decryption and always return the same error. Note: this attack
7692 needs on average 2^20 messages so it only affects automated senders. The
7693 old behaviour can be re-enabled in the CMS code by setting the
7694 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7695 an MMA defence is not necessary.
7696 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7697 this issue. ([CVE-2012-0884])
7701 * Fix CVE-2011-4619: make sure we really are receiving a
7702 client hello before rejecting multiple SGC restarts. Thanks to
7703 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7707 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7709 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7710 Thanks to Antonio Martin, Enterprise Secure Access Research and
7711 Development, Cisco Systems, Inc. for discovering this bug and
7712 preparing a fix. ([CVE-2012-0050])
7716 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7718 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7719 of the Vaudenay padding oracle attack on CBC mode encryption
7720 which enables an efficient plaintext recovery attack against
7721 the OpenSSL implementation of DTLS. Their attack exploits timing
7722 differences arising during decryption processing. A research
7723 paper describing this attack can be found at:
7724 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7725 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7726 Security Group at Royal Holloway, University of London
7727 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7728 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7729 for preparing the fix. ([CVE-2011-4108])
7731 *Robin Seggelmann, Michael Tuexen*
7733 * Clear bytes used for block padding of SSL 3.0 records.
7736 *Adam Langley (Google)*
7738 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7739 Kadianakis <desnacked@gmail.com> for discovering this issue and
7740 Adam Langley for preparing the fix. ([CVE-2011-4619])
7742 *Adam Langley (Google)*
7744 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7746 *Andrey Kulikov <amdeich@gmail.com>*
7748 * Prevent malformed RFC3779 data triggering an assertion failure.
7749 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7750 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7752 *Rob Austein <sra@hactrn.net>*
7754 * Improved PRNG seeding for VOS.
7756 *Paul Green <Paul.Green@stratus.com>*
7758 * Fix ssl_ciph.c set-up race.
7760 *Adam Langley (Google)*
7762 * Fix spurious failures in ecdsatest.c.
7764 *Emilia Käsper (Google)*
7766 * Fix the BIO_f_buffer() implementation (which was mixing different
7767 interpretations of the `..._len` fields).
7769 *Adam Langley (Google)*
7771 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7772 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7773 threads won't reuse the same blinding coefficients.
7775 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7776 lock to call BN_BLINDING_invert_ex, and avoids one use of
7777 BN_BLINDING_update for each BN_BLINDING structure (previously,
7778 the last update always remained unused).
7780 *Emilia Käsper (Google)*
7782 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7784 *Bob Buckholz (Google)*
7786 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7788 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7789 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7791 *Kaspar Brand <ossl@velox.ch>*
7793 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7794 for multi-threaded use of ECDH. ([CVE-2011-3210])
7796 *Adam Langley (Google)*
7798 * Fix x509_name_ex_d2i memory leak on bad inputs.
7802 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7803 signature public key algorithm by using OID xref utilities instead.
7804 Before this you could only use some ECC ciphersuites with SHA1 only.
7808 * Add protection against ECDSA timing attacks as mentioned in the paper
7809 by Billy Bob Brumley and Nicola Tuveri, see:
7810 <http://eprint.iacr.org/2011/232.pdf>
7812 *Billy Bob Brumley and Nicola Tuveri*
7814 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7816 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7818 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7820 * Fix bug in string printing code: if *any* escaping is enabled we must
7821 escape the escape character (backslash) or the resulting string is
7826 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7828 * Disable code workaround for ancient and obsolete Netscape browsers
7829 and servers: an attacker can use it in a ciphersuite downgrade attack.
7830 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7834 * Fixed J-PAKE implementation error, originally discovered by
7835 Sebastien Martini, further info and confirmation from Stefan
7836 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7840 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7842 * Fix extension code to avoid race conditions which can result in a buffer
7843 overrun vulnerability: resumed sessions must not be modified as they can
7844 be shared by multiple threads. CVE-2010-3864
7848 * Fix WIN32 build system to correctly link an ENGINE directory into
7853 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7855 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7858 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7860 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7862 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7863 context. The operation can be customised via the ctrl mechanism in
7864 case ENGINEs want to include additional functionality.
7868 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7872 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7873 output hashes compatible with older versions of OpenSSL.
7875 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7877 * Fix compression algorithm handling: if resuming a session use the
7878 compression algorithm of the resumed session instead of determining
7879 it from client hello again. Don't allow server to change algorithm.
7883 * Add load_crls() function to apps tidying load_certs() too. Add option
7884 to verify utility to allow additional CRLs to be included.
7888 * Update OCSP request code to permit adding custom headers to the request:
7889 some responders need this.
7893 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7896 *Julia Lawall <julia@diku.dk>*
7898 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7899 needlessly dereferenced structures, used obsolete functions and
7900 didn't handle all updated verify codes correctly.
7904 * Disable MD2 in the default configuration.
7908 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7909 indicate the initial BIO being pushed or popped. This makes it possible
7910 to determine whether the BIO is the one explicitly called or as a result
7911 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7912 it handles reference counts correctly and doesn't zero out the I/O bio
7913 when it is not being explicitly popped. WARNING: applications which
7914 included workarounds for the old buggy behaviour will need to be modified
7915 or they could free up already freed BIOs.
7919 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7920 renaming to all platforms (within the 0.9.8 branch, this was
7921 done conditionally on Netware platforms to avoid a name clash).
7923 *Guenter <lists@gknw.net>*
7925 * Add ECDHE and PSK support to DTLS.
7927 *Michael Tuexen <tuexen@fh-muenster.de>*
7929 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7934 * Add "missing" function EVP_MD_flags() (without this the only way to
7935 retrieve a digest flags is by accessing the structure directly. Update
7936 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7937 or cipher is registered as in the "from" argument. Print out all
7938 registered digests in the dgst usage message instead of manually
7939 attempting to work them out.
7943 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7944 this allows the use of compression and extensions. Change default cipher
7945 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7946 by default unless an application cipher string requests it.
7950 * Alter match criteria in PKCS12_parse(). It used to try to use local
7951 key ids to find matching certificates and keys but some PKCS#12 files
7952 don't follow the (somewhat unwritten) rules and this strategy fails.
7953 Now just gather all certificates together and the first private key
7954 then look for the first certificate that matches the key.
7958 * Support use of registered digest and cipher names for dgst and cipher
7959 commands instead of having to add each one as a special case. So now
7966 openssl dgst -sha256 foo
7968 and this works for ENGINE based algorithms too.
7972 * Update Gost ENGINE to support parameter files.
7974 *Victor B. Wagner <vitus@cryptocom.ru>*
7976 * Support GeneralizedTime in ca utility.
7978 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7980 * Enhance the hash format used for certificate directory links. The new
7981 form uses the canonical encoding (meaning equivalent names will work
7982 even if they aren't identical) and uses SHA1 instead of MD5. This form
7983 is incompatible with the older format and as a result c_rehash should
7984 be used to rebuild symbolic links.
7988 * Make PKCS#8 the default write format for private keys, replacing the
7989 traditional format. This form is standardised, more secure and doesn't
7990 include an implicit MD5 dependency.
7994 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7995 committed to OpenSSL should pass this lot as a minimum.
7999 * Add session ticket override functionality for use by EAP-FAST.
8001 *Jouni Malinen <j@w1.fi>*
8003 * Modify HMAC functions to return a value. Since these can be implemented
8004 in an ENGINE errors can occur.
8008 * Type-checked OBJ_bsearch_ex.
8012 * Type-checked OBJ_bsearch. Also some constification necessitated
8013 by type-checking. Still to come: TXT_DB, bsearch(?),
8014 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8019 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8020 seconds to a tm structure directly, instead of going through OS
8021 specific date routines. This avoids any issues with OS routines such
8022 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8023 and X509_time_adj_ex() to cover the extended range. The existing
8024 X509_time_adj() is still usable and will no longer have any date issues.
8028 * Delta CRL support. New use deltas option which will attempt to locate
8029 and search any appropriate delta CRLs available.
8031 This work was sponsored by Google.
8035 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8036 code and add additional score elements. Validate alternate CRL paths
8037 as part of the CRL checking and indicate a new error "CRL path validation
8038 error" in this case. Applications wanting additional details can use
8039 the verify callback and check the new "parent" field. If this is not
8040 NULL CRL path validation is taking place. Existing applications won't
8041 see this because it requires extended CRL support which is off by
8044 This work was sponsored by Google.
8048 * Support for freshest CRL extension.
8050 This work was sponsored by Google.
8054 * Initial indirect CRL support. Currently only supported in the CRLs
8055 passed directly and not via lookup. Process certificate issuer
8056 CRL entry extension and lookup CRL entries by bother issuer name
8057 and serial number. Check and process CRL issuer entry in IDP extension.
8059 This work was sponsored by Google.
8063 * Add support for distinct certificate and CRL paths. The CRL issuer
8064 certificate is validated separately in this case. Only enabled if
8065 an extended CRL support flag is set: this flag will enable additional
8066 CRL functionality in future.
8068 This work was sponsored by Google.
8072 * Add support for policy mappings extension.
8074 This work was sponsored by Google.
8078 * Fixes to pathlength constraint, self issued certificate handling,
8079 policy processing to align with RFC3280 and PKITS tests.
8081 This work was sponsored by Google.
8085 * Support for name constraints certificate extension. DN, email, DNS
8086 and URI types are currently supported.
8088 This work was sponsored by Google.
8092 * To cater for systems that provide a pointer-based thread ID rather
8093 than numeric, deprecate the current numeric thread ID mechanism and
8094 replace it with a structure and associated callback type. This
8095 mechanism allows a numeric "hash" to be extracted from a thread ID in
8096 either case, and on platforms where pointers are larger than 'long',
8097 mixing is done to help ensure the numeric 'hash' is usable even if it
8098 can't be guaranteed unique. The default mechanism is to use "&errno"
8099 as a pointer-based thread ID to distinguish between threads.
8101 Applications that want to provide their own thread IDs should now use
8102 CRYPTO_THREADID_set_callback() to register a callback that will call
8103 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8105 Note that ERR_remove_state() is now deprecated, because it is tied
8106 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8107 to free the current thread's error state should be replaced by
8108 ERR_remove_thread_state(NULL).
8110 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8111 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8112 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8113 application was previously providing a numeric thread callback that
8114 was inappropriate for distinguishing threads, then uniqueness might
8115 have been obtained with &errno that happened immediately in the
8116 intermediate development versions of OpenSSL; this is no longer the
8117 case, the numeric thread callback will now override the automatic use
8120 *Geoff Thorpe, with help from Bodo Moeller*
8122 * Initial support for different CRL issuing certificates. This covers a
8123 simple case where the self issued certificates in the chain exist and
8124 the real CRL issuer is higher in the existing chain.
8126 This work was sponsored by Google.
8130 * Removed effectively defunct crypto/store from the build.
8134 * Revamp of STACK to provide stronger type-checking. Still to come:
8135 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8136 ASN1_STRING, CONF_VALUE.
8140 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8141 RAM on SSL connections. This option can save about 34k per idle SSL.
8145 * Revamp of LHASH to provide stronger type-checking. Still to come:
8146 STACK, TXT_DB, bsearch, qsort.
8150 * Initial support for Cryptographic Message Syntax (aka CMS) based
8151 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8152 support for data, signedData, compressedData, digestedData and
8153 encryptedData, envelopedData types included. Scripts to check against
8154 RFC4134 examples draft and interop and consistency checks of many
8155 content types and variants.
8159 * Add options to enc utility to support use of zlib compression BIO.
8163 * Extend mk1mf to support importing of options and assembly language
8164 files from Configure script, currently only included in VC-WIN32.
8165 The assembly language rules can now optionally generate the source
8166 files from the associated perl scripts.
8170 * Implement remaining functionality needed to support GOST ciphersuites.
8171 Interop testing has been performed using CryptoPro implementations.
8173 *Victor B. Wagner <vitus@cryptocom.ru>*
8175 * s390x assembler pack.
8179 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8184 * Implement Opaque PRF Input TLS extension as specified in
8185 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8186 official specification yet and no extension type assignment by
8187 IANA exists, this extension (for now) will have to be explicitly
8188 enabled when building OpenSSL by providing the extension number
8189 to use. For example, specify an option
8191 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8193 to the "config" or "Configure" script to enable the extension,
8194 assuming extension number 0x9527 (which is a completely arbitrary
8195 and unofficial assignment based on the MD5 hash of the Internet
8196 Draft). Note that by doing so, you potentially lose
8197 interoperability with other TLS implementations since these might
8198 be using the same extension number for other purposes.
8200 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8201 opaque PRF input value to use in the handshake. This will create
8202 an internal copy of the length-'len' string at 'src', and will
8203 return non-zero for success.
8205 To get more control and flexibility, provide a callback function
8208 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8209 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8213 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8216 Callback function 'cb' will be called in handshakes, and is
8217 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8218 Argument 'arg' is for application purposes (the value as given to
8219 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8220 be provided to the callback function). The callback function
8221 has to return non-zero to report success: usually 1 to use opaque
8222 PRF input just if possible, or 2 to enforce use of the opaque PRF
8223 input. In the latter case, the library will abort the handshake
8224 if opaque PRF input is not successfully negotiated.
8226 Arguments 'peerinput' and 'len' given to the callback function
8227 will always be NULL and 0 in the case of a client. A server will
8228 see the client's opaque PRF input through these variables if
8229 available (NULL and 0 otherwise). Note that if the server
8230 provides an opaque PRF input, the length must be the same as the
8231 length of the client's opaque PRF input.
8233 Note that the callback function will only be called when creating
8234 a new session (session resumption can resume whatever was
8235 previously negotiated), and will not be called in SSL 2.0
8236 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8237 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8238 for applications that need to enforce opaque PRF input.
8242 * Update ssl code to support digests other than SHA1+MD5 for handshake
8245 *Victor B. Wagner <vitus@cryptocom.ru>*
8247 * Add RFC4507 support to OpenSSL. This includes the corrections in
8248 RFC4507bis. The encrypted ticket format is an encrypted encoded
8249 SSL_SESSION structure, that way new session features are automatically
8252 If a client application caches session in an SSL_SESSION structure
8253 support is transparent because tickets are now stored in the encoded
8256 The SSL_CTX structure automatically generates keys for ticket
8257 protection in servers so again support should be possible
8258 with no application modification.
8260 If a client or server wishes to disable RFC4507 support then the option
8261 SSL_OP_NO_TICKET can be set.
8263 Add a TLS extension debugging callback to allow the contents of any client
8264 or server extensions to be examined.
8266 This work was sponsored by Google.
8270 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8271 OpenSSL should now compile cleanly on gcc 4.2
8273 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8275 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8276 support including streaming MAC support: this is required for GOST
8277 ciphersuite support.
8279 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8281 * Add option -stream to use PKCS#7 streaming in smime utility. New
8282 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8283 to output in BER and PEM format.
8287 * Experimental support for use of HMAC via EVP_PKEY interface. This
8288 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8289 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8290 ENGINE support for HMAC keys which are unextractable. New -mac and
8291 -macopt options to dgst utility.
8295 * New option -sigopt to dgst utility. Update dgst to use
8296 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8297 alternative signing parameters such as X9.31 or PSS in the dgst
8302 * Change ssl_cipher_apply_rule(), the internal function that does
8303 the work each time a ciphersuite string requests enabling
8304 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8305 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8306 the order of disabled ciphersuites such that those ciphersuites
8307 that most recently went from enabled to disabled not only stay
8308 in order with respect to each other, but also have higher priority
8309 than other disabled ciphersuites the next time ciphersuites are
8312 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8313 the same ciphersuites as with "HIGH" alone, but in a specific
8314 order where the PSK ciphersuites come first (since they are the
8315 most recently disabled ciphersuites when "HIGH" is parsed).
8317 Also, change ssl_create_cipher_list() (using this new
8318 functionality) such that between otherwise identical
8319 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8324 * Change ssl_create_cipher_list() so that it automatically
8325 arranges the ciphersuites in reasonable order before starting
8326 to process the rule string. Thus, the definition for "DEFAULT"
8327 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8328 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8329 This makes it much easier to arrive at a reasonable default order
8330 in applications for which anonymous ciphers are OK (meaning
8331 that you can't actually use DEFAULT).
8333 *Bodo Moeller; suggested by Victor Duchovni*
8335 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8336 processing) into multiple integers instead of setting
8337 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8338 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8339 (These masks as well as the individual bit definitions are hidden
8340 away into the non-exported interface ssl/ssl_locl.h, so this
8341 change to the definition of the SSL_CIPHER structure shouldn't
8342 affect applications.) This give us more bits for each of these
8343 categories, so there is no longer a need to coagulate AES128 and
8344 AES256 into a single algorithm bit, and to coagulate Camellia128
8345 and Camellia256 into a single algorithm bit, which has led to all
8348 Thus, among other things, the kludge introduced in 0.9.7m and
8349 0.9.8e for masking out AES256 independently of AES128 or masking
8350 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8352 With the change, we also introduce new ciphersuite aliases that
8353 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8358 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8359 Use the leftmost N bytes of the signature input if the input is
8360 larger than the prime q (with N being the size in bytes of q).
8364 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8365 it yet and it is largely untested.
8369 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8373 * Initial incomplete changes to avoid need for function casts in OpenSSL
8374 some compilers (gcc 4.2 and later) reject their use. Safestack is
8375 reimplemented. Update ASN1 to avoid use of legacy functions.
8379 * Win32/64 targets are linked with Winsock2.
8383 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8384 to external functions. This can be used to increase CRL handling
8385 efficiency especially when CRLs are very large by (for example) storing
8386 the CRL revoked certificates in a database.
8390 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8391 new CRLs added to a directory can be used. New command line option
8392 -verify_return_error to s_client and s_server. This causes real errors
8393 to be returned by the verify callback instead of carrying on no matter
8394 what. This reflects the way a "real world" verify callback would behave.
8398 * GOST engine, supporting several GOST algorithms and public key formats.
8399 Kindly donated by Cryptocom.
8403 * Partial support for Issuing Distribution Point CRL extension. CRLs
8404 partitioned by DP are handled but no indirect CRL or reason partitioning
8405 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8406 selected via a scoring technique which handles IDP and AKID in CRLs.
8410 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8411 will ultimately be used for all verify operations: this will remove the
8412 X509_STORE dependency on certificate verification and allow alternative
8413 lookup methods. X509_STORE based implementations of these two callbacks.
8417 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8418 Modify get_crl() to find a valid (unexpired) CRL if possible.
8422 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8423 this would be called X509_CRL_cmp() but that name is already used by
8424 a function that just compares CRL issuer names. Cache several CRL
8425 extensions in X509_CRL structure and cache CRLDP in X509.
8429 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8430 this maps equivalent X509_NAME structures into a consistent structure.
8431 Name comparison can then be performed rapidly using memcmp().
8435 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8440 * Allow digests to supply their own micalg string for S/MIME type using
8441 the ctrl EVP_MD_CTRL_MICALG.
8445 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8446 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8447 ctrl. It can then customise the structure before and/or after signing
8452 * New function OBJ_add_sigid() to allow application defined signature OIDs
8453 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8454 to free up any added signature OIDs.
8458 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8459 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8460 digest and cipher tables. New options added to openssl utility:
8461 list-message-digest-algorithms and list-cipher-algorithms.
8465 * Change the array representation of binary polynomials: the list
8466 of degrees of non-zero coefficients is now terminated with -1.
8467 Previously it was terminated with 0, which was also part of the
8468 value; thus, the array representation was not applicable to
8469 polynomials where t^0 has coefficient zero. This change makes
8470 the array representation useful in a more general context.
8474 * Various modifications and fixes to SSL/TLS cipher string
8475 handling. For ECC, the code now distinguishes between fixed ECDH
8476 with RSA certificates on the one hand and with ECDSA certificates
8477 on the other hand, since these are separate ciphersuites. The
8478 unused code for Fortezza ciphersuites has been removed.
8480 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8481 (not "ECDHE"). For consistency with the code for DH
8482 certificates, use of ECDH certificates is now considered ECDH
8483 authentication, not RSA or ECDSA authentication (the latter is
8484 merely the CA's signing algorithm and not actively used in the
8487 The temporary ciphersuite alias "ECCdraft" is no longer
8488 available, and ECC ciphersuites are no longer excluded from "ALL"
8489 and "DEFAULT". The following aliases now exist for RFC 4492
8490 ciphersuites, most of these by analogy with the DH case:
8492 kECDHr - ECDH cert, signed with RSA
8493 kECDHe - ECDH cert, signed with ECDSA
8494 kECDH - ECDH cert (signed with either RSA or ECDSA)
8495 kEECDH - ephemeral ECDH
8496 ECDH - ECDH cert or ephemeral ECDH
8502 AECDH - anonymous ECDH
8503 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8507 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8508 Use correct micalg parameters depending on digest(s) in signed message.
8512 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8513 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8517 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8518 an engine to register a method. Add ENGINE lookups for methods and
8519 functional reference processing.
8523 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8524 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8529 * New -resign option to smime utility. This adds one or more signers
8530 to an existing PKCS#7 signedData structure. Also -md option to use an
8531 alternative message digest algorithm for signing.
8535 * Tidy up PKCS#7 routines and add new functions to make it easier to
8536 create PKCS7 structures containing multiple signers. Update smime
8537 application to support multiple signers.
8541 * New -macalg option to pkcs12 utility to allow setting of an alternative
8546 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8547 Reorganize PBE internals to lookup from a static table using NIDs,
8548 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8549 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8550 PRF which will be automatically used with PBES2.
8554 * Replace the algorithm specific calls to generate keys in "req" with the
8559 * Update PKCS#7 enveloped data routines to use new API. This is now
8560 supported by any public key method supporting the encrypt operation. A
8561 ctrl is added to allow the public key algorithm to examine or modify
8562 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8567 * Add a ctrl to asn1 method to allow a public key algorithm to express
8568 a default digest type to use. In most cases this will be SHA1 but some
8569 algorithms (such as GOST) need to specify an alternative digest. The
8570 return value indicates how strong the preference is 1 means optional and
8571 2 is mandatory (that is it is the only supported type). Modify
8572 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8573 use the default md. Update openssl utilities to use the default digest
8574 type for signing if it is not explicitly indicated.
8578 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8579 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8580 signing method from the key type. This effectively removes the link
8581 between digests and public key types.
8585 * Add an OID cross reference table and utility functions. Its purpose is to
8586 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8587 rsaEncryption. This will allow some of the algorithm specific hackery
8588 needed to use the correct OID to be removed.
8592 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8593 structures for PKCS7_sign(). They are now set up by the relevant public
8598 * Add provisional EC pkey method with support for ECDSA and ECDH.
8602 * Add support for key derivation (agreement) in the API, DH method and
8607 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8608 public and private key formats. As a side effect these add additional
8609 command line functionality not previously available: DSA signatures can be
8610 generated and verified using pkeyutl and DH key support and generation in
8617 *Oliver Tappe <zooey@hirschkaefer.de>*
8619 * New make target "install_html_docs" installs HTML renditions of the
8622 *Oliver Tappe <zooey@hirschkaefer.de>*
8624 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8625 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8626 support key and parameter generation and add initial key generation
8627 functionality for RSA.
8631 * Add functions for main EVP_PKEY_method operations. The undocumented
8632 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8633 `EVP_PKEY_{encrypt,decrypt}_old`.
8637 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8638 key API, doesn't do much yet.
8642 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8643 public key algorithms. New option to openssl utility:
8644 "list-public-key-algorithms" to print out info.
8648 * Implement the Supported Elliptic Curves Extension for
8649 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8653 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8654 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8658 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8659 utilities such as rsa, dsa, dsaparam etc except they process any key
8664 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8665 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8666 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8671 * Initial support for pluggable public key ASN1.
8672 De-spaghettify the public key ASN1 handling. Move public and private
8673 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8674 algorithm specific handling to a single module within the relevant
8675 algorithm directory. Add functions to allow (near) opaque processing
8676 of public and private key structures.
8680 * Implement the Supported Point Formats Extension for
8681 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8685 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8686 for the psk identity [hint] and the psk callback functions to the
8687 SSL_SESSION, SSL and SSL_CTX structure.
8690 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8694 SSL_CTX_use_psk_identity_hint
8695 SSL_get_psk_identity_hint
8696 SSL_get_psk_identity
8697 SSL_use_psk_identity_hint
8699 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8701 * Add RFC 3161 compliant time stamp request creation, response generation
8702 and response verification functionality.
8704 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8706 * Add initial support for TLS extensions, specifically for the server_name
8707 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8708 have new members for a host name. The SSL data structure has an
8709 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8710 stored in that context to allow for session resumption, even after the
8711 SSL has been switched to a new SSL_CTX in reaction to a client's
8712 server_name extension.
8714 New functions (subject to change):
8716 SSL_get_servername()
8717 SSL_get_servername_type()
8720 New CTRL codes and macros (subject to change):
8722 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8723 - SSL_CTX_set_tlsext_servername_callback()
8724 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8725 - SSL_CTX_set_tlsext_servername_arg()
8726 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8728 openssl s_client has a new '-servername ...' option.
8730 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8731 '-key2 ...', '-servername_fatal' (subject to change). This allows
8732 testing the HostName extension for a specific single host name ('-cert'
8733 and '-key' remain fallbacks for handshakes without HostName
8734 negotiation). If the unrecognized_name alert has to be sent, this by
8735 default is a warning; it becomes fatal with the '-servername_fatal'
8738 *Peter Sylvester, Remy Allais, Christophe Renou*
8740 * Whirlpool hash implementation is added.
8744 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8745 bn(64,32). Because of instruction set limitations it doesn't have
8746 any negative impact on performance. This was done mostly in order
8747 to make it possible to share assembler modules, such as bn_mul_mont
8748 implementations, between 32- and 64-bit builds without hassle.
8752 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8753 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8758 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8759 dedicated Montgomery multiplication procedure, is introduced.
8760 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8761 "64-bit" performance on certain 32-bit targets.
8765 * New option SSL_OP_NO_COMP to disable use of compression selectively
8766 in SSL structures. New SSL ctrl to set maximum send fragment size.
8767 Save memory by setting the I/O buffer sizes dynamically instead of
8768 using the maximum available value.
8772 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8773 in addition to the text details.
8777 * Very, very preliminary EXPERIMENTAL support for printing of general
8778 ASN1 structures. This currently produces rather ugly output and doesn't
8779 handle several customised structures at all.
8783 * Integrated support for PVK file format and some related formats such
8784 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8785 these in the 'rsa' and 'dsa' utilities.
8789 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8793 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8794 place for the (very old) "NETSCAPE" format certificates which are now
8795 handled using new ASN1 code equivalents.
8799 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8800 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8801 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8805 * Modify CRL distribution points extension code to print out previously
8806 unsupported fields. Enhance extension setting code to allow setting of
8811 * Add print and set support for Issuing Distribution Point CRL extension.
8815 * Change 'Configure' script to enable Camellia by default.
8822 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8824 * When rejecting SSL/TLS records due to an incorrect version number, never
8825 update s->server with a new major version number. As of
8826 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8827 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8828 the previous behavior could result in a read attempt at NULL when
8829 receiving specific incorrect SSL/TLS records once record payload
8830 protection is active. ([CVE-2010-0740])
8832 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8834 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8835 could be crashed if the relevant tables were not present (e.g. chrooted).
8837 *Tomas Hoger <thoger@redhat.com>*
8839 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8841 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8843 *Martin Olsson, Neel Mehta*
8845 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8846 accommodate for stack sorting, always a write lock!).
8850 * On some versions of WIN32 Heap32Next is very slow. This can cause
8851 excessive delays in the RAND_poll(): over a minute. As a workaround
8852 include a time check in the inner Heap32Next loop too.
8856 * The code that handled flushing of data in SSL/TLS originally used the
8857 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8858 the problem outlined in PR#1949. The fix suggested there however can
8859 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8860 of Apache). So instead simplify the code to flush unconditionally.
8861 This should be fine since flushing with no data to flush is a no op.
8865 * Handle TLS versions 2.0 and later properly and correctly use the
8866 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8867 off ancient servers have a habit of sticking around for a while...
8871 * Modify compression code so it frees up structures without using the
8872 ex_data callbacks. This works around a problem where some applications
8873 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8874 restarting) then use compression (e.g. SSL with compression) later.
8875 This results in significant per-connection memory leaks and
8876 has caused some security issues including CVE-2008-1678 and
8881 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8882 change when encrypting or decrypting.
8886 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8887 connect and renegotiate with servers which do not support RI.
8888 Until RI is more widely deployed this option is enabled by default.
8892 * Add "missing" ssl ctrls to clear options and mode.
8896 * If client attempts to renegotiate and doesn't support RI respond with
8897 a no_renegotiation alert as required by RFC5746. Some renegotiating
8898 TLS clients will continue a connection gracefully when they receive
8899 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8900 waiting for a server hello which it will never receive. Now we treat a
8901 received no_renegotiation alert as a fatal error. This is because
8902 applications requesting a renegotiation might well expect it to succeed
8903 and would have no code in place to handle the server denying it so the
8904 only safe thing to do is to terminate the connection.
8908 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8909 peer supports secure renegotiation and 0 otherwise. Print out peer
8910 renegotiation support in s_client/s_server.
8914 * Replace the highly broken and deprecated SPKAC certification method with
8915 the updated NID creation version. This should correctly handle UTF8.
8919 * Implement RFC5746. Re-enable renegotiation but require the extension
8920 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8921 turns out to be a bad idea. It has been replaced by
8922 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8923 SSL_CTX_set_options(). This is really not recommended unless you
8924 know what you are doing.
8926 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8928 * Fixes to stateless session resumption handling. Use initial_ctx when
8929 issuing and attempting to decrypt tickets in case it has changed during
8930 servername handling. Use a non-zero length session ID when attempting
8931 stateless session resumption: this makes it possible to determine if
8932 a resumption has occurred immediately after receiving server hello
8933 (several places in OpenSSL subtly assume this) instead of later in
8938 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8939 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8940 fixes for a few places where the return code is not checked
8943 *Julia Lawall <julia@diku.dk>*
8945 * Add --strict-warnings option to Configure script to include devteam
8946 warnings in other configurations.
8950 * Add support for --libdir option and LIBDIR variable in makefiles. This
8951 makes it possible to install openssl libraries in locations which
8952 have names other than "lib", for example "/usr/lib64" which some
8955 *Steve Henson, based on patch from Jeremy Utley*
8957 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8958 X690 8.9.12 and can produce some misleading textual output of OIDs.
8960 *Steve Henson, reported by Dan Kaminsky*
8962 * Delete MD2 from algorithm tables. This follows the recommendation in
8963 several standards that it is not used in new applications due to
8964 several cryptographic weaknesses. For binary compatibility reasons
8965 the MD2 API is still compiled in by default.
8969 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8974 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8975 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8978 *Guenter <lists@gknw.net>*
8980 * Fix the server certificate chain building code to use X509_verify_cert(),
8981 it used to have an ad-hoc builder which was unable to cope with anything
8982 other than a simple chain.
8984 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8986 * Don't check self signed certificate signatures in X509_verify_cert()
8987 by default (a flag can override this): it just wastes time without
8988 adding any security. As a useful side effect self signed root CAs
8989 with non-FIPS digests are now usable in FIPS mode.
8993 * In dtls1_process_out_of_seq_message() the check if the current message
8994 is already buffered was missing. For every new message was memory
8995 allocated, allowing an attacker to perform an denial of service attack
8996 with sending out of seq handshake messages until there is no memory
8997 left. Additionally every future message was buffered, even if the
8998 sequence number made no sense and would be part of another handshake.
8999 So only messages with sequence numbers less than 10 in advance will be
9000 buffered. ([CVE-2009-1378])
9002 *Robin Seggelmann, discovered by Daniel Mentz*
9004 * Records are buffered if they arrive with a future epoch to be
9005 processed after finishing the corresponding handshake. There is
9006 currently no limitation to this buffer allowing an attacker to perform
9007 a DOS attack with sending records with future epochs until there is no
9008 memory left. This patch adds the pqueue_size() function to determine
9009 the size of a buffer and limits the record buffer to 100 entries.
9012 *Robin Seggelmann, discovered by Daniel Mentz*
9014 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9015 parent structure is freed. ([CVE-2009-1379])
9019 * Handle non-blocking I/O properly in SSL_shutdown() call.
9021 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9023 * Add `2.5.4.*` OIDs
9025 *Ilya O. <vrghost@gmail.com>*
9027 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9029 * Disable renegotiation completely - this fixes a severe security
9030 problem ([CVE-2009-3555]) at the cost of breaking all
9031 renegotiation. Renegotiation can be re-enabled by setting
9032 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9033 run-time. This is really not recommended unless you know what
9038 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9040 * Don't set val to NULL when freeing up structures, it is freed up by
9041 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9042 zeroing past the valid field. ([CVE-2009-0789])
9044 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9046 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9047 checked correctly. This would allow some invalid signed attributes to
9048 appear to verify correctly. ([CVE-2009-0591])
9050 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9052 * Reject UniversalString and BMPString types with invalid lengths. This
9053 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9054 a legal length. ([CVE-2009-0590])
9058 * Set S/MIME signing as the default purpose rather than setting it
9059 unconditionally. This allows applications to override it at the store
9064 * Permit restricted recursion of ASN1 strings. This is needed in practice
9065 to handle some structures.
9069 * Improve efficiency of mem_gets: don't search whole buffer each time
9072 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9074 * New -hex option for openssl rand.
9078 * Print out UTF8String and NumericString when parsing ASN1.
9082 * Support NumericString type for name components.
9086 * Allow CC in the environment to override the automatically chosen
9087 compiler. Note that nothing is done to ensure flags work with the
9092 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9094 * Properly check EVP_VerifyFinal() and similar return values
9097 *Ben Laurie, Bodo Moeller, Google Security Team*
9099 * Enable TLS extensions by default.
9103 * Allow the CHIL engine to be loaded, whether the application is
9104 multithreaded or not. (This does not release the developer from the
9105 obligation to set up the dynamic locking callbacks.)
9107 *Sander Temme <sander@temme.net>*
9109 * Use correct exit code if there is an error in dgst command.
9111 *Steve Henson; problem pointed out by Roland Dirlewanger*
9113 * Tweak Configure so that you need to say "experimental-jpake" to enable
9114 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9118 * Add experimental JPAKE support, including demo authentication in
9119 s_client and s_server.
9123 * Set the comparison function in v3_addr_canonize().
9125 *Rob Austein <sra@hactrn.net>*
9127 * Add support for XMPP STARTTLS in s_client.
9129 *Philip Paeps <philip@freebsd.org>*
9131 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9132 to ensure that even with this option, only ciphersuites in the
9133 server's preference list will be accepted. (Note that the option
9134 applies only when resuming a session, so the earlier behavior was
9135 just about the algorithm choice for symmetric cryptography.)
9139 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9141 * Fix NULL pointer dereference if a DTLS server received
9142 ChangeCipherSpec as first record ([CVE-2009-1386]).
9146 * Fix a state transition in s3_srvr.c and d1_srvr.c
9147 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9151 * The fix in 0.9.8c that supposedly got rid of unsafe
9152 double-checked locking was incomplete for RSA blinding,
9153 addressing just one layer of what turns out to have been
9154 doubly unsafe triple-checked locking.
9156 So now fix this for real by retiring the MONT_HELPER macro
9157 in crypto/rsa/rsa_eay.c.
9159 *Bodo Moeller; problem pointed out by Marius Schilder*
9161 * Various precautionary measures:
9163 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9165 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9166 (NB: This would require knowledge of the secret session ticket key
9167 to exploit, in which case you'd be SOL either way.)
9169 - Change bn_nist.c so that it will properly handle input BIGNUMs
9170 outside the expected range.
9172 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9175 *Neel Mehta, Bodo Moeller*
9177 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9178 the load fails. Useful for distros.
9180 *Ben Laurie and the FreeBSD team*
9182 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9186 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9190 * Expand ENGINE to support engine supplied SSL client certificate functions.
9192 This work was sponsored by Logica.
9196 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9197 keystores. Support for SSL/TLS client authentication too.
9198 Not compiled unless enable-capieng specified to Configure.
9200 This work was sponsored by Logica.
9204 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9205 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9206 attribute creation routines such as certificate requests and PKCS#12
9211 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9213 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9214 handshake which could lead to a client crash as found using the
9215 Codenomicon TLS test suite ([CVE-2008-1672])
9217 *Steve Henson, Mark Cox*
9219 * Fix double free in TLS server name extensions which could lead to
9220 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9224 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9226 Clear the error queue to ensure that error entries left from
9227 older function calls do not interfere with the correct operation.
9229 *Lutz Jaenicke, Erik de Castro Lopo*
9231 * Remove root CA certificates of commercial CAs:
9233 The OpenSSL project does not recommend any specific CA and does not
9234 have any policy with respect to including or excluding any CA.
9235 Therefore it does not make any sense to ship an arbitrary selection
9236 of root CA certificates with the OpenSSL software.
9240 * RSA OAEP patches to fix two separate invalid memory reads.
9241 The first one involves inputs when 'lzero' is greater than
9242 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9243 before the beginning of from). The second one involves inputs where
9244 the 'db' section contains nothing but zeroes (there is a one-byte
9245 invalid read after the end of 'db').
9247 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9249 * Partial backport from 0.9.9-dev:
9251 Introduce bn_mul_mont (dedicated Montgomery multiplication
9252 procedure) as a candidate for BIGNUM assembler implementation.
9253 While 0.9.9-dev uses assembler for various architectures, only
9254 x86_64 is available by default here in the 0.9.8 branch, and
9255 32-bit x86 is available through a compile-time setting.
9257 To try the 32-bit x86 assembler implementation, use Configure
9258 option "enable-montasm" (which exists only for this backport).
9260 As "enable-montasm" for 32-bit x86 disclaims code stability
9261 anyway, in this constellation we activate additional code
9262 backported from 0.9.9-dev for further performance improvements,
9263 namely BN_from_montgomery_word. (To enable this otherwise,
9264 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9266 *Andy Polyakov (backport partially by Bodo Moeller)*
9268 * Add TLS session ticket callback. This allows an application to set
9269 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9270 values. This is useful for key rollover for example where several key
9271 sets may exist with different names.
9275 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9276 This was broken until now in 0.9.8 releases, such that the only way
9277 a registered ENGINE could be used (assuming it initialises
9278 successfully on the host) was to explicitly set it as the default
9279 for the relevant algorithms. This is in contradiction with 0.9.7
9280 behaviour and the documentation. With this fix, when an ENGINE is
9281 registered into a given algorithm's table of implementations, the
9282 'uptodate' flag is reset so that auto-discovery will be used next
9283 time a new context for that algorithm attempts to select an
9286 *Ian Lister (tweaked by Geoff Thorpe)*
9288 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9289 implementation in the following ways:
9291 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9294 Lack of BER streaming support means one pass streaming processing is
9295 only supported if data is detached: setting the streaming flag is
9296 ignored for embedded content.
9298 CMS support is disabled by default and must be explicitly enabled
9299 with the enable-cms configuration option.
9303 * Update the GMP engine glue to do direct copies between BIGNUM and
9304 mpz_t when openssl and GMP use the same limb size. Otherwise the
9305 existing "conversion via a text string export" trick is still used.
9307 *Paul Sheer <paulsheer@gmail.com>*
9309 * Zlib compression BIO. This is a filter BIO which compressed and
9310 uncompresses any data passed through it.
9314 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9315 RFC3394 compatible AES key wrapping.
9319 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9320 sets string data without copying. X509_ALGOR_set0() and
9321 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9322 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9323 from an X509_ATTRIBUTE structure optionally checking it occurs only
9324 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9329 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9330 to get the expected BN_FLG_CONSTTIME behavior.
9332 *Bodo Moeller (Google)*
9336 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9337 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9338 - added some more tests to do_tests.pl
9339 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9340 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9341 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9342 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9343 - various changes to netware.pl to enable gcc-cross builds on Win32
9345 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9346 - various changes to fix missing prototype warnings
9347 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9348 - added AES, WHIRLPOOL and CPUID assembler code to build files
9349 - added missing AES assembler make rules to mk1mf.pl
9350 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9352 *Guenter Knauf <eflash@gmx.net>*
9354 * Implement certificate status request TLS extension defined in RFC3546.
9355 A client can set the appropriate parameters and receive the encoded
9356 OCSP response via a callback. A server can query the supplied parameters
9357 and set the encoded OCSP response in the callback. Add simplified examples
9358 to s_client and s_server.
9362 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9365 + Binary incompatibility of ssl_ctx_st structure
9366 + DTLS interoperation with non-compliant servers
9367 + Don't call get_session_cb() without proposed session
9368 + Fix ia64 assembler code
9370 *Andy Polyakov, Steve Henson*
9372 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9374 * DTLS Handshake overhaul. There were longstanding issues with
9375 OpenSSL DTLS implementation, which were making it impossible for
9376 RFC 4347 compliant client to communicate with OpenSSL server.
9377 Unfortunately just fixing these incompatibilities would "cut off"
9378 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9379 server keeps tolerating non RFC compliant syntax. The opposite is
9380 not true, 0.9.8f client can not communicate with earlier server.
9381 This update even addresses CVE-2007-4995.
9385 * Changes to avoid need for function casts in OpenSSL: some compilers
9386 (gcc 4.2 and later) reject their use.
9387 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9390 * Add RFC4507 support to OpenSSL. This includes the corrections in
9391 RFC4507bis. The encrypted ticket format is an encrypted encoded
9392 SSL_SESSION structure, that way new session features are automatically
9395 If a client application caches session in an SSL_SESSION structure
9396 support is transparent because tickets are now stored in the encoded
9399 The SSL_CTX structure automatically generates keys for ticket
9400 protection in servers so again support should be possible
9401 with no application modification.
9403 If a client or server wishes to disable RFC4507 support then the option
9404 SSL_OP_NO_TICKET can be set.
9406 Add a TLS extension debugging callback to allow the contents of any client
9407 or server extensions to be examined.
9409 This work was sponsored by Google.
9413 * Add initial support for TLS extensions, specifically for the server_name
9414 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9415 have new members for a host name. The SSL data structure has an
9416 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9417 stored in that context to allow for session resumption, even after the
9418 SSL has been switched to a new SSL_CTX in reaction to a client's
9419 server_name extension.
9421 New functions (subject to change):
9423 SSL_get_servername()
9424 SSL_get_servername_type()
9427 New CTRL codes and macros (subject to change):
9429 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9430 - SSL_CTX_set_tlsext_servername_callback()
9431 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9432 - SSL_CTX_set_tlsext_servername_arg()
9433 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9435 openssl s_client has a new '-servername ...' option.
9437 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9438 '-key2 ...', '-servername_fatal' (subject to change). This allows
9439 testing the HostName extension for a specific single host name ('-cert'
9440 and '-key' remain fallbacks for handshakes without HostName
9441 negotiation). If the unrecognized_name alert has to be sent, this by
9442 default is a warning; it becomes fatal with the '-servername_fatal'
9445 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9447 * Add AES and SSE2 assembly language support to VC++ build.
9451 * Mitigate attack on final subtraction in Montgomery reduction.
9455 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9456 (which previously caused an internal error).
9460 * Squeeze another 10% out of IGE mode when in != out.
9464 * AES IGE mode speedup.
9466 *Dean Gaudet (Google)*
9468 * Add the Korean symmetric 128-bit cipher SEED (see
9469 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9470 add SEED ciphersuites from RFC 4162:
9472 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9473 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9474 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9475 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9477 To minimize changes between patchlevels in the OpenSSL 0.9.8
9478 series, SEED remains excluded from compilation unless OpenSSL
9479 is configured with 'enable-seed'.
9481 *KISA, Bodo Moeller*
9483 * Mitigate branch prediction attacks, which can be practical if a
9484 single processor is shared, allowing a spy process to extract
9485 information. For detailed background information, see
9486 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9487 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9488 and Necessary Software Countermeasures"). The core of the change
9489 are new versions BN_div_no_branch() and
9490 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9491 respectively, which are slower, but avoid the security-relevant
9492 conditional branches. These are automatically called by BN_div()
9493 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9494 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9495 remove a conditional branch.
9497 BN_FLG_CONSTTIME is the new name for the previous
9498 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9499 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9500 in the exponent causes BN_mod_exp_mont() to use the alternative
9501 implementation in BN_mod_exp_mont_consttime().) The old name
9502 remains as a deprecated alias.
9504 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9505 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9506 constant-time implementations for more than just exponentiation.
9507 Here too the old name is kept as a deprecated alias.
9509 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9510 the BN_BLINDING structure gets an independent copy of the
9511 modulus. This means that the previous `BIGNUM *m` argument to
9512 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9513 essentially becomes `const BIGNUM *m`, although we can't actually
9514 change this in the header file before 0.9.9. It allows
9515 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9516 enable BN_FLG_CONSTTIME.
9518 *Matthew D Wood (Intel Corp)*
9520 * In the SSL/TLS server implementation, be strict about session ID
9521 context matching (which matters if an application uses a single
9522 external cache for different purposes). Previously,
9523 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9524 set. This did ensure strict client verification, but meant that,
9525 with applications using a single external cache for quite
9526 different requirements, clients could circumvent ciphersuite
9527 restrictions for a given session ID context by starting a session
9528 in a different context.
9532 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9533 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9534 authentication-only ciphersuites.
9538 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9539 not complete and could lead to a possible single byte overflow
9540 ([CVE-2007-5135]) [Ben Laurie]
9542 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9544 * Since AES128 and AES256 (and similarly Camellia128 and
9545 Camellia256) share a single mask bit in the logic of
9546 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9547 kludge to work properly if AES128 is available and AES256 isn't
9548 (or if Camellia128 is available and Camellia256 isn't).
9552 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9553 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9554 When a point or a seed is encoded in a BIT STRING, we need to
9555 prevent the removal of trailing zero bits to get the proper DER
9556 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9557 of a NamedBitList, for which trailing 0 bits need to be removed.)
9561 * Have SSL/TLS server implementation tolerate "mismatched" record
9562 protocol version while receiving ClientHello even if the
9563 ClientHello is fragmented. (The server can't insist on the
9564 particular protocol version it has chosen before the ServerHello
9565 message has informed the client about his choice.)
9569 * Add RFC 3779 support.
9571 *Rob Austein for ARIN, Ben Laurie*
9573 * Load error codes if they are not already present instead of using a
9574 static variable. This allows them to be cleanly unloaded and reloaded.
9575 Improve header file function name parsing.
9579 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9580 or CAPABILITY handshake as required by RFCs.
9584 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9586 * Introduce limits to prevent malicious keys being able to
9587 cause a denial of service. ([CVE-2006-2940])
9589 *Steve Henson, Bodo Moeller*
9591 * Fix ASN.1 parsing of certain invalid structures that can result
9592 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9594 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9595 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9597 * Fix SSL client code which could crash if connecting to a
9598 malicious SSLv2 server. ([CVE-2006-4343])
9600 *Tavis Ormandy and Will Drewry, Google Security Team*
9602 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9603 match only those. Before that, "AES256-SHA" would be interpreted
9604 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9605 the same strength classification in 0.9.7h) as we currently only
9606 have a single AES bit in the ciphersuite description bitmap.
9607 That change, however, also applied to ciphersuite strings such as
9608 "RC4-MD5" that intentionally matched multiple ciphersuites --
9609 namely, SSL 2.0 ciphersuites in addition to the more common ones
9610 from SSL 3.0/TLS 1.0.
9612 So we change the selection algorithm again: Naming an explicit
9613 ciphersuite selects this one ciphersuite, and any other similar
9614 ciphersuite (same bitmap) from *other* protocol versions.
9615 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9616 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9618 Since SSL 2.0 does not have any ciphersuites for which the
9619 128/256 bit distinction would be relevant, this works for now.
9620 The proper fix will be to use different bits for AES128 and
9621 AES256, which would have avoided the problems from the beginning;
9622 however, bits are scarce, so we can only do this in a new release
9623 (not just a patchlevel) when we can change the SSL_CIPHER
9624 definition to split the single 'unsigned long mask' bitmap into
9625 multiple values to extend the available space.
9629 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9631 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9632 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9634 * Add AES IGE and biIGE modes.
9638 * Change the Unix randomness entropy gathering to use poll() when
9639 possible instead of select(), since the latter has some
9640 undesirable limitations.
9642 *Darryl Miles via Richard Levitte and Bodo Moeller*
9644 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9645 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9646 cannot be implicitly activated as part of, e.g., the "AES" alias.
9647 However, please upgrade to OpenSSL 0.9.9[-dev] for
9648 non-experimental use of the ECC ciphersuites to get TLS extension
9649 support, which is required for curve and point format negotiation
9650 to avoid potential handshake problems.
9654 * Disable rogue ciphersuites:
9656 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9657 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9658 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9660 The latter two were purportedly from
9661 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9664 Also deactivate the remaining ciphersuites from
9665 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9666 unofficial, and the ID has long expired.
9670 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9671 dual-core machines) and other potential thread-safety issues.
9675 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9676 versions), which is now available for royalty-free use
9677 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9678 Also, add Camellia TLS ciphersuites from RFC 4132.
9680 To minimize changes between patchlevels in the OpenSSL 0.9.8
9681 series, Camellia remains excluded from compilation unless OpenSSL
9682 is configured with 'enable-camellia'.
9686 * Disable the padding bug check when compression is in use. The padding
9687 bug check assumes the first packet is of even length, this is not
9688 necessarily true if compression is enabled and can result in false
9689 positives causing handshake failure. The actual bug test is ancient
9690 code so it is hoped that implementations will either have fixed it by
9691 now or any which still have the bug do not support compression.
9695 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9697 * When applying a cipher rule check to see if string match is an explicit
9698 cipher suite and only match that one cipher suite if it is.
9702 * Link in manifests for VC++ if needed.
9704 *Austin Ziegler <halostatue@gmail.com>*
9706 * Update support for ECC-based TLS ciphersuites according to
9707 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9708 TLS extensions, which are supported starting with the 0.9.9
9709 branch, not in the OpenSSL 0.9.8 branch).
9713 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9714 opaque EVP_CIPHER_CTX handling.
9718 * Fixes and enhancements to zlib compression code. We now only use
9719 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9720 to conform with the standards mentioned here:
9721 <http://www.zlib.net/DLL_FAQ.txt>
9722 Static zlib linking now works on Windows and the new --with-zlib-include
9723 --with-zlib-lib options to Configure can be used to supply the location
9724 of the headers and library. Gracefully handle case where zlib library
9729 * Several fixes and enhancements to the OID generation code. The old code
9730 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9731 handle numbers larger than ULONG_MAX, truncated printing and had a
9732 non standard OBJ_obj2txt() behaviour.
9736 * Add support for building of engines under engine/ as shared libraries
9737 under VC++ build system.
9741 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9742 Hopefully, we will not see any false combination of paths any more.
9746 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9748 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9749 (part of SSL_OP_ALL). This option used to disable the
9750 countermeasure against man-in-the-middle protocol-version
9751 rollback in the SSL 2.0 server implementation, which is a bad
9752 idea. ([CVE-2005-2969])
9754 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9755 for Information Security, National Institute of Advanced Industrial
9756 Science and Technology [AIST], Japan)*
9758 * Add two function to clear and return the verify parameter flags.
9762 * Keep cipherlists sorted in the source instead of sorting them at
9763 runtime, thus removing the need for a lock.
9767 * Avoid some small subgroup attacks in Diffie-Hellman.
9769 *Nick Mathewson and Ben Laurie*
9771 * Add functions for well-known primes.
9775 * Extended Windows CE support.
9777 *Satoshi Nakamura and Andy Polyakov*
9779 * Initialize SSL_METHOD structures at compile time instead of during
9780 runtime, thus removing the need for a lock.
9784 * Make PKCS7_decrypt() work even if no certificate is supplied by
9785 attempting to decrypt each encrypted key in turn. Add support to
9790 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9792 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9795 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9799 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9800 key into the same file any more.
9804 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9808 * Add -utf8 command line and config file option to 'ca'.
9810 *Stefan <stf@udoma.org*
9812 * Removed the macro des_crypt(), as it seems to conflict with some
9813 libraries. Use DES_crypt().
9817 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9818 involves renaming the source and generated shared-libs for
9819 both. The engines will accept the corrected or legacy ids
9820 ('ncipher' and '4758_cca' respectively) when binding. NB,
9821 this only applies when building 'shared'.
9823 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9825 * Add attribute functions to EVP_PKEY structure. Modify
9826 PKCS12_create() to recognize a CSP name attribute and
9827 use it. Make -CSP option work again in pkcs12 utility.
9831 * Add new functionality to the bn blinding code:
9832 - automatic re-creation of the BN_BLINDING parameters after
9833 a fixed number of uses (currently 32)
9834 - add new function for parameter creation
9835 - introduce flags to control the update behaviour of the
9836 BN_BLINDING parameters
9837 - hide BN_BLINDING structure
9838 Add a second BN_BLINDING slot to the RSA structure to improve
9839 performance when a single RSA object is shared among several
9844 * Add support for DTLS.
9846 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9848 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9849 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9853 * Remove buggy and incomplete DH cert support from
9854 ssl/ssl_rsa.c and ssl/s3_both.c
9858 * Use SHA-1 instead of MD5 as the default digest algorithm for
9859 the apps/openssl applications.
9863 * Compile clean with "-Wall -Wmissing-prototypes
9864 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9865 DEBUG_SAFESTACK must also be set.
9869 * Change ./Configure so that certain algorithms can be disabled by default.
9870 The new counterpiece to "no-xxx" is "enable-xxx".
9872 The patented RC5 and MDC2 algorithms will now be disabled unless
9873 "enable-rc5" and "enable-mdc2", respectively, are specified.
9875 (IDEA remains enabled despite being patented. This is because IDEA
9876 is frequently required for interoperability, and there is no license
9877 fee for non-commercial use. As before, "no-idea" can be used to
9878 avoid this algorithm.)
9882 * Add processing of proxy certificates (see RFC 3820). This work was
9883 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9884 EGEE (Enabling Grids for E-science in Europe).
9888 * RC4 performance overhaul on modern architectures/implementations, such
9889 as Intel P4, IA-64 and AMD64.
9893 * New utility extract-section.pl. This can be used specify an alternative
9894 section number in a pod file instead of having to treat each file as
9895 a separate case in Makefile. This can be done by adding two lines to the
9898 =for comment openssl_section:XXX
9900 The blank line is mandatory.
9904 * New arguments -certform, -keyform and -pass for s_client and s_server
9905 to allow alternative format key and certificate files and passphrase
9910 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9911 update associated structures and add various utility functions.
9913 Add new policy related verify parameters, include policy checking in
9914 standard verify code. Enhance 'smime' application with extra parameters
9915 to support policy checking and print out.
9919 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9920 Nehemiah processors. These extensions support AES encryption in hardware
9921 as well as RNG (though RNG support is currently disabled).
9923 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9925 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9929 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9931 *Andy Polyakov and a number of other people*
9933 * Improved PowerPC platform support. Most notably BIGNUM assembler
9934 implementation contributed by IBM.
9936 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9938 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9939 exponent rather than 'unsigned long'. There is a corresponding change to
9940 the new 'rsa_keygen' element of the RSA_METHOD structure.
9942 *Jelte Jansen, Geoff Thorpe*
9944 * Functionality for creating the initial serial number file is now
9945 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9947 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9948 number file to 1, which is bound to cause problems. To avoid
9949 the problems while respecting compatibility between different 0.9.7
9950 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9951 CA.pl for serial number initialization. With the new release 0.9.8,
9952 we can fix the problem directly in the 'ca' utility.)
9956 * Reduced header interdependencies by declaring more opaque objects in
9957 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9958 give fewer recursive includes, which could break lazy source code - so
9959 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9960 developers should define this symbol when building and using openssl to
9961 ensure they track the recommended behaviour, interfaces, [etc], but
9962 backwards-compatible behaviour prevails when this isn't defined.
9966 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9970 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9971 This will generate a random key of the appropriate length based on the
9972 cipher context. The EVP_CIPHER can provide its own random key generation
9973 routine to support keys of a specific form. This is used in the des and
9974 3des routines to generate a key of the correct parity. Update S/MIME
9975 code to use new functions and hence generate correct parity DES keys.
9976 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9977 valid (weak or incorrect parity).
9981 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9982 as looking them up. This is useful when the verified structure may contain
9983 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9984 present unless the new PKCS7_NO_CRL flag is asserted.
9988 * Extend ASN1 oid configuration module. It now additionally accepts the
9991 shortName = some long name, 1.2.3.4
9995 * Reimplemented the BN_CTX implementation. There is now no more static
9996 limitation on the number of variables it can handle nor the depth of the
9997 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9998 information can now expand as required, and rather than having a single
9999 static array of bignums, BN_CTX now uses a linked-list of such arrays
10000 allowing it to expand on demand whilst maintaining the usefulness of
10001 BN_CTX's "bundling".
10005 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10006 to allow all RSA operations to function using a single BN_CTX.
10010 * Preliminary support for certificate policy evaluation and checking. This
10011 is initially intended to pass the tests outlined in "Conformance Testing
10012 of Relying Party Client Certificate Path Processing Logic" v1.07.
10016 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10017 remained unused and not that useful. A variety of other little bignum
10018 tweaks and fixes have also been made continuing on from the audit (see
10023 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10024 associated ASN1, EVP and SSL functions and old ASN1 macros.
10028 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10029 and this should never fail. So the return value from the use of
10030 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10031 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10035 * BN_CTX_get() should return zero-valued bignums, providing the same
10036 initialised value as BN_new().
10038 *Geoff Thorpe, suggested by Ulf Möller*
10040 * Support for inhibitAnyPolicy certificate extension.
10044 * An audit of the BIGNUM code is underway, for which debugging code is
10045 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10046 is considered valid when processing BIGNUMs, and causes execution to
10047 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10048 further steps are taken to deliberately pollute unused data in BIGNUM
10049 structures to try and expose faulty code further on. For now, openssl will
10050 (in its default mode of operation) continue to tolerate the inconsistent
10051 forms that it has tolerated in the past, but authors and packagers should
10052 consider trying openssl and their own applications when compiled with
10053 these debugging symbols defined. It will help highlight potential bugs in
10054 their own code, and will improve the test coverage for OpenSSL itself. At
10055 some point, these tighter rules will become openssl's default to improve
10056 maintainability, though the assert()s and other overheads will remain only
10057 in debugging configurations. See bn.h for more details.
10059 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10061 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10062 that can only be obtained through BN_CTX_new() (which implicitly
10063 initialises it). The presence of this function only made it possible
10064 to overwrite an existing structure (and cause memory leaks).
10068 * Because of the callback-based approach for implementing LHASH as a
10069 template type, lh_insert() adds opaque objects to hash-tables and
10070 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10071 to clean up those corresponding objects before destroying the hash table
10072 (and losing the object pointers). So some over-zealous constifications in
10073 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10074 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10075 prototyped to have "const" restrictions on the object pointers they are
10076 given (and so aren't required to cast them away any more).
10080 * The tmdiff.h API was so ugly and minimal that our own timing utility
10081 (speed) prefers to use its own implementation. The two implementations
10082 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10083 its object type properly exposed (MS_TM) instead of casting to/from
10084 `char *`. This may still change yet if someone realises MS_TM and
10086 aren't necessarily the greatest nomenclatures - but this is what was used
10087 internally to the implementation so I've used that for now.
10091 * Ensure that deprecated functions do not get compiled when
10092 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10093 the self-tests were still using deprecated key-generation functions so
10094 these have been updated also.
10098 * Reorganise PKCS#7 code to separate the digest location functionality
10099 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10100 New function PKCS7_set_digest() to set the digest type for PKCS#7
10101 digestedData type. Add additional code to correctly generate the
10102 digestedData type and add support for this type in PKCS7 initialization
10107 * New function PKCS7_set0_type_other() this initializes a PKCS7
10108 structure of type "other".
10112 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10113 sure the loop does correctly stop and breaking ("division by zero")
10114 modulus operations are not performed. The (pre-generated) prime
10115 table crypto/bn/bn_prime.h was already correct, but it could not be
10116 re-generated on some platforms because of the "division by zero"
10117 situation in the script.
10119 *Ralf S. Engelschall*
10121 * Update support for ECC-based TLS ciphersuites according to
10122 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10123 SHA-1 now is only used for "small" curves (where the
10124 representation of a field element takes up to 24 bytes); for
10125 larger curves, the field element resulting from ECDH is directly
10126 used as premaster secret.
10128 *Douglas Stebila (Sun Microsystems Laboratories)*
10130 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10131 curve secp160r1 to the tests.
10133 *Douglas Stebila (Sun Microsystems Laboratories)*
10135 * Add the possibility to load symbols globally with DSO.
10137 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10139 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10140 control of the error stack.
10144 * Add support for STORE in ENGINE.
10148 * Add the STORE type. The intention is to provide a common interface
10149 to certificate and key stores, be they simple file-based stores, or
10150 HSM-type store, or LDAP stores, or...
10151 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10155 * Add a generic structure called OPENSSL_ITEM. This can be used to
10156 pass a list of arguments to any function as well as provide a way
10157 for a function to pass data back to the caller.
10161 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10162 works like BUF_strdup() but can be used to duplicate a portion of
10163 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10168 * Add the function sk_find_ex() which works like sk_find(), but will
10169 return an index to an element even if an exact match couldn't be
10170 found. The index is guaranteed to point at the element where the
10171 searched-for key would be inserted to preserve sorting order.
10175 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10176 takes an extra flags argument for optional functionality. Currently,
10177 the following flags are defined:
10179 OBJ_BSEARCH_VALUE_ON_NOMATCH
10180 This one gets OBJ_bsearch_ex() to return a pointer to the first
10181 element where the comparing function returns a negative or zero
10184 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10185 This one gets OBJ_bsearch_ex() to return a pointer to the first
10186 element where the comparing function returns zero. This is useful
10187 if there are more than one element where the comparing function
10192 * Make it possible to create self-signed certificates with 'openssl ca'
10193 in such a way that the self-signed certificate becomes part of the
10194 CA database and uses the same mechanisms for serial number generation
10195 as all other certificate signing. The new flag '-selfsign' enables
10196 this functionality. Adapt CA.sh and CA.pl.in.
10200 * Add functionality to check the public key of a certificate request
10201 against a given private. This is useful to check that a certificate
10202 request can be signed by that key (self-signing).
10206 * Make it possible to have multiple active certificates with the same
10207 subject in the CA index file. This is done only if the keyword
10208 'unique_subject' is set to 'no' in the main CA section (default
10209 if 'CA_default') of the configuration file. The value is saved
10210 with the database itself in a separate index attribute file,
10211 named like the index file with '.attr' appended to the name.
10215 * Generate multi-valued AVAs using '+' notation in config files for
10220 * Support for nameConstraints certificate extension.
10224 * Support for policyConstraints certificate extension.
10228 * Support for policyMappings certificate extension.
10232 * Make sure the default DSA_METHOD implementation only uses its
10233 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10234 and change its own handlers to be NULL so as to remove unnecessary
10235 indirection. This lets alternative implementations fallback to the
10236 default implementation more easily.
10240 * Support for directoryName in GeneralName related extensions
10245 * Make it possible to link applications using Makefile.shared.
10246 Make that possible even when linking against static libraries!
10250 * Support for single pass processing for S/MIME signing. This now
10251 means that S/MIME signing can be done from a pipe, in addition
10252 cleartext signing (multipart/signed type) is effectively streaming
10253 and the signed data does not need to be all held in memory.
10255 This is done with a new flag PKCS7_STREAM. When this flag is set
10256 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10257 is done after the data is output (and digests calculated) in
10258 SMIME_write_PKCS7().
10262 * Add full support for -rpath/-R, both in shared libraries and
10263 applications, at least on the platforms where it's known how
10268 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10269 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10270 will now compute a table of multiples of the generator that
10271 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10272 faster (notably in the case of a single point multiplication,
10273 scalar * generator).
10275 *Nils Larsch, Bodo Moeller*
10277 * IPv6 support for certificate extensions. The various extensions
10278 which use the IP:a.b.c.d can now take IPv6 addresses using the
10279 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10284 * Added an ENGINE that implements RSA by performing private key
10285 exponentiations with the GMP library. The conversions to and from
10286 GMP's mpz_t format aren't optimised nor are any montgomery forms
10287 cached, and on x86 it appears OpenSSL's own performance has caught up.
10288 However there are likely to be other architectures where GMP could
10289 provide a boost. This ENGINE is not built in by default, but it can be
10290 specified at Configure time and should be accompanied by the necessary
10291 linker additions, eg;
10292 ./config -DOPENSSL_USE_GMP -lgmp
10296 * "openssl engine" will not display ENGINE/DSO load failure errors when
10297 testing availability of engines with "-t" - the old behaviour is
10298 produced by increasing the feature's verbosity with "-tt".
10302 * ECDSA routines: under certain error conditions uninitialized BN objects
10303 could be freed. Solution: make sure initialization is performed early
10304 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10309 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10310 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10311 software implementations. For DSA and DH, parameter generation can
10312 also be overridden by providing the appropriate method callbacks.
10316 * Change the "progress" mechanism used in key-generation and
10317 primality testing to functions that take a new BN_GENCB pointer in
10318 place of callback/argument pairs. The new API functions have `_ex`
10319 postfixes and the older functions are reimplemented as wrappers for
10320 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10321 declarations of the old functions to help (graceful) attempts to
10322 migrate to the new functions. Also, the new key-generation API
10323 functions operate on a caller-supplied key-structure and return
10324 success/failure rather than returning a key or NULL - this is to
10325 help make "keygen" another member function of RSA_METHOD etc.
10327 Example for using the new callback interface:
10329 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10330 void *my_arg = ...;
10333 BN_GENCB_set(&my_cb, my_callback, my_arg);
10335 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10336 /* For the meaning of a, b in calls to my_callback(), see the
10337 * documentation of the function that calls the callback.
10338 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10339 * my_callback should return 1 if it wants BN_is_prime_ex()
10340 * to continue, or 0 to stop.
10345 * Change the ZLIB compression method to be stateful, and make it
10346 available to TLS with the number defined in
10347 draft-ietf-tls-compression-04.txt.
10351 * Add the ASN.1 structures and functions for CertificatePair, which
10352 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10354 CertificatePair ::= SEQUENCE {
10355 forward [0] Certificate OPTIONAL,
10356 reverse [1] Certificate OPTIONAL,
10357 -- at least one of the pair shall be present -- }
10359 Also implement the PEM functions to read and write certificate
10360 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10362 This needed to be defined, mostly for the sake of the LDAP
10363 attribute crossCertificatePair, but may prove useful elsewhere as
10368 * Make it possible to inhibit symlinking of shared libraries in
10369 Makefile.shared, for Cygwin's sake.
10373 * Extend the BIGNUM API by creating a function
10374 void BN_set_negative(BIGNUM *a, int neg);
10375 and a macro that behave like
10376 int BN_is_negative(const BIGNUM *a);
10378 to avoid the need to access 'a->neg' directly in applications.
10382 * Implement fast modular reduction for pseudo-Mersenne primes
10383 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10384 EC_GROUP_new_curve_GFp() will now automatically use this
10387 *Nils Larsch <nla@trustcenter.de>*
10389 * Add new lock type (CRYPTO_LOCK_BN).
10393 * Change the ENGINE framework to automatically load engines
10394 dynamically from specific directories unless they could be
10395 found to already be built in or loaded. Move all the
10396 current engines except for the cryptodev one to a new
10397 directory engines/.
10398 The engines in engines/ are built as shared libraries if
10399 the "shared" options was given to ./Configure or ./config.
10400 Otherwise, they are inserted in libcrypto.a.
10401 /usr/local/ssl/engines is the default directory for dynamic
10402 engines, but that can be overridden at configure time through
10403 the usual use of --prefix and/or --openssldir, and at run
10404 time with the environment variable OPENSSL_ENGINES.
10406 *Geoff Thorpe and Richard Levitte*
10408 * Add Makefile.shared, a helper makefile to build shared
10409 libraries. Adapt Makefile.org.
10413 * Add version info to Win32 DLLs.
10415 *Peter 'Luna' Runestig" <peter@runestig.com>*
10417 * Add new 'medium level' PKCS#12 API. Certificates and keys
10418 can be added using this API to created arbitrary PKCS#12
10419 files while avoiding the low level API.
10421 New options to PKCS12_create(), key or cert can be NULL and
10422 will then be omitted from the output file. The encryption
10423 algorithm NIDs can be set to -1 for no encryption, the mac
10424 iteration count can be set to 0 to omit the mac.
10426 Enhance pkcs12 utility by making the -nokeys and -nocerts
10427 options work when creating a PKCS#12 file. New option -nomac
10428 to omit the mac, NONE can be set for an encryption algorithm.
10429 New code is modified to use the enhanced PKCS12_create()
10430 instead of the low level API.
10434 * Extend ASN1 encoder to support indefinite length constructed
10435 encoding. This can output sequences tags and octet strings in
10436 this form. Modify pk7_asn1.c to support indefinite length
10437 encoding. This is experimental and needs additional code to
10438 be useful, such as an ASN1 bio and some enhanced streaming
10441 Extend template encode functionality so that tagging is passed
10442 down to the template encoder.
10446 * Let 'openssl req' fail if an argument to '-newkey' is not
10447 recognized instead of using RSA as a default.
10451 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10452 As these are not official, they are not included in "ALL";
10453 the "ECCdraft" ciphersuite group alias can be used to select them.
10455 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10457 * Add ECDH engine support.
10459 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10461 * Add ECDH in new directory crypto/ecdh/.
10463 *Douglas Stebila (Sun Microsystems Laboratories)*
10465 * Let BN_rand_range() abort with an error after 100 iterations
10466 without success (which indicates a broken PRNG).
10470 * Change BN_mod_sqrt() so that it verifies that the input value
10471 is really the square of the return value. (Previously,
10472 BN_mod_sqrt would show GIGO behaviour.)
10476 * Add named elliptic curves over binary fields from X9.62, SECG,
10477 and WAP/WTLS; add OIDs that were still missing.
10479 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10481 * Extend the EC library for elliptic curves over binary fields
10482 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10485 EC_GF2m_simple_method
10489 EC_GROUP_new_curve_GF2m
10490 EC_GROUP_set_curve_GF2m
10491 EC_GROUP_get_curve_GF2m
10492 EC_POINT_set_affine_coordinates_GF2m
10493 EC_POINT_get_affine_coordinates_GF2m
10494 EC_POINT_set_compressed_coordinates_GF2m
10496 Point compression for binary fields is disabled by default for
10497 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10500 As binary polynomials are represented as BIGNUMs, various members
10501 of the EC_GROUP and EC_POINT data structures can be shared
10502 between the implementations for prime fields and binary fields;
10503 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10504 are essentially identical to their `..._GFp` counterparts.
10505 (For simplicity, the `..._GFp` prefix has been dropped from
10506 various internal method names.)
10508 An internal 'field_div' method (similar to 'field_mul' and
10509 'field_sqr') has been added; this is used only for binary fields.
10511 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10513 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10514 through methods ('mul', 'precompute_mult').
10516 The generic implementations (now internally called 'ec_wNAF_mul'
10517 and 'ec_wNAF_precomputed_mult') remain the default if these
10518 methods are undefined.
10520 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10522 * New function EC_GROUP_get_degree, which is defined through
10523 EC_METHOD. For curves over prime fields, this returns the bit
10524 length of the modulus.
10526 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10528 * New functions EC_GROUP_dup, EC_POINT_dup.
10529 (These simply call ..._new and ..._copy).
10531 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10533 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10534 Polynomials are represented as BIGNUMs (where the sign bit is not
10535 used) in the following functions [macros]:
10538 BN_GF2m_sub [= BN_GF2m_add]
10539 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10540 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10541 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10543 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10544 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10545 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10546 BN_GF2m_cmp [= BN_ucmp]
10548 (Note that only the 'mod' functions are actually for fields GF(2^m).
10549 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10551 For some functions, an the irreducible polynomial defining a
10552 field can be given as an 'unsigned int[]' with strictly
10553 decreasing elements giving the indices of those bits that are set;
10554 i.e., p[] represents the polynomial
10555 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10557 p[0] > p[1] > ... > p[k] = 0.
10558 This applies to the following functions:
10561 BN_GF2m_mod_mul_arr
10562 BN_GF2m_mod_sqr_arr
10563 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10564 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10565 BN_GF2m_mod_exp_arr
10566 BN_GF2m_mod_sqrt_arr
10567 BN_GF2m_mod_solve_quad_arr
10571 Conversion can be performed by the following functions:
10576 bntest.c has additional tests for binary polynomial arithmetic.
10578 Two implementations for BN_GF2m_mod_div() are available.
10579 The default algorithm simply uses BN_GF2m_mod_inv() and
10580 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10581 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10582 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10584 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10586 * Add new error code 'ERR_R_DISABLED' that can be used when some
10587 functionality is disabled at compile-time.
10589 *Douglas Stebila <douglas.stebila@sun.com>*
10591 * Change default behaviour of 'openssl asn1parse' so that more
10592 information is visible when viewing, e.g., a certificate:
10594 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10595 mode the content of non-printable OCTET STRINGs is output in a
10596 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10597 avoid the appearance of a printable string.
10599 *Nils Larsch <nla@trustcenter.de>*
10601 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10603 EC_GROUP_set_asn1_flag()
10604 EC_GROUP_get_asn1_flag()
10605 EC_GROUP_set_point_conversion_form()
10606 EC_GROUP_get_point_conversion_form()
10607 These control ASN1 encoding details:
10608 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10609 has been set to OPENSSL_EC_NAMED_CURVE.
10610 - Points are encoded in uncompressed form by default; options for
10611 asn1_for are as for point2oct, namely
10612 POINT_CONVERSION_COMPRESSED
10613 POINT_CONVERSION_UNCOMPRESSED
10614 POINT_CONVERSION_HYBRID
10616 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10618 EC_GROUP_set_seed()
10619 EC_GROUP_get0_seed()
10620 EC_GROUP_get_seed_len()
10621 This is used only for ASN1 purposes (so far).
10623 *Nils Larsch <nla@trustcenter.de>*
10625 * Add 'field_type' member to EC_METHOD, which holds the NID
10626 of the appropriate field type OID. The new function
10627 EC_METHOD_get_field_type() returns this value.
10629 *Nils Larsch <nla@trustcenter.de>*
10632 EC_POINT_point2bn()
10633 EC_POINT_bn2point()
10634 EC_POINT_point2hex()
10635 EC_POINT_hex2point()
10636 providing useful interfaces to EC_POINT_point2oct() and
10637 EC_POINT_oct2point().
10639 *Nils Larsch <nla@trustcenter.de>*
10641 * Change internals of the EC library so that the functions
10642 EC_GROUP_set_generator()
10643 EC_GROUP_get_generator()
10644 EC_GROUP_get_order()
10645 EC_GROUP_get_cofactor()
10646 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10647 to methods, which would lead to unnecessary code duplication when
10648 adding different types of curves.
10650 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10652 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10653 arithmetic, and such that modified wNAFs are generated
10654 (which avoid length expansion in many cases).
10658 * Add a function EC_GROUP_check_discriminant() (defined via
10659 EC_METHOD) that verifies that the curve discriminant is non-zero.
10661 Add a function EC_GROUP_check() that makes some sanity tests
10662 on a EC_GROUP, its generator and order. This includes
10663 EC_GROUP_check_discriminant().
10665 *Nils Larsch <nla@trustcenter.de>*
10667 * Add ECDSA in new directory crypto/ecdsa/.
10669 Add applications 'openssl ecparam' and 'openssl ecdsa'
10670 (these are based on 'openssl dsaparam' and 'openssl dsa').
10672 ECDSA support is also included in various other files across the
10673 library. Most notably,
10674 - 'openssl req' now has a '-newkey ecdsa:file' option;
10675 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10676 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10677 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10678 them suitable for ECDSA where domain parameters must be
10679 extracted before the specific public key;
10680 - ECDSA engine support has been added.
10682 *Nils Larsch <nla@trustcenter.de>*
10684 * Include some named elliptic curves, and add OIDs from X9.62,
10685 SECG, and WAP/WTLS. Each curve can be obtained from the new
10687 EC_GROUP_new_by_curve_name(),
10688 and the list of available named curves can be obtained with
10689 EC_get_builtin_curves().
10690 Also add a 'curve_name' member to EC_GROUP objects, which can be
10692 EC_GROUP_set_curve_name()
10693 EC_GROUP_get_curve_name()
10695 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10697 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10698 was actually never needed) and in BN_mul(). The removal in BN_mul()
10699 required a small change in bn_mul_part_recursive() and the addition
10700 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10701 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10702 bn_sub_words() and bn_add_words() except they take arrays with
10707 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10709 * Cleanse PEM buffers before freeing them since they may contain
10712 *Benjamin Bennett <ben@psc.edu>*
10714 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10715 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10716 authentication-only ciphersuites.
10720 * Since AES128 and AES256 share a single mask bit in the logic of
10721 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10722 kludge to work properly if AES128 is available and AES256 isn't.
10726 * Expand security boundary to match 1.1.1 module.
10730 * Remove redundant features: hash file source, editing of test vectors
10731 modify fipsld to use external fips_premain.c signature.
10735 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10736 run algorithm test programs.
10740 * Make algorithm test programs more tolerant of whitespace.
10744 * Have SSL/TLS server implementation tolerate "mismatched" record
10745 protocol version while receiving ClientHello even if the
10746 ClientHello is fragmented. (The server can't insist on the
10747 particular protocol version it has chosen before the ServerHello
10748 message has informed the client about his choice.)
10752 * Load error codes if they are not already present instead of using a
10753 static variable. This allows them to be cleanly unloaded and reloaded.
10757 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10759 * Introduce limits to prevent malicious keys being able to
10760 cause a denial of service. ([CVE-2006-2940])
10762 *Steve Henson, Bodo Moeller*
10764 * Fix ASN.1 parsing of certain invalid structures that can result
10765 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10767 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10768 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10770 * Fix SSL client code which could crash if connecting to a
10771 malicious SSLv2 server. ([CVE-2006-4343])
10773 *Tavis Ormandy and Will Drewry, Google Security Team*
10775 * Change ciphersuite string processing so that an explicit
10776 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10777 will no longer include "AES128-SHA"), and any other similar
10778 ciphersuite (same bitmap) from *other* protocol versions (so that
10779 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10780 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10781 changes from 0.9.8b and 0.9.8d.
10785 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10787 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10788 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10790 * Change the Unix randomness entropy gathering to use poll() when
10791 possible instead of select(), since the latter has some
10792 undesirable limitations.
10794 *Darryl Miles via Richard Levitte and Bodo Moeller*
10796 * Disable rogue ciphersuites:
10798 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10799 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10800 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10802 The latter two were purportedly from
10803 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10806 Also deactivate the remaining ciphersuites from
10807 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10808 unofficial, and the ID has long expired.
10812 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10813 dual-core machines) and other potential thread-safety issues.
10817 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10819 * Adapt fipsld and the build system to link against the validated FIPS
10820 module in FIPS mode.
10824 * Fixes for VC++ 2005 build under Windows.
10828 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10829 from a Windows bash shell such as MSYS. It is autodetected from the
10830 "config" script when run from a VC++ environment. Modify standard VC++
10831 build to use fipscanister.o from the GNU make build.
10835 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10837 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10838 The value now differs depending on if you build for FIPS or not.
10839 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10840 safely run with a non-FIPSed libcrypto, as it may crash because of
10841 the difference induced by this change.
10845 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10847 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10848 (part of SSL_OP_ALL). This option used to disable the
10849 countermeasure against man-in-the-middle protocol-version
10850 rollback in the SSL 2.0 server implementation, which is a bad
10851 idea. ([CVE-2005-2969])
10853 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10854 for Information Security, National Institute of Advanced Industrial
10855 Science and Technology [AIST, Japan)]*
10857 * Minimal support for X9.31 signatures and PSS padding modes. This is
10858 mainly for FIPS compliance and not fully integrated at this stage.
10862 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10863 the exponentiation using a fixed-length exponent. (Otherwise,
10864 the information leaked through timing could expose the secret key
10865 after many signatures; cf. Bleichenbacher's attack on DSA with
10870 * Make a new fixed-window mod_exp implementation the default for
10871 RSA, DSA, and DH private-key operations so that the sequence of
10872 squares and multiplies and the memory access pattern are
10873 independent of the particular secret key. This will mitigate
10874 cache-timing and potential related attacks.
10876 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10877 and this is automatically used by BN_mod_exp_mont() if the new flag
10878 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10879 will use this BN flag for private exponents unless the flag
10880 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10881 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10883 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10885 * Change the client implementation for SSLv23_method() and
10886 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10887 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10888 (Previously, the SSL 2.0 backwards compatible Client Hello
10889 message format would be used even with SSL_OP_NO_SSLv2.)
10893 * Add support for smime-type MIME parameter in S/MIME messages which some
10898 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10899 a threadsafe manner. Modify rsa code to use new function and add calls
10900 to dsa and dh code (which had race conditions before).
10904 * Include the fixed error library code in the C error file definitions
10905 instead of fixing them up at runtime. This keeps the error code
10906 structures constant.
10910 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10912 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10915 * Fixes for newer kerberos headers. NB: the casts are needed because
10916 the 'length' field is signed on one version and unsigned on another
10917 with no (?) obvious way to tell the difference, without these VC++
10918 complains. Also the "definition" of FAR (blank) is no longer included
10919 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10920 some needed definitions.
10924 * Undo Cygwin change.
10928 * Added support for proxy certificates according to RFC 3820.
10929 Because they may be a security thread to unaware applications,
10930 they must be explicitly allowed in run-time. See
10931 docs/HOWTO/proxy_certificates.txt for further information.
10935 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10937 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10938 server and client random values. Previously
10939 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10940 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10942 This change has negligible security impact because:
10944 1. Server and client random values still have 24 bytes of pseudo random
10947 2. Server and client random values are sent in the clear in the initial
10950 3. The master secret is derived using the premaster secret (48 bytes in
10951 size for static RSA ciphersuites) as well as client server and random
10954 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10957 *Stephen Henson, reported by UK NISCC*
10959 * Use Windows randomness collection on Cygwin.
10963 * Fix hang in EGD/PRNGD query when communication socket is closed
10964 prematurely by EGD/PRNGD.
10966 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10968 * Prompt for pass phrases when appropriate for PKCS12 input format.
10972 * Back-port of selected performance improvements from development
10973 branch, as well as improved support for PowerPC platforms.
10977 * Add lots of checks for memory allocation failure, error codes to indicate
10978 failure and freeing up memory if a failure occurs.
10980 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10982 * Add new -passin argument to dgst.
10986 * Perform some character comparisons of different types in X509_NAME_cmp:
10987 this is needed for some certificates that re-encode DNs into UTF8Strings
10988 (in violation of RFC3280) and can't or won't issue name rollover
10993 * Make an explicit check during certificate validation to see that
10994 the CA setting in each certificate on the chain is correct. As a
10995 side effect always do the following basic checks on extensions,
10996 not just when there's an associated purpose to the check:
10998 - if there is an unhandled critical extension (unless the user
10999 has chosen to ignore this fault)
11000 - if the path length has been exceeded (if one is set at all)
11001 - that certain extensions fit the associated purpose (if one has
11006 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11008 * Avoid a race condition when CRLs are checked in a multi threaded
11009 environment. This would happen due to the reordering of the revoked
11010 entries during signature checking and serial number lookup. Now the
11011 encoding is cached and the serial number sort performed under a lock.
11012 Add new STACK function sk_is_sorted().
11016 * Add Delta CRL to the extension code.
11020 * Various fixes to s3_pkt.c so alerts are sent properly.
11022 *David Holmes <d.holmes@f5.com>*
11024 * Reduce the chances of duplicate issuer name and serial numbers (in
11025 violation of RFC3280) using the OpenSSL certificate creation utilities.
11026 This is done by creating a random 64 bit value for the initial serial
11027 number when a serial number file is created or when a self signed
11028 certificate is created using 'openssl req -x509'. The initial serial
11029 number file is created using 'openssl x509 -next_serial' in CA.pl
11030 rather than being initialized to 1.
11034 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11036 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11037 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11039 *Joe Orton, Steve Henson*
11041 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11044 *Joe Orton, Steve Henson*
11046 * Make it possible to have multiple active certificates with the same
11047 subject in the CA index file. This is done only if the keyword
11048 'unique_subject' is set to 'no' in the main CA section (default
11049 if 'CA_default') of the configuration file. The value is saved
11050 with the database itself in a separate index attribute file,
11051 named like the index file with '.attr' appended to the name.
11055 * X509 verify fixes. Disable broken certificate workarounds when
11056 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11057 keyUsage extension present. Don't accept CRLs with unhandled critical
11058 extensions: since verify currently doesn't process CRL extensions this
11059 rejects a CRL with *any* critical extensions. Add new verify error codes
11064 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11065 A clarification of RFC2560 will require the use of OCTET STRINGs and
11066 some implementations cannot handle the current raw format. Since OpenSSL
11067 copies and compares OCSP nonces as opaque blobs without any attempt at
11068 parsing them this should not create any compatibility issues.
11072 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11073 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11074 this HMAC (and other) operations are several times slower than OpenSSL
11079 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11081 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11083 * Use the correct content when signing type "other".
11087 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11089 * Fix various bugs revealed by running the NISCC test suite:
11091 Stop out of bounds reads in the ASN1 code when presented with
11092 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11094 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11096 If verify callback ignores invalid public key errors don't try to check
11097 certificate signature with the NULL public key.
11101 * New -ignore_err option in ocsp application to stop the server
11102 exiting on the first error in a request.
11106 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11107 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11112 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11113 extra data after the compression methods not only for TLS 1.0
11114 but also for SSL 3.0 (as required by the specification).
11116 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11118 * Change X509_certificate_type() to mark the key as exported/exportable
11119 when it's 512 *bits* long, not 512 bytes.
11123 * Change AES_cbc_encrypt() so it outputs exact multiple of
11124 blocks during encryption.
11128 * Various fixes to base64 BIO and non blocking I/O. On write
11129 flushes were not handled properly if the BIO retried. On read
11130 data was not being buffered properly and had various logic bugs.
11131 This also affects blocking I/O when the data being decoded is a
11136 * Various S/MIME bugfixes and compatibility changes:
11137 output correct application/pkcs7 MIME type if
11138 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11139 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11140 of files as .eml work). Correctly handle very long lines in MIME
11145 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11147 * Countermeasure against the Klima-Pokorny-Rosa extension of
11148 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11149 a protocol version number mismatch like a decryption error
11150 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11154 * Turn on RSA blinding by default in the default implementation
11155 to avoid a timing attack. Applications that don't want it can call
11156 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11157 They would be ill-advised to do so in most cases.
11159 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11161 * Change RSA blinding code so that it works when the PRNG is not
11162 seeded (in this case, the secret RSA exponent is abused as
11163 an unpredictable seed -- if it is not unpredictable, there
11164 is no point in blinding anyway). Make RSA blinding thread-safe
11165 by remembering the creator's thread ID in rsa->blinding and
11166 having all other threads use local one-time blinding factors
11167 (this requires more computation than sharing rsa->blinding, but
11168 avoids excessive locking; and if an RSA object is not shared
11169 between threads, blinding will still be very fast).
11173 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11174 ENGINE as defaults for all supported algorithms irrespective of
11175 the 'flags' parameter. 'flags' is now honoured, so applications
11176 should make sure they are passing it correctly.
11180 * Target "mingw" now allows native Windows code to be generated in
11181 the Cygwin environment as well as with the MinGW compiler.
11185 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11187 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11188 via timing by performing a MAC computation even if incorrect
11189 block cipher padding has been found. This is a countermeasure
11190 against active attacks where the attacker has to distinguish
11191 between bad padding and a MAC verification error. ([CVE-2003-0078])
11193 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11194 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11195 Martin Vuagnoux (EPFL, Ilion)*
11197 * Make the no-err option work as intended. The intention with no-err
11198 is not to have the whole error stack handling routines removed from
11199 libcrypto, it's only intended to remove all the function name and
11200 reason texts, thereby removing some of the footprint that may not
11201 be interesting if those errors aren't displayed anyway.
11203 NOTE: it's still possible for any application or module to have its
11204 own set of error texts inserted. The routines are there, just not
11205 used by default when no-err is given.
11209 * Add support for FreeBSD on IA64.
11211 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11213 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11214 Kerberos function mit_des_cbc_cksum(). Before this change,
11215 the value returned by DES_cbc_cksum() was like the one from
11216 mit_des_cbc_cksum(), except the bytes were swapped.
11218 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11220 * Allow an application to disable the automatic SSL chain building.
11221 Before this a rather primitive chain build was always performed in
11222 ssl3_output_cert_chain(): an application had no way to send the
11223 correct chain if the automatic operation produced an incorrect result.
11225 Now the chain builder is disabled if either:
11227 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11229 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11231 The reasoning behind this is that an application would not want the
11232 auto chain building to take place if extra chain certificates are
11233 present and it might also want a means of sending no additional
11234 certificates (for example the chain has two certificates and the
11239 * Add the possibility to build without the ENGINE framework.
11241 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11243 * Under Win32 gmtime() can return NULL: check return value in
11244 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11248 * DSA routines: under certain error conditions uninitialized BN objects
11249 could be freed. Solution: make sure initialization is performed early
11250 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11251 Nils Larsch <nla@trustcenter.de> via PR#459)
11255 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11256 checked on reconnect on the client side, therefore session resumption
11257 could still fail with a "ssl session id is different" error. This
11258 behaviour is masked when SSL_OP_ALL is used due to
11259 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11260 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11261 followup to PR #377.
11265 * IA-32 assembler support enhancements: unified ELF targets, support
11266 for SCO/Caldera platforms, fix for Cygwin shared build.
11270 * Add support for FreeBSD on sparc64. As a consequence, support for
11271 FreeBSD on non-x86 processors is separate from x86 processors on
11272 the config script, much like the NetBSD support.
11274 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11276 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11278 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11281 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11282 code (06) was taken as the first octet of the session ID and the last
11283 octet was ignored consequently. As a result SSLv2 client side session
11284 caching could not have worked due to the session ID mismatch between
11286 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11291 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11292 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11297 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11298 seems that in spite of existing for more than a year, many application
11299 author have done nothing to provide the necessary callbacks, which
11300 means that this particular engine will not work properly anywhere.
11301 This is a very unfortunate situation which forces us, in the name
11302 of usability, to give the hw_ncipher.c a static lock, which is part
11304 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11305 appear in 0.9.8 or later. We EXPECT application authors to have
11306 dealt properly with this when 0.9.8 is released (unless we actually
11307 make such changes in the libcrypto locking code that changes will
11308 have to be made anyway).
11312 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11313 octets have been read, EOF or an error occurs. Without this change
11314 some truncated ASN1 structures will not produce an error.
11318 * Disable Heimdal support, since it hasn't been fully implemented.
11319 Still give the possibility to force the use of Heimdal, but with
11320 warnings and a request that patches get sent to openssl-dev.
11324 * Add the VC-CE target, introduce the WINCE sysname, and add
11325 INSTALL.WCE and appropriate conditionals to make it build.
11327 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11329 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11330 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11331 edit numbers of the version.
11333 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11335 * Introduce safe string copy and catenation functions
11336 (BUF_strlcpy() and BUF_strlcat()).
11338 *Ben Laurie (CHATS) and Richard Levitte*
11340 * Avoid using fixed-size buffers for one-line DNs.
11342 *Ben Laurie (CHATS)*
11344 * Add BUF_MEM_grow_clean() to avoid information leakage when
11345 resizing buffers containing secrets, and use where appropriate.
11347 *Ben Laurie (CHATS)*
11349 * Avoid using fixed size buffers for configuration file location.
11351 *Ben Laurie (CHATS)*
11353 * Avoid filename truncation for various CA files.
11355 *Ben Laurie (CHATS)*
11357 * Use sizeof in preference to magic numbers.
11359 *Ben Laurie (CHATS)*
11361 * Avoid filename truncation in cert requests.
11363 *Ben Laurie (CHATS)*
11365 * Add assertions to check for (supposedly impossible) buffer
11368 *Ben Laurie (CHATS)*
11370 * Don't cache truncated DNS entries in the local cache (this could
11371 potentially lead to a spoofing attack).
11373 *Ben Laurie (CHATS)*
11375 * Fix various buffers to be large enough for hex/decimal
11376 representations in a platform independent manner.
11378 *Ben Laurie (CHATS)*
11380 * Add CRYPTO_realloc_clean() to avoid information leakage when
11381 resizing buffers containing secrets, and use where appropriate.
11383 *Ben Laurie (CHATS)*
11385 * Add BIO_indent() to avoid much slightly worrying code to do
11388 *Ben Laurie (CHATS)*
11390 * Convert sprintf()/BIO_puts() to BIO_printf().
11392 *Ben Laurie (CHATS)*
11394 * buffer_gets() could terminate with the buffer only half
11397 *Ben Laurie (CHATS)*
11399 * Add assertions to prevent user-supplied crypto functions from
11400 overflowing internal buffers by having large block sizes, etc.
11402 *Ben Laurie (CHATS)*
11404 * New OPENSSL_assert() macro (similar to assert(), but enabled
11407 *Ben Laurie (CHATS)*
11409 * Eliminate unused copy of key in RC4.
11411 *Ben Laurie (CHATS)*
11413 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11415 *Ben Laurie (CHATS)*
11417 * Fix off-by-one error in EGD path.
11419 *Ben Laurie (CHATS)*
11421 * If RANDFILE path is too long, ignore instead of truncating.
11423 *Ben Laurie (CHATS)*
11425 * Eliminate unused and incorrectly sized X.509 structure
11428 *Ben Laurie (CHATS)*
11430 * Eliminate unused and dangerous function knumber().
11432 *Ben Laurie (CHATS)*
11434 * Eliminate unused and dangerous structure, KSSL_ERR.
11436 *Ben Laurie (CHATS)*
11438 * Protect against overlong session ID context length in an encoded
11439 session object. Since these are local, this does not appear to be
11442 *Ben Laurie (CHATS)*
11444 * Change from security patch (see 0.9.6e below) that did not affect
11445 the 0.9.6 release series:
11447 Remote buffer overflow in SSL3 protocol - an attacker could
11448 supply an oversized master key in Kerberos-enabled versions.
11451 *Ben Laurie (CHATS)*
11453 * Change the SSL kerb5 codes to match RFC 2712.
11457 * Make -nameopt work fully for req and add -reqopt switch.
11459 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11461 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11463 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11465 * Make sure tests can be performed even if the corresponding algorithms
11466 have been removed entirely. This was also the last step to make
11467 OpenSSL compilable with DJGPP under all reasonable conditions.
11469 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11471 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11472 to allow version independent disabling of normally unselected ciphers,
11473 which may be activated as a side-effect of selecting a single cipher.
11475 (E.g., cipher list string "RSA" enables ciphersuites that are left
11476 out of "ALL" because they do not provide symmetric encryption.
11477 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11479 *Lutz Jaenicke, Bodo Moeller*
11481 * Add appropriate support for separate platform-dependent build
11482 directories. The recommended way to make a platform-dependent
11483 build directory is the following (tested on Linux), maybe with
11486 # Place yourself outside of the OpenSSL source tree. In
11487 # this example, the environment variable OPENSSL_SOURCE
11488 # is assumed to contain the absolute OpenSSL source directory.
11489 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11490 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11491 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11492 mkdir -p `dirname $F`
11493 ln -s $OPENSSL_SOURCE/$F $F
11496 To be absolutely sure not to disturb the source tree, a "make clean"
11497 is a good thing. If it isn't successful, don't worry about it,
11498 it probably means the source directory is very clean.
11502 * Make sure any ENGINE control commands make local copies of string
11503 pointers passed to them whenever necessary. Otherwise it is possible
11504 the caller may have overwritten (or deallocated) the original string
11505 data when a later ENGINE operation tries to use the stored values.
11507 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11509 * Improve diagnostics in file reading and command-line digests.
11511 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11513 * Add AES modes CFB and OFB to the object database. Correct an
11514 error in AES-CFB decryption.
11518 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11519 allows existing EVP_CIPHER_CTX structures to be reused after
11520 calling `EVP_*Final()`. This behaviour is used by encryption
11521 BIOs and some applications. This has the side effect that
11522 applications must explicitly clean up cipher contexts with
11523 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11527 * Check the values of dna and dnb in bn_mul_recursive before calling
11528 bn_mul_comba (a non zero value means the a or b arrays do not contain
11529 n2 elements) and fallback to bn_mul_normal if either is not zero.
11533 * Fix escaping of non-ASCII characters when using the -subj option
11534 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11538 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11539 form for "surname", serialNumber has no short form.
11540 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11541 therefore remove "mail" short name for "internet 7".
11542 The OID for unique identifiers in X509 certificates is
11543 x500UniqueIdentifier, not uniqueIdentifier.
11544 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11548 * Add an "init" command to the ENGINE config module and auto initialize
11549 ENGINEs. Without any "init" command the ENGINE will be initialized
11550 after all ctrl commands have been executed on it. If init=1 the
11551 ENGINE is initialized at that point (ctrls before that point are run
11552 on the uninitialized ENGINE and after on the initialized one). If
11553 init=0 then the ENGINE will not be initialized at all.
11557 * Fix the 'app_verify_callback' interface so that the user-defined
11558 argument is actually passed to the callback: In the
11559 SSL_CTX_set_cert_verify_callback() prototype, the callback
11560 declaration has been changed from
11563 int (*cb)(X509_STORE_CTX *,void *);
11564 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11565 i=s->ctx->app_verify_callback(&ctx)
11566 has been changed into
11567 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11569 To update applications using SSL_CTX_set_cert_verify_callback(),
11570 a dummy argument can be added to their callback functions.
11572 *D. K. Smetters <smetters@parc.xerox.com>*
11574 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11576 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11578 * Add and OPENSSL_LOAD_CONF define which will cause
11579 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11580 This allows older applications to transparently support certain
11581 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11582 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11583 load the config file and OPENSSL_add_all_algorithms_conf() which will
11584 always load it have also been added.
11588 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11589 Adjust NIDs and EVP layer.
11591 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11593 * Config modules support in openssl utility.
11595 Most commands now load modules from the config file,
11596 though in a few (such as version) this isn't done
11597 because it couldn't be used for anything.
11599 In the case of ca and req the config file used is
11600 the same as the utility itself: that is the -config
11601 command line option can be used to specify an
11606 * Move default behaviour from OPENSSL_config(). If appname is NULL
11607 use "openssl_conf" if filename is NULL use default openssl config file.
11611 * Add an argument to OPENSSL_config() to allow the use of an alternative
11612 config section name. Add a new flag to tolerate a missing config file
11613 and move code to CONF_modules_load_file().
11617 * Support for crypto accelerator cards from Accelerated Encryption
11618 Processing, www.aep.ie. (Use engine 'aep')
11619 The support was copied from 0.9.6c [engine] and adapted/corrected
11620 to work with the new engine framework.
11622 *AEP Inc. and Richard Levitte*
11624 * Support for SureWare crypto accelerator cards from Baltimore
11625 Technologies. (Use engine 'sureware')
11626 The support was copied from 0.9.6c [engine] and adapted
11627 to work with the new engine framework.
11631 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11632 make the newer ENGINE framework commands for the CHIL engine work.
11634 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11636 * Make it possible to produce shared libraries on ReliantUNIX.
11638 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11640 * Add the configuration target debug-linux-ppro.
11641 Make 'openssl rsa' use the general key loading routines
11642 implemented in apps.c, and make those routines able to
11643 handle the key format FORMAT_NETSCAPE and the variant
11646 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11648 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11650 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11652 * Add -keyform to rsautl, and document -engine.
11654 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11656 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11657 BIO_R_NO_SUCH_FILE error code rather than the generic
11658 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11662 * Add new functions
11663 ERR_peek_last_error
11664 ERR_peek_last_error_line
11665 ERR_peek_last_error_line_data.
11666 These are similar to
11668 ERR_peek_error_line
11669 ERR_peek_error_line_data,
11670 but report on the latest error recorded rather than the first one
11671 still in the error queue.
11673 *Ben Laurie, Bodo Moeller*
11675 * default_algorithms option in ENGINE config module. This allows things
11677 default_algorithms = ALL
11678 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11682 * Preliminary ENGINE config module.
11686 * New experimental application configuration code.
11690 * Change the AES code to follow the same name structure as all other
11691 symmetric ciphers, and behave the same way. Move everything to
11692 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11694 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11696 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11698 *Ben Laurie and Theo de Raadt*
11700 * Add option to output public keys in req command.
11702 *Massimiliano Pala madwolf@openca.org*
11704 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11705 (up to about 10% better than before for P-192 and P-224).
11709 * New functions/macros
11711 SSL_CTX_set_msg_callback(ctx, cb)
11712 SSL_CTX_set_msg_callback_arg(ctx, arg)
11713 SSL_set_msg_callback(ssl, cb)
11714 SSL_set_msg_callback_arg(ssl, arg)
11716 to request calling a callback function
11718 void cb(int write_p, int version, int content_type,
11719 const void *buf, size_t len, SSL *ssl, void *arg)
11721 whenever a protocol message has been completely received
11722 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11723 protocol version according to which the SSL library interprets
11724 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11725 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11726 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11727 specification (change_cipher_spec(20), alert(21), handshake(22)).
11728 'buf' and 'len' point to the actual message, 'ssl' to the
11729 SSL object, and 'arg' is the application-defined value set by
11730 SSL[_CTX]_set_msg_callback_arg().
11732 'openssl s_client' and 'openssl s_server' have new '-msg' options
11733 to enable a callback that displays all protocol messages.
11737 * Change the shared library support so shared libraries are built as
11738 soon as the corresponding static library is finished, and thereby get
11739 openssl and the test programs linked against the shared library.
11740 This still only happens when the keyword "shard" has been given to
11741 the configuration scripts.
11743 NOTE: shared library support is still an experimental thing, and
11744 backward binary compatibility is still not guaranteed.
11746 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11748 * Add support for Subject Information Access extension.
11750 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11752 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11753 additional bytes when new memory had to be allocated, not just
11754 when reusing an existing buffer.
11758 * New command line and configuration option 'utf8' for the req command.
11759 This allows field values to be specified as UTF8 strings.
11763 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11764 runs for the former and machine-readable output for the latter.
11768 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11769 of the e-mail address in the DN (i.e., it will go into a certificate
11770 extension only). The new configuration file option 'email_in_dn = no'
11771 has the same effect.
11773 *Massimiliano Pala madwolf@openca.org*
11775 * Change all functions with names starting with `des_` to be starting
11776 with `DES_` instead. Add wrappers that are compatible with libdes,
11777 but are named `_ossl_old_des_*`. Finally, add macros that map the
11778 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11779 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11780 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11783 Since we provide two compatibility mappings, the user needs to
11784 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11785 compatibility is desired. The default (i.e., when that macro
11786 isn't defined) is OpenSSL 0.9.6c compatibility.
11788 There are also macros that enable and disable the support of old
11789 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11790 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11791 are defined, the default will apply: to support the old des routines.
11793 In either case, one must include openssl/des.h to get the correct
11794 definitions. Do not try to just include openssl/des_old.h, that
11797 NOTE: This is a major break of an old API into a new one. Software
11798 authors are encouraged to switch to the `DES_` style functions. Some
11799 time in the future, des_old.h and the libdes compatibility functions
11800 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11801 default), and then completely removed.
11805 * Test for certificates which contain unsupported critical extensions.
11806 If such a certificate is found during a verify operation it is
11807 rejected by default: this behaviour can be overridden by either
11808 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11809 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11810 X509_supported_extension() has also been added which returns 1 if a
11811 particular extension is supported.
11815 * Modify the behaviour of EVP cipher functions in similar way to digests
11816 to retain compatibility with existing code.
11820 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11821 compatibility with existing code. In particular the 'ctx' parameter does
11822 not have to be to be initialized before the call to EVP_DigestInit() and
11823 it is tidied up after a call to EVP_DigestFinal(). New function
11824 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11825 EVP_MD_CTX_copy() changed to not require the destination to be
11826 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11827 requires the destination to be valid.
11829 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11830 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11834 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11835 so that complete 'Handshake' protocol structures are kept in memory
11836 instead of overwriting 'msg_type' and 'length' with 'body' data.
11840 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11842 *Massimo Santin via Richard Levitte*
11844 * Major restructuring to the underlying ENGINE code. This includes
11845 reduction of linker bloat, separation of pure "ENGINE" manipulation
11846 (initialisation, etc) from functionality dealing with implementations
11847 of specific crypto interfaces. This change also introduces integrated
11848 support for symmetric ciphers and digest implementations - so ENGINEs
11849 can now accelerate these by providing EVP_CIPHER and EVP_MD
11850 implementations of their own. This is detailed in
11851 [crypto/engine/README.md](crypto/engine/README.md)
11852 as it couldn't be adequately described here. However, there are a few
11853 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11854 were changed in the original introduction of ENGINE code have now
11855 reverted back - the hooking from this code to ENGINE is now a good
11856 deal more passive and at run-time, operations deal directly with
11857 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11858 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11859 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11860 they were not being used by the framework as there is no concept of a
11861 BIGNUM_METHOD and they could not be generalised to the new
11862 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11863 ENGINE_cpy() has been removed as it cannot be consistently defined in
11868 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11872 * Change mkdef.pl to sort symbols that get the same entry number,
11873 and make sure the automatically generated functions `ERR_load_*`
11874 become part of libeay.num as well.
11878 * New function SSL_renegotiate_pending(). This returns true once
11879 renegotiation has been requested (either SSL_renegotiate() call
11880 or HelloRequest/ClientHello received from the peer) and becomes
11881 false once a handshake has been completed.
11882 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11883 sends a HelloRequest, but does not ensure that a handshake takes
11884 place. SSL_renegotiate_pending() is useful for checking if the
11885 client has followed the request.)
11889 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11890 By default, clients may request session resumption even during
11891 renegotiation (if session ID contexts permit); with this option,
11892 session resumption is possible only in the first handshake.
11894 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11895 more bits available for options that should not be part of
11896 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11900 * Add some demos for certificate and certificate request creation.
11904 * Make maximum certificate chain size accepted from the peer application
11905 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11906 "Douglas E. Engert" <deengert@anl.gov>.
11910 * Add support for shared libraries for Unixware-7
11911 (Boyd Lynn Gerber <gerberb@zenez.com>).
11915 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11916 be done prior to destruction. Use this to unload error strings from
11917 ENGINEs that load their own error strings. NB: This adds two new API
11918 functions to "get" and "set" this destroy handler in an ENGINE.
11922 * Alter all existing ENGINE implementations (except "openssl" and
11923 "openbsd") to dynamically instantiate their own error strings. This
11924 makes them more flexible to be built both as statically-linked ENGINEs
11925 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11926 Also, add stub code to each that makes building them as self-contained
11927 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11931 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11932 implementations into applications that are completely implemented in
11933 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11934 commands that can be used to configure what shared-library to load and
11935 to control aspects of the way it is handled. Also, made an update to
11936 the [README-Engine.md](README-Engine.md) file
11937 that brings its information up-to-date and
11938 provides some information and instructions on the "dynamic" ENGINE
11939 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11943 * Make it possible to unload ranges of ERR strings with a new
11944 "ERR_unload_strings" function.
11948 * Add a copy() function to EVP_MD.
11952 * Make EVP_MD routines take a context pointer instead of just the
11953 md_data void pointer.
11957 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11958 that the digest can only process a single chunk of data
11959 (typically because it is provided by a piece of
11960 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11961 is only going to provide a single chunk of data, and hence the
11962 framework needn't accumulate the data for oneshot drivers.
11966 * As with "ERR", make it possible to replace the underlying "ex_data"
11967 functions. This change also alters the storage and management of global
11968 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11969 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11970 index counters. The API functions that use this state have been changed
11971 to take a "class_index" rather than pointers to the class's local STACK
11972 and counter, and there is now an API function to dynamically create new
11973 classes. This centralisation allows us to (a) plug a lot of the
11974 thread-safety problems that existed, and (b) makes it possible to clean
11975 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11976 such data would previously have always leaked in application code and
11977 workarounds were in place to make the memory debugging turn a blind eye
11978 to it. Application code that doesn't use this new function will still
11979 leak as before, but their memory debugging output will announce it now
11980 rather than letting it slide.
11982 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11983 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11984 has a return value to indicate success or failure.
11988 * Make it possible to replace the underlying "ERR" functions such that the
11989 global state (2 LHASH tables and 2 locks) is only used by the "default"
11990 implementation. This change also adds two functions to "get" and "set"
11991 the implementation prior to it being automatically set the first time
11992 any other ERR function takes place. Ie. an application can call "get",
11993 pass the return value to a module it has just loaded, and that module
11994 can call its own "set" function using that value. This means the
11995 module's "ERR" operations will use (and modify) the error state in the
11996 application and not in its own statically linked copy of OpenSSL code.
12000 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12001 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12002 the operation, and provides a more encapsulated way for external code
12003 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12004 to use these functions rather than manually incrementing the counts.
12006 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12010 * Add EVP test program.
12014 * Add symmetric cipher support to ENGINE. Expect the API to change!
12018 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12019 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12020 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12021 These allow a CRL to be built without having to access X509_CRL fields
12022 directly. Modify 'ca' application to use new functions.
12026 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12027 bug workarounds. Rollback attack detection is a security feature.
12028 The problem will only arise on OpenSSL servers when TLSv1 is not
12029 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12030 Software authors not wanting to support TLSv1 will have special reasons
12031 for their choice and can explicitly enable this option.
12033 *Bodo Moeller, Lutz Jaenicke*
12035 * Rationalise EVP so it can be extended: don't include a union of
12036 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12037 (similar to those existing for EVP_CIPHER_CTX).
12042 EVP_MD_CTX_init(&md); /* new function call */
12043 EVP_DigestInit(&md, EVP_sha1());
12044 EVP_DigestUpdate(&md, in, len);
12045 EVP_DigestFinal(&md, out, NULL);
12046 EVP_MD_CTX_cleanup(&md); /* new function call */
12050 * Make DES key schedule conform to the usual scheme, as well as
12051 correcting its structure. This means that calls to DES functions
12052 now have to pass a pointer to a des_key_schedule instead of a
12053 plain des_key_schedule (which was actually always a pointer
12056 des_key_schedule ks;
12058 des_set_key_checked(..., &ks);
12059 des_ncbc_encrypt(..., &ks, ...);
12061 (Note that a later change renames 'des_...' into 'DES_...'.)
12065 * Initial reduction of linker bloat: the use of some functions, such as
12066 PEM causes large amounts of unused functions to be linked in due to
12067 poor organisation. For example pem_all.c contains every PEM function
12068 which has a knock on effect of linking in large amounts of (unused)
12069 ASN1 code. Grouping together similar functions and splitting unrelated
12070 functions prevents this.
12074 * Cleanup of EVP macros.
12078 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12079 correct `_ecb suffix`.
12083 * Add initial OCSP responder support to ocsp application. The
12084 revocation information is handled using the text based index
12085 use by the ca application. The responder can either handle
12086 requests generated internally, supplied in files (for example
12087 via a CGI script) or using an internal minimal server.
12091 * Add configuration choices to get zlib compression for TLS.
12095 * Changes to Kerberos SSL for RFC 2712 compliance:
12096 1. Implemented real KerberosWrapper, instead of just using
12097 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12098 2. Implemented optional authenticator field of KerberosWrapper.
12100 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12101 and authenticator structs; see crypto/krb5/.
12103 Generalized Kerberos calls to support multiple Kerberos libraries.
12104 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12105 via Richard Levitte*
12107 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12108 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12109 values for each of the key sizes rather than having just
12110 parameters (and 'speed' generating keys each time).
12114 * Speed up EVP routines.
12117 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12118 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12119 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12120 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12122 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12123 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12124 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12127 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12129 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12133 * Added the OS2-EMX target.
12135 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12137 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12138 to support NCONF routines in extension code. New function CONF_set_nconf()
12139 to allow functions which take an NCONF to also handle the old LHASH
12140 structure: this means that the old CONF compatible routines can be
12141 retained (in particular wrt extensions) without having to duplicate the
12142 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12146 * Enhance the general user interface with mechanisms for inner control
12147 and with possibilities to have yes/no kind of prompts.
12151 * Change all calls to low level digest routines in the library and
12152 applications to use EVP. Add missing calls to HMAC_cleanup() and
12153 don't assume HMAC_CTX can be copied using memcpy().
12155 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12157 * Add the possibility to control engines through control names but with
12158 arbitrary arguments instead of just a string.
12159 Change the key loaders to take a UI_METHOD instead of a callback
12160 function pointer. NOTE: this breaks binary compatibility with earlier
12161 versions of OpenSSL [engine].
12162 Adapt the nCipher code for these new conditions and add a card insertion
12167 * Enhance the general user interface with mechanisms to better support
12168 dialog box interfaces, application-defined prompts, the possibility
12169 to use defaults (for example default passwords from somewhere else)
12170 and interrupts/cancellations.
12174 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12175 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12179 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12180 tidy up some unnecessarily weird code in 'sk_new()').
12182 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12184 * Change the key loading routines for ENGINEs to use the same kind
12185 callback (pem_password_cb) as all other routines that need this
12190 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12191 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12192 than this minimum value is recommended.
12196 * New random seeder for OpenVMS, using the system process statistics
12197 that are easily reachable.
12201 * Windows apparently can't transparently handle global
12202 variables defined in DLLs. Initialisations such as:
12204 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12206 won't compile. This is used by the any applications that need to
12207 declare their own ASN1 modules. This was fixed by adding the option
12208 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12209 needed for static libraries under Win32.
12213 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12214 setting of purpose and trust fields. New X509_STORE trust and
12215 purpose functions and tidy up setting in other SSL functions.
12219 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12220 structure. These are inherited by X509_STORE_CTX when it is
12221 initialised. This allows various defaults to be set in the
12222 X509_STORE structure (such as flags for CRL checking and custom
12223 purpose or trust settings) for functions which only use X509_STORE_CTX
12224 internally such as S/MIME.
12226 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12227 trust settings if they are not set in X509_STORE. This allows X509_STORE
12228 purposes and trust (in S/MIME for example) to override any set by default.
12230 Add command line options for CRL checking to smime, s_client and s_server
12235 * Initial CRL based revocation checking. If the CRL checking flag(s)
12236 are set then the CRL is looked up in the X509_STORE structure and
12237 its validity and signature checked, then if the certificate is found
12238 in the CRL the verify fails with a revoked error.
12240 Various new CRL related callbacks added to X509_STORE_CTX structure.
12242 Command line options added to 'verify' application to support this.
12244 This needs some additional work, such as being able to handle multiple
12245 CRLs with different times, extension based lookup (rather than just
12246 by subject name) and ultimately more complete V2 CRL extension
12251 * Add a general user interface API (crypto/ui/). This is designed
12252 to replace things like des_read_password and friends (backward
12253 compatibility functions using this new API are provided).
12254 The purpose is to remove prompting functions from the DES code
12255 section as well as provide for prompting through dialog boxes in
12256 a window system and the like.
12260 * Add "ex_data" support to ENGINE so implementations can add state at a
12261 per-structure level rather than having to store it globally.
12265 * Make it possible for ENGINE structures to be copied when retrieved by
12266 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12267 This causes the "original" ENGINE structure to act like a template,
12268 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12269 operational state can be localised to each ENGINE structure, despite the
12270 fact they all share the same "methods". New ENGINE structures returned in
12271 this case have no functional references and the return value is the single
12272 structural reference. This matches the single structural reference returned
12273 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12278 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12279 needs to match any other type at all we need to manually clear the
12284 * Changes to the "openssl engine" utility to include;
12285 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12286 about an ENGINE's available control commands.
12287 - executing control commands from command line arguments using the
12288 '-pre' and '-post' switches. '-post' is only used if '-t' is
12289 specified and the ENGINE is successfully initialised. The syntax for
12290 the individual commands are colon-separated, for example;
12291 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12295 * New dynamic control command support for ENGINEs. ENGINEs can now
12296 declare their own commands (numbers), names (strings), descriptions,
12297 and input types for run-time discovery by calling applications. A
12298 subset of these commands are implicitly classed as "executable"
12299 depending on their input type, and only these can be invoked through
12300 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12301 can be based on user input, config files, etc). The distinction is
12302 that "executable" commands cannot return anything other than a boolean
12303 result and can only support numeric or string input, whereas some
12304 discoverable commands may only be for direct use through
12305 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12306 pointers, or other custom uses. The "executable" commands are to
12307 support parameterisations of ENGINE behaviour that can be
12308 unambiguously defined by ENGINEs and used consistently across any
12309 OpenSSL-based application. Commands have been added to all the
12310 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12311 control over shared-library paths without source code alterations.
12315 * Changed all ENGINE implementations to dynamically allocate their
12316 ENGINEs rather than declaring them statically. Apart from this being
12317 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12318 this also allows the implementations to compile without using the
12319 internal engine_int.h header.
12323 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12324 'const' value. Any code that should be able to modify a RAND_METHOD
12325 should already have non-const pointers to it (ie. they should only
12326 modify their own ones).
12330 * Made a variety of little tweaks to the ENGINE code.
12331 - "atalla" and "ubsec" string definitions were moved from header files
12332 to C code. "nuron" string definitions were placed in variables
12333 rather than hard-coded - allowing parameterisation of these values
12334 later on via ctrl() commands.
12335 - Removed unused "#if 0"'d code.
12336 - Fixed engine list iteration code so it uses ENGINE_free() to release
12337 structural references.
12338 - Constified the RAND_METHOD element of ENGINE structures.
12339 - Constified various get/set functions as appropriate and added
12340 missing functions (including a catch-all ENGINE_cpy that duplicates
12341 all ENGINE values onto a new ENGINE except reference counts/state).
12342 - Removed NULL parameter checks in get/set functions. Setting a method
12343 or function to NULL is a way of cancelling out a previously set
12344 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12345 and doesn't justify the extra error symbols and code.
12346 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12347 flags from engine_int.h to engine.h.
12348 - Changed prototypes for ENGINE handler functions (init(), finish(),
12349 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12353 * Implement binary inversion algorithm for BN_mod_inverse in addition
12354 to the algorithm using long division. The binary algorithm can be
12355 used only if the modulus is odd. On 32-bit systems, it is faster
12356 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12357 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12358 up to 450 bits. In 64-bit environments, the binary algorithm
12359 appears to be advantageous for much longer moduli; here we use it
12360 for moduli up to 2048 bits.
12364 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12365 could not support the combine flag in choice fields.
12369 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12370 extensions from a certificate request to the certificate.
12374 * Allow multiple 'certopt' and 'nameopt' options to be separated
12375 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12376 file: this allows the display of the certificate about to be
12377 signed to be customised, to allow certain fields to be included
12378 or excluded and extension details. The old system didn't display
12379 multicharacter strings properly, omitted fields not in the policy
12380 and couldn't display additional details such as extensions.
12384 * Function EC_POINTs_mul for multiple scalar multiplication
12385 of an arbitrary number of elliptic curve points
12386 \sum scalars[i]*points[i],
12387 optionally including the generator defined for the EC_GROUP:
12388 scalar*generator + \sum scalars[i]*points[i].
12390 EC_POINT_mul is a simple wrapper function for the typical case
12391 that the point list has just one item (besides the optional
12396 * First EC_METHODs for curves over GF(p):
12398 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12399 operations and provides various method functions that can also
12400 operate with faster implementations of modular arithmetic.
12402 EC_GFp_mont_method() reuses most functions that are part of
12403 EC_GFp_simple_method, but uses Montgomery arithmetic.
12405 *Bodo Moeller; point addition and point doubling
12406 implementation directly derived from source code provided by
12407 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12409 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12410 crypto/ec/ec_lib.c):
12412 Curves are EC_GROUP objects (with an optional group generator)
12413 based on EC_METHODs that are built into the library.
12415 Points are EC_POINT objects based on EC_GROUP objects.
12417 Most of the framework would be able to handle curves over arbitrary
12418 finite fields, but as there are no obvious types for fields other
12419 than GF(p), some functions are limited to that for now.
12423 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12424 that the file contains a complete HTTP response.
12428 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12429 change the def and num file printf format specifier from "%-40sXXX"
12430 to "%-39s XXX". The latter will always guarantee a space after the
12431 field while the former will cause them to run together if the field
12432 is 40 of more characters long.
12436 * Constify the cipher and digest 'method' functions and structures
12437 and modify related functions to take constant EVP_MD and EVP_CIPHER
12442 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12443 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12447 * Modify `EVP_Digest*()` routines so they now return values. Although the
12448 internal software routines can never fail additional hardware versions
12453 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12455 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12456 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12459 ERR_R_NESTED_ASN1_ERROR
12461 ERR_R_MISSING_ASN1_EOS
12462 were 4 .. 9, conflicting with
12463 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12465 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12466 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12468 Add new error code 'ERR_R_INTERNAL_ERROR'.
12472 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12477 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12478 sets the subject name for a new request or supersedes the
12479 subject name in a given request. Formats that can be parsed are
12480 'CN=Some Name, OU=myOU, C=IT'
12482 'CN=Some Name/OU=myOU/C=IT'.
12484 Add options '-batch' and '-verbose' to 'openssl req'.
12486 *Massimiliano Pala <madwolf@hackmasters.net>*
12488 * Introduce the possibility to access global variables through
12489 functions on platform were that's the best way to handle exporting
12490 global variables in shared libraries. To enable this functionality,
12491 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12492 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12493 is normally done by Configure or something similar).
12495 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12496 in the source file (foo.c) like this:
12498 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12499 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12501 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12502 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12504 OPENSSL_DECLARE_GLOBAL(int,foo);
12505 #define foo OPENSSL_GLOBAL_REF(foo)
12506 OPENSSL_DECLARE_GLOBAL(double,bar);
12507 #define bar OPENSSL_GLOBAL_REF(bar)
12509 The #defines are very important, and therefore so is including the
12510 header file everywhere where the defined globals are used.
12512 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12513 of ASN.1 items, but that structure is a bit different.
12515 The largest change is in util/mkdef.pl which has been enhanced with
12516 better and easier to understand logic to choose which symbols should
12517 go into the Windows .def files as well as a number of fixes and code
12518 cleanup (among others, algorithm keywords are now sorted
12519 lexicographically to avoid constant rewrites).
12523 * In BN_div() keep a copy of the sign of 'num' before writing the
12524 result to 'rm' because if rm==num the value will be overwritten
12525 and produce the wrong result if 'num' is negative: this caused
12526 problems with BN_mod() and BN_nnmod().
12530 * Function OCSP_request_verify(). This checks the signature on an
12531 OCSP request and verifies the signer certificate. The signer
12532 certificate is just checked for a generic purpose and OCSP request
12537 * Add OCSP_check_validity() function to check the validity of OCSP
12538 responses. OCSP responses are prepared in real time and may only
12539 be a few seconds old. Simply checking that the current time lies
12540 between thisUpdate and nextUpdate max reject otherwise valid responses
12541 caused by either OCSP responder or client clock inaccuracy. Instead
12542 we allow thisUpdate and nextUpdate to fall within a certain period of
12543 the current time. The age of the response can also optionally be
12544 checked. Two new options -validity_period and -status_age added to
12549 * If signature or public key algorithm is unrecognized print out its
12550 OID rather that just UNKNOWN.
12554 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12555 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12556 ID to be generated from the issuer certificate alone which can then be
12557 passed to OCSP_id_issuer_cmp().
12561 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12562 ASN1 modules to export functions returning ASN1_ITEM pointers
12563 instead of the ASN1_ITEM structures themselves. This adds several
12564 new macros which allow the underlying ASN1 function/structure to
12565 be accessed transparently. As a result code should not use ASN1_ITEM
12566 references directly (such as &X509_it) but instead use the relevant
12567 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12568 use of the new ASN1 code on platforms where exporting structures
12569 is problematical (for example in shared libraries) but exporting
12570 functions returning pointers to structures is not.
12574 * Add support for overriding the generation of SSL/TLS session IDs.
12575 These callbacks can be registered either in an SSL_CTX or per SSL.
12576 The purpose of this is to allow applications to control, if they wish,
12577 the arbitrary values chosen for use as session IDs, particularly as it
12578 can be useful for session caching in multiple-server environments. A
12579 command-line switch for testing this (and any client code that wishes
12580 to use such a feature) has been added to "s_server".
12582 *Geoff Thorpe, Lutz Jaenicke*
12584 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12585 of the form `#if defined(...) || defined(...) || ...` and
12586 `#if !defined(...) && !defined(...) && ...`. This also avoids
12587 the growing number of special cases it was previously handling.
12591 * Make all configuration macros available for application by making
12592 sure they are available in opensslconf.h, by giving them names starting
12593 with `OPENSSL_` to avoid conflicts with other packages and by making
12594 sure e_os2.h will cover all platform-specific cases together with
12596 Additionally, it is now possible to define configuration/platform-
12597 specific names (called "system identities"). In the C code, these
12598 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12599 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12600 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12605 * New option -set_serial to 'req' and 'x509' this allows the serial
12606 number to use to be specified on the command line. Previously self
12607 signed certificates were hard coded with serial number 0 and the
12608 CA options of 'x509' had to use a serial number in a file which was
12613 * New options to 'ca' utility to support V2 CRL entry extensions.
12614 Currently CRL reason, invalidity date and hold instruction are
12615 supported. Add new CRL extensions to V3 code and some new objects.
12619 * New function EVP_CIPHER_CTX_set_padding() this is used to
12620 disable standard block padding (aka PKCS#5 padding) in the EVP
12621 API, which was previously mandatory. This means that the data is
12622 not padded in any way and so the total length much be a multiple
12623 of the block size, otherwise an error occurs.
12627 * Initial (incomplete) OCSP SSL support.
12631 * New function OCSP_parse_url(). This splits up a URL into its host,
12632 port and path components: primarily to parse OCSP URLs. New -url
12633 option to ocsp utility.
12637 * New nonce behavior. The return value of OCSP_check_nonce() now
12638 reflects the various checks performed. Applications can decide
12639 whether to tolerate certain situations such as an absent nonce
12640 in a response when one was present in a request: the ocsp application
12641 just prints out a warning. New function OCSP_add1_basic_nonce()
12642 this is to allow responders to include a nonce in a response even if
12643 the request is nonce-less.
12647 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12648 skipped when using openssl x509 multiple times on a single input file,
12649 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12653 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12654 set string type: to handle setting ASN1_TIME structures. Fix ca
12655 utility to correctly initialize revocation date of CRLs.
12659 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12660 the clients preferred ciphersuites and rather use its own preferences.
12661 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12662 Internet Explorer by ensuring unchanged hash method during stepup.
12663 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12667 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12668 to aes and add a new 'exist' option to print out symbols that don't
12673 * Additional options to ocsp utility to allow flags to be set and
12674 additional certificates supplied.
12678 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12679 OCSP client a number of certificate to only verify the response
12684 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12685 handle the new API. Currently only ECB, CBC modes supported. Add new
12688 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12689 Encryption Standard (AES) Ciphersuites for Transport Layer
12690 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12691 not enabled by default and were not part of the "ALL" ciphersuite
12692 alias because they were not yet official; they could be
12693 explicitly requested by specifying the "AESdraft" ciphersuite
12694 group alias. In the final release of OpenSSL 0.9.7, the group
12695 alias is called "AES" and is part of "ALL".)
12697 *Ben Laurie, Steve Henson, Bodo Moeller*
12699 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12700 request to response.
12704 * Functions for OCSP responders. OCSP_request_onereq_count(),
12705 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12706 extract information from a certificate request. OCSP_response_create()
12707 creates a response and optionally adds a basic response structure.
12708 OCSP_basic_add1_status() adds a complete single response to a basic
12709 response and returns the OCSP_SINGLERESP structure just added (to allow
12710 extensions to be included for example). OCSP_basic_add1_cert() adds a
12711 certificate to a basic response and OCSP_basic_sign() signs a basic
12712 response with various flags. New helper functions ASN1_TIME_check()
12713 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12714 (converts ASN1_TIME to GeneralizedTime).
12718 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12719 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12720 structure from a certificate. X509_pubkey_digest() digests the public_key
12721 contents: this is used in various key identifiers.
12725 * Make sk_sort() tolerate a NULL argument.
12727 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12729 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12730 passed by the function are trusted implicitly. If any of them signed the
12731 response then it is assumed to be valid and is not verified.
12735 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12736 to data. This was previously part of the PKCS7 ASN1 code. This
12737 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12738 *Steve Henson, reported by Kenneth R. Robinette
12739 <support@securenetterm.com>*
12741 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12742 routines: without these tracing memory leaks is very painful.
12743 Fix leaks in PKCS12 and PKCS7 routines.
12747 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12748 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12749 effectively meant GeneralizedTime would never be used. Now it
12750 is initialised to -1 but X509_time_adj() now has to check the value
12751 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12752 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12753 *Steve Henson, reported by Kenneth R. Robinette
12754 <support@securenetterm.com>*
12756 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12757 result in a zero length in the ASN1_INTEGER structure which was
12758 not consistent with the structure when d2i_ASN1_INTEGER() was used
12759 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12760 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12761 where it did not print out a minus for negative ASN1_INTEGER.
12765 * Add summary printout to ocsp utility. The various functions which
12766 convert status values to strings have been renamed to:
12767 OCSP_response_status_str(), OCSP_cert_status_str() and
12768 OCSP_crl_reason_str() and are no longer static. New options
12769 to verify nonce values and to disable verification. OCSP response
12770 printout format cleaned up.
12774 * Add additional OCSP certificate checks. These are those specified
12775 in RFC2560. This consists of two separate checks: the CA of the
12776 certificate being checked must either be the OCSP signer certificate
12777 or the issuer of the OCSP signer certificate. In the latter case the
12778 OCSP signer certificate must contain the OCSP signing extended key
12779 usage. This check is performed by attempting to match the OCSP
12780 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12781 in the OCSP_CERTID structures of the response.
12785 * Initial OCSP certificate verification added to OCSP_basic_verify()
12786 and related routines. This uses the standard OpenSSL certificate
12787 verify routines to perform initial checks (just CA validity) and
12788 to obtain the certificate chain. Then additional checks will be
12789 performed on the chain. Currently the root CA is checked to see
12790 if it is explicitly trusted for OCSP signing. This is used to set
12791 a root CA as a global signing root: that is any certificate that
12792 chains to that CA is an acceptable OCSP signing certificate.
12796 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12797 extensions from a separate configuration file.
12798 As when reading extensions from the main configuration file,
12799 the '-extensions ...' option may be used for specifying the
12802 *Massimiliano Pala <madwolf@comune.modena.it>*
12804 * New OCSP utility. Allows OCSP requests to be generated or
12805 read. The request can be sent to a responder and the output
12806 parsed, outputted or printed in text form. Not complete yet:
12807 still needs to check the OCSP response validity.
12811 * New subcommands for 'openssl ca':
12812 `openssl ca -status <serial>` prints the status of the cert with
12813 the given serial number (according to the index file).
12814 `openssl ca -updatedb` updates the expiry status of certificates
12817 *Massimiliano Pala <madwolf@comune.modena.it>*
12819 * New '-newreq-nodes' command option to CA.pl. This is like
12820 '-newreq', but calls 'openssl req' with the '-nodes' option
12821 so that the resulting key is not encrypted.
12823 *Damien Miller <djm@mindrot.org>*
12825 * New configuration for the GNU Hurd.
12827 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12829 * Initial code to implement OCSP basic response verify. This
12830 is currently incomplete. Currently just finds the signer's
12831 certificate and verifies the signature on the response.
12835 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12836 value of OPENSSLDIR. This is available via the new '-d' option
12837 to 'openssl version', and is also included in 'openssl version -a'.
12841 * Allowing defining memory allocation callbacks that will be given
12842 file name and line number information in additional arguments
12843 (a `const char*` and an int). The basic functionality remains, as
12844 well as the original possibility to just replace malloc(),
12845 realloc() and free() by functions that do not know about these
12846 additional arguments. To register and find out the current
12847 settings for extended allocation functions, the following
12848 functions are provided:
12850 CRYPTO_set_mem_ex_functions
12851 CRYPTO_set_locked_mem_ex_functions
12852 CRYPTO_get_mem_ex_functions
12853 CRYPTO_get_locked_mem_ex_functions
12855 These work the same way as CRYPTO_set_mem_functions and friends.
12856 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12857 extended allocation function is enabled.
12858 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12859 a conventional allocation function is enabled.
12861 *Richard Levitte, Bodo Moeller*
12863 * Finish off removing the remaining LHASH function pointer casts.
12864 There should no longer be any prototype-casting required when using
12865 the LHASH abstraction, and any casts that remain are "bugs". See
12866 the callback types and macros at the head of lhash.h for details
12867 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12871 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12872 If /dev/[u]random devices are not available or do not return enough
12873 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12875 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12876 /etc/entropy will be queried once each in this sequence, querying stops
12877 when enough entropy was collected without querying more sockets.
12881 * Change the Unix RAND_poll() variant to be able to poll several
12882 random devices, as specified by DEVRANDOM, until a sufficient amount
12883 of data has been collected. We spend at most 10 ms on each file
12884 (select timeout) and read in non-blocking mode. DEVRANDOM now
12885 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12886 (previously it was just the string "/dev/urandom"), so on typical
12887 platforms the 10 ms delay will never occur.
12888 Also separate out the Unix variant to its own file, rand_unix.c.
12889 For VMS, there's a currently-empty rand_vms.c.
12893 * Move OCSP client related routines to ocsp_cl.c. These
12894 provide utility functions which an application needing
12895 to issue a request to an OCSP responder and analyse the
12896 response will typically need: as opposed to those which an
12897 OCSP responder itself would need which will be added later.
12899 OCSP_request_sign() signs an OCSP request with an API similar
12900 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12901 response. OCSP_response_get1_basic() extracts basic response
12902 from response. OCSP_resp_find_status(): finds and extracts status
12903 information from an OCSP_CERTID structure (which will be created
12904 when the request structure is built). These are built from lower
12905 level functions which work on OCSP_SINGLERESP structures but
12906 won't normally be used unless the application wishes to examine
12907 extensions in the OCSP response for example.
12909 Replace nonce routines with a pair of functions.
12910 OCSP_request_add1_nonce() adds a nonce value and optionally
12911 generates a random value. OCSP_check_nonce() checks the
12912 validity of the nonce in an OCSP response.
12916 * Change function OCSP_request_add() to OCSP_request_add0_id().
12917 This doesn't copy the supplied OCSP_CERTID and avoids the
12918 need to free up the newly created id. Change return type
12919 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12920 This can then be used to add extensions to the request.
12921 Deleted OCSP_request_new(), since most of its functionality
12922 is now in OCSP_REQUEST_new() (and the case insensitive name
12923 clash) apart from the ability to set the request name which
12924 will be added elsewhere.
12928 * Update OCSP API. Remove obsolete extensions argument from
12929 various functions. Extensions are now handled using the new
12930 OCSP extension code. New simple OCSP HTTP function which
12931 can be used to send requests and parse the response.
12935 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12936 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12937 uses the special reorder version of SET OF to sort the attributes
12938 and reorder them to match the encoded order. This resolves a long
12939 standing problem: a verify on a PKCS7 structure just after signing
12940 it used to fail because the attribute order did not match the
12941 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12942 it uses the received order. This is necessary to tolerate some broken
12943 software that does not order SET OF. This is handled by encoding
12944 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12945 to produce the required SET OF.
12949 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12950 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12951 files to get correct declarations of the ASN.1 item variables.
12955 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12956 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12957 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12958 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12959 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12960 ASN1_ITEM and no wrapper functions.
12964 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12965 replace the old function pointer based I/O routines. Change most of
12966 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12970 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12971 lines, recognize more "algorithms" that can be deselected, and make
12972 it complain about algorithm deselection that isn't recognised.
12976 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12977 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12978 to use new functions. Add NO_ASN1_OLD which can be set to remove
12979 some old style ASN1 functions: this can be used to determine if old
12980 code will still work when these eventually go away.
12984 * New extension functions for OCSP structures, these follow the
12985 same conventions as certificates and CRLs.
12989 * New function X509V3_add1_i2d(). This automatically encodes and
12990 adds an extension. Its behaviour can be customised with various
12991 flags to append, replace or delete. Various wrappers added for
12992 certificates and CRLs.
12996 * Fix to avoid calling the underlying ASN1 print routine when
12997 an extension cannot be parsed. Correct a typo in the
12998 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13002 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13003 entries for variables.
13007 * Add functionality to apps/openssl.c for detecting locking
13008 problems: As the program is single-threaded, all we have
13009 to do is register a locking callback using an array for
13010 storing which locks are currently held by the program.
13014 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13015 SSL_get_ex_data_X509_STORE_idx(), which is used in
13016 ssl_verify_cert_chain() and thus can be called at any time
13017 during TLS/SSL handshakes so that thread-safety is essential.
13018 Unfortunately, the ex_data design is not at all suited
13019 for multi-threaded use, so it probably should be abolished.
13023 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13025 *Broadcom, tweaked and integrated by Geoff Thorpe*
13027 * Move common extension printing code to new function
13028 X509V3_print_extensions(). Reorganise OCSP print routines and
13029 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13033 * New function X509_signature_print() to remove duplication in some
13038 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13039 set (this was treated exactly the same as SET OF previously). This
13040 is used to reorder the STACK representing the structure to match the
13041 encoding. This will be used to get round a problem where a PKCS7
13042 structure which was signed could not be verified because the STACK
13043 order did not reflect the encoded order.
13047 * Reimplement the OCSP ASN1 module using the new code.
13051 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13052 for its ASN1 operations. The old style function pointers still exist
13053 for now but they will eventually go away.
13057 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13058 completely replaces the old ASN1 functionality with a table driven
13059 encoder and decoder which interprets an ASN1_ITEM structure describing
13060 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13061 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13062 has also been converted to the new form.
13066 * Change BN_mod_exp_recp so that negative moduli are tolerated
13067 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13068 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13069 for negative moduli.
13073 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13074 of not touching the result's sign bit.
13078 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13083 * Changed the LHASH code to use prototypes for callbacks, and created
13084 macros to declare and implement thin (optionally static) functions
13085 that provide type-safety and avoid function pointer casting for the
13086 type-specific callbacks.
13090 * Added Kerberos Cipher Suites to be used with TLS, as written in
13092 *Veers Staats <staatsvr@asc.hpc.mil>,
13093 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13095 * Reformat the FAQ so the different questions and answers can be divided
13096 in sections depending on the subject.
13100 * Have the zlib compression code load ZLIB.DLL dynamically under
13105 * New function BN_mod_sqrt for computing square roots modulo a prime
13106 (using the probabilistic Tonelli-Shanks algorithm unless
13107 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13108 be handled deterministically).
13110 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13112 * Make BN_mod_inverse faster by explicitly handling small quotients
13113 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13114 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13118 * New function BN_kronecker.
13122 * Fix BN_gcd so that it works on negative inputs; the result is
13123 positive unless both parameters are zero.
13124 Previously something reasonably close to an infinite loop was
13125 possible because numbers could be growing instead of shrinking
13126 in the implementation of Euclid's algorithm.
13130 * Fix BN_is_word() and BN_is_one() macros to take into account the
13131 sign of the number in question.
13133 Fix BN_is_word(a,w) to work correctly for w == 0.
13135 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13136 because its test if the absolute value of 'a' equals 'w'.
13137 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13138 it exists mostly for use in the implementations of BN_is_zero(),
13139 BN_is_one(), and BN_is_word().
13143 * New function BN_swap.
13147 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13148 the exponentiation functions are more likely to produce reasonable
13149 results on negative inputs.
13153 * Change BN_mod_mul so that the result is always non-negative.
13154 Previously, it could be negative if one of the factors was negative;
13155 I don't think anyone really wanted that behaviour.
13159 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13160 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13161 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13162 and add new functions:
13171 BN_mod_lshift1_quick
13173 BN_mod_lshift_quick
13175 These functions always generate non-negative results.
13177 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13178 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13180 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13181 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13182 be reduced modulo `m`.
13184 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13187 The following entry accidentally appeared in the CHANGES file
13188 distributed with OpenSSL 0.9.7. The modifications described in
13189 it do *not* apply to OpenSSL 0.9.7.
13191 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13192 was actually never needed) and in BN_mul(). The removal in BN_mul()
13193 required a small change in bn_mul_part_recursive() and the addition
13194 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13195 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13196 bn_sub_words() and bn_add_words() except they take arrays with
13202 * In 'openssl passwd', verify passwords read from the terminal
13203 unless the '-salt' option is used (which usually means that
13204 verification would just waste user's time since the resulting
13205 hash is going to be compared with some given password hash)
13206 or the new '-noverify' option is used.
13208 This is an incompatible change, but it does not affect
13209 non-interactive use of 'openssl passwd' (passwords on the command
13210 line, '-stdin' option, '-in ...' option) and thus should not
13211 cause any problems.
13215 * Remove all references to RSAref, since there's no more need for it.
13219 * Make DSO load along a path given through an environment variable
13220 (SHLIB_PATH) with shl_load().
13224 * Constify the ENGINE code as a result of BIGNUM constification.
13225 Also constify the RSA code and most things related to it. In a
13226 few places, most notable in the depth of the ASN.1 code, ugly
13227 casts back to non-const were required (to be solved at a later
13232 * Make it so the openssl application has all engines loaded by default.
13236 * Constify the BIGNUM routines a little more.
13240 * Add the following functions:
13242 ENGINE_load_cswift()
13244 ENGINE_load_atalla()
13245 ENGINE_load_nuron()
13246 ENGINE_load_builtin_engines()
13248 That way, an application can itself choose if external engines that
13249 are built-in in OpenSSL shall ever be used or not. The benefit is
13250 that applications won't have to be linked with libdl or other dso
13251 libraries unless it's really needed.
13253 Changed 'openssl engine' to load all engines on demand.
13254 Changed the engine header files to avoid the duplication of some
13255 declarations (they differed!).
13259 * 'openssl engine' can now list capabilities.
13263 * Better error reporting in 'openssl engine'.
13267 * Never call load_dh_param(NULL) in s_server.
13271 * Add engine application. It can currently list engines by name and
13272 identity, and test if they are actually available.
13276 * Improve RPM specification file by forcing symbolic linking and making
13277 sure the installed documentation is also owned by root.root.
13279 *Damien Miller <djm@mindrot.org>*
13281 * Give the OpenSSL applications more possibilities to make use of
13282 keys (public as well as private) handled by engines.
13286 * Add OCSP code that comes from CertCo.
13290 * Add VMS support for the Rijndael code.
13294 * Added untested support for Nuron crypto accelerator.
13298 * Add support for external cryptographic devices. This code was
13299 previously distributed separately as the "engine" branch.
13301 *Geoff Thorpe, Richard Levitte*
13303 * Rework the filename-translation in the DSO code. It is now possible to
13304 have far greater control over how a "name" is turned into a filename
13305 depending on the operating environment and any oddities about the
13306 different shared library filenames on each system.
13310 * Support threads on FreeBSD-elf in Configure.
13314 * Fix for SHA1 assembly problem with MASM: it produces
13315 warnings about corrupt line number information when assembling
13316 with debugging information. This is caused by the overlapping
13319 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13322 NCONF_get_number() has no error checking at all. As a replacement,
13323 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13324 promoted strongly. The old NCONF_get_number is kept around for
13325 binary backward compatibility.
13326 Make it possible for methods to load from something other than a BIO,
13327 by providing a function pointer that is given a name instead of a BIO.
13328 For example, this could be used to load configuration data from an
13333 * Fix for non blocking accept BIOs. Added new I/O special reason
13334 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13335 with non blocking I/O was not possible because no retry code was
13336 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13341 * Added the beginnings of Rijndael support.
13345 * Fix for bug in DirectoryString mask setting. Add support for
13346 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13347 to allow certificate printing to more controllable, additional
13348 'certopt' option to 'x509' to allow new printing options to be
13353 * Clean old EAY MD5 hack from e_os.h.
13357 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13359 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13360 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13362 *Joe Orton, Steve Henson*
13364 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13366 * Fix additional bug revealed by the NISCC test suite:
13368 Stop bug triggering large recursion when presented with
13369 certain ASN.1 tags ([CVE-2003-0851])
13373 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13375 * Fix various bugs revealed by running the NISCC test suite:
13377 Stop out of bounds reads in the ASN1 code when presented with
13378 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13380 If verify callback ignores invalid public key errors don't try to check
13381 certificate signature with the NULL public key.
13385 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13386 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13391 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13392 extra data after the compression methods not only for TLS 1.0
13393 but also for SSL 3.0 (as required by the specification).
13395 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13397 * Change X509_certificate_type() to mark the key as exported/exportable
13398 when it's 512 *bits* long, not 512 bytes.
13402 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13404 * Countermeasure against the Klima-Pokorny-Rosa extension of
13405 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13406 a protocol version number mismatch like a decryption error
13407 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13411 * Turn on RSA blinding by default in the default implementation
13412 to avoid a timing attack. Applications that don't want it can call
13413 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13414 They would be ill-advised to do so in most cases.
13416 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13418 * Change RSA blinding code so that it works when the PRNG is not
13419 seeded (in this case, the secret RSA exponent is abused as
13420 an unpredictable seed -- if it is not unpredictable, there
13421 is no point in blinding anyway). Make RSA blinding thread-safe
13422 by remembering the creator's thread ID in rsa->blinding and
13423 having all other threads use local one-time blinding factors
13424 (this requires more computation than sharing rsa->blinding, but
13425 avoids excessive locking; and if an RSA object is not shared
13426 between threads, blinding will still be very fast).
13430 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13432 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13433 via timing by performing a MAC computation even if incorrect
13434 block cipher padding has been found. This is a countermeasure
13435 against active attacks where the attacker has to distinguish
13436 between bad padding and a MAC verification error. ([CVE-2003-0078])
13438 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13439 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13440 Martin Vuagnoux (EPFL, Ilion)*
13442 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13444 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13445 memory from its contents. This is done with a counter that will
13446 place alternating values in each byte. This can be used to solve
13447 two issues: 1) the removal of calls to memset() by highly optimizing
13448 compilers, and 2) cleansing with other values than 0, since those can
13449 be read through on certain media, for example a swap space on disk.
13453 * Bugfix: client side session caching did not work with external caching,
13454 because the session->cipher setting was not restored when reloading
13455 from the external cache. This problem was masked, when
13456 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13457 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13461 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13462 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13464 *Zeev Lieber <zeev-l@yahoo.com>*
13466 * Undo an undocumented change introduced in 0.9.6e which caused
13467 repeated calls to OpenSSL_add_all_ciphers() and
13468 OpenSSL_add_all_digests() to be ignored, even after calling
13473 * Change the default configuration reader to deal with last line not
13474 being properly terminated.
13478 * Change X509_NAME_cmp() so it applies the special rules on handling
13479 DN values that are of type PrintableString, as well as RDNs of type
13480 emailAddress where the value has the type ia5String.
13482 *stefank@valicert.com via Richard Levitte*
13484 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13485 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13486 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13487 the bitwise-OR of the two for use by the majority of applications
13488 wanting this behaviour, and update the docs. The documented
13489 behaviour and actual behaviour were inconsistent and had been
13490 changing anyway, so this is more a bug-fix than a behavioural
13493 *Geoff Thorpe, diagnosed by Nadav Har'El*
13495 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13496 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13500 * Fix initialization code race conditions in
13501 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13502 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13503 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13504 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13505 ssl2_get_cipher_by_char(),
13506 ssl3_get_cipher_by_char().
13508 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13510 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13511 the cached sessions are flushed, as the remove_cb() might use ex_data
13512 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13513 (see [openssl.org #212]).
13515 *Geoff Thorpe, Lutz Jaenicke*
13517 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13518 length, instead of the encoding length to d2i_ASN1_OBJECT.
13522 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13524 * [In 0.9.6g-engine release:]
13525 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13527 *Lynn Gazis <lgazis@rainbow.com>*
13529 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13531 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13532 and get fix the header length calculation.
13533 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13534 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13536 * Use proper error handling instead of 'assertions' in buffer
13537 overflow checks added in 0.9.6e. This prevents DoS (the
13538 assertions could call abort()).
13540 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13542 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13544 * Add various sanity checks to asn1_get_length() to reject
13545 the ASN1 length bytes if they exceed sizeof(long), will appear
13546 negative or the content length exceeds the length of the
13549 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13551 * Fix cipher selection routines: ciphers without encryption had no flags
13552 for the cipher strength set and where therefore not handled correctly
13553 by the selection routines (PR #130).
13557 * Fix EVP_dsa_sha macro.
13562 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13563 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13564 that was added in OpenSSL 0.9.6d.
13566 As the countermeasure turned out to be incompatible with some
13567 broken SSL implementations, the new option is part of SSL_OP_ALL.
13568 SSL_OP_ALL is usually employed when compatibility with weird SSL
13569 implementations is desired (e.g. '-bugs' option to 's_client' and
13570 's_server'), so the new option is automatically set in many
13575 * Changes in security patch:
13577 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13578 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13579 Air Force Materiel Command, USAF, under agreement number
13582 * Add various sanity checks to asn1_get_length() to reject
13583 the ASN1 length bytes if they exceed sizeof(long), will appear
13584 negative or the content length exceeds the length of the
13585 supplied buffer. ([CVE-2002-0659])
13587 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13589 * Assertions for various potential buffer overflows, not known to
13590 happen in practice.
13592 *Ben Laurie (CHATS)*
13594 * Various temporary buffers to hold ASCII versions of integers were
13595 too small for 64 bit platforms. ([CVE-2002-0655])
13596 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13598 * Remote buffer overflow in SSL3 protocol - an attacker could
13599 supply an oversized session ID to a client. ([CVE-2002-0656])
13601 *Ben Laurie (CHATS)*
13603 * Remote buffer overflow in SSL2 protocol - an attacker could
13604 supply an oversized client master key. ([CVE-2002-0656])
13606 *Ben Laurie (CHATS)*
13608 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13610 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13611 encoded as NULL) with id-dsa-with-sha1.
13613 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13615 * Check various `X509_...()` return values in apps/req.c.
13617 *Nils Larsch <nla@trustcenter.de>*
13619 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13620 an end-of-file condition would erroneously be flagged, when the CRLF
13621 was just at the end of a processed block. The bug was discovered when
13622 processing data through a buffering memory BIO handing the data to a
13623 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13624 <ptsekov@syntrex.com> and Nedelcho Stanev.
13628 * Implement a countermeasure against a vulnerability recently found
13629 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13630 before application data chunks to avoid the use of known IVs
13631 with data potentially chosen by the attacker.
13635 * Fix length checks in ssl3_get_client_hello().
13639 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13640 to prevent ssl3_read_internal() from incorrectly assuming that
13641 ssl3_read_bytes() found application data while handshake
13642 processing was enabled when in fact s->s3->in_read_app_data was
13643 merely automatically cleared during the initial handshake.
13645 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13647 * Fix object definitions for Private and Enterprise: they were not
13648 recognized in their shortname (=lowercase) representation. Extend
13649 obj_dat.pl to issue an error when using undefined keywords instead
13650 of silently ignoring the problem (Svenning Sorensen
13651 <sss@sss.dnsalias.net>).
13655 * Fix DH_generate_parameters() so that it works for 'non-standard'
13656 generators, i.e. generators other than 2 and 5. (Previously, the
13657 code did not properly initialise the 'add' and 'rem' values to
13658 BN_generate_prime().)
13660 In the new general case, we do not insist that 'generator' is
13661 actually a primitive root: This requirement is rather pointless;
13662 a generator of the order-q subgroup is just as good, if not
13667 * Map new X509 verification errors to alerts. Discovered and submitted by
13668 Tom Wu <tom@arcot.com>.
13672 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13673 returning non-zero before the data has been completely received
13674 when using non-blocking I/O.
13676 *Bodo Moeller; problem pointed out by John Hughes*
13678 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13680 *Ben Laurie, Lutz Jaenicke*
13682 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13683 Yoram Zahavi <YoramZ@gilian.com>).
13687 * Add information about CygWin 1.3 and on, and preserve proper
13688 configuration for the versions before that.
13690 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13692 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13693 check whether we deal with a copy of a session and do not delete from
13694 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13695 <izhar@checkpoint.com>.
13699 * Do not store session data into the internal session cache, if it
13700 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13701 flag is set). Proposed by Aslam <aslam@funk.com>.
13705 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13710 * [In 0.9.6d-engine release:]
13711 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13713 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13715 * Add the configuration target linux-s390x.
13717 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13719 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13720 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13721 variable as an indication that a ClientHello message has been
13722 received. As the flag value will be lost between multiple
13723 invocations of ssl3_accept when using non-blocking I/O, the
13724 function may not be aware that a handshake has actually taken
13725 place, thus preventing a new session from being added to the
13728 To avoid this problem, we now set s->new_session to 2 instead of
13729 using a local variable.
13731 *Lutz Jaenicke, Bodo Moeller*
13733 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13734 if the SSL_R_LENGTH_MISMATCH error is detected.
13736 *Geoff Thorpe, Bodo Moeller*
13738 * New 'shared_ldflag' column in Configure platform table.
13742 * Fix EVP_CIPHER_mode macro.
13744 *"Dan S. Camper" <dan@bti.net>*
13746 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13747 type, we must throw them away by setting rr->length to 0.
13749 *D P Chang <dpc@qualys.com>*
13751 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13753 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13754 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13755 worked incorrectly for those cases where range = `10..._2` and
13756 `3*range` is two bits longer than range.)
13760 * Only add signing time to PKCS7 structures if it is not already
13765 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13766 OBJ_ld_ce should be OBJ_id_ce.
13767 Also some ip-pda OIDs in crypto/objects/objects.txt were
13768 incorrect (cf. RFC 3039).
13770 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13772 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13773 returns early because it has nothing to do.
13775 *Andy Schneider <andy.schneider@bjss.co.uk>*
13777 * [In 0.9.6c-engine release:]
13778 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13780 *Andy Schneider <andy.schneider@bjss.co.uk>*
13782 * [In 0.9.6c-engine release:]
13783 Add support for Cryptographic Appliance's keyserver technology.
13784 (Use engine 'keyclient')
13786 *Cryptographic Appliances and Geoff Thorpe*
13788 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13789 is called via tools/c89.sh because arguments have to be
13790 rearranged (all '-L' options must appear before the first object
13793 *Richard Shapiro <rshapiro@abinitio.com>*
13795 * [In 0.9.6c-engine release:]
13796 Add support for Broadcom crypto accelerator cards, backported
13799 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13801 * [In 0.9.6c-engine release:]
13802 Add support for SureWare crypto accelerator cards from
13803 Baltimore Technologies. (Use engine 'sureware')
13805 *Baltimore Technologies and Mark Cox*
13807 * [In 0.9.6c-engine release:]
13808 Add support for crypto accelerator cards from Accelerated
13809 Encryption Processing, www.aep.ie. (Use engine 'aep')
13811 *AEP Inc. and Mark Cox*
13813 * Add a configuration entry for gcc on UnixWare.
13815 *Gary Benson <gbenson@redhat.com>*
13817 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13818 messages are stored in a single piece (fixed-length part and
13819 variable-length part combined) and fix various bugs found on the way.
13823 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13824 instead. BIO_gethostbyname() does not know what timeouts are
13825 appropriate, so entries would stay in cache even when they have
13827 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13829 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13830 faced with a pathologically small ClientHello fragment that does
13831 not contain client_version: Instead of aborting with an error,
13832 simply choose the highest available protocol version (i.e.,
13833 TLS 1.0 unless it is disabled). In practice, ClientHello
13834 messages are never sent like this, but this change gives us
13835 strictly correct behaviour at least for TLS.
13839 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13840 never resets s->method to s->ctx->method when called from within
13841 one of the SSL handshake functions.
13843 *Bodo Moeller; problem pointed out by Niko Baric*
13845 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13846 (sent using the client's version number) if client_version is
13847 smaller than the protocol version in use. Also change
13848 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13849 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13850 the client will at least see that alert.
13854 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13859 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13860 client receives HelloRequest while in a handshake.
13862 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13864 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13865 should end in 'break', not 'goto end' which circumvents various
13866 cleanups done in state SSL_ST_OK. But session related stuff
13867 must be disabled for SSL_ST_OK in the case that we just sent a
13870 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13871 before just sending a HelloRequest.
13873 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13875 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13876 reveal whether illegal block cipher padding was found or a MAC
13877 verification error occurred. (Neither SSLerr() codes nor alerts
13878 are directly visible to potential attackers, but the information
13879 may leak via logfiles.)
13881 Similar changes are not required for the SSL 2.0 implementation
13882 because the number of padding bytes is sent in clear for SSL 2.0,
13883 and the extra bytes are just ignored. However ssl/s2_pkt.c
13884 failed to verify that the purported number of padding bytes is in
13889 * Add OpenUNIX-8 support including shared libraries
13890 (Boyd Lynn Gerber <gerberb@zenez.com>).
13894 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13895 'wristwatch attack' using huge encoding parameters (cf.
13896 James H. Manger's CRYPTO 2001 paper). Note that the
13897 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13898 encoding parameters and hence was not vulnerable.
13902 * BN_sqr() bug fix.
13904 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13906 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13907 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13908 followed by modular reduction.
13910 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13912 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13913 equivalent based on BN_pseudo_rand() instead of BN_rand().
13917 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13918 This function was broken, as the check for a new client hello message
13919 to handle SGC did not allow these large messages.
13920 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13924 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13928 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13929 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13933 * Rework the configuration and shared library support for Tru64 Unix.
13934 The configuration part makes use of modern compiler features and
13935 still retains old compiler behavior for those that run older versions
13936 of the OS. The shared library support part includes a variant that
13937 uses the RPATH feature, and is available through the special
13938 configuration target "alpha-cc-rpath", which will never be selected
13941 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13943 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13944 with the same message size as in ssl3_get_certificate_request().
13945 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13946 messages might inadvertently be reject as too long.
13948 *Petr Lampa <lampa@fee.vutbr.cz>*
13950 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13954 * Modified SSL library such that the verify_callback that has been set
13955 specifically for an SSL object with SSL_set_verify() is actually being
13956 used. Before the change, a verify_callback set with this function was
13957 ignored and the verify_callback() set in the SSL_CTX at the time of
13958 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13959 to allow the necessary settings.
13963 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13964 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13965 done automatically (in contradiction to the requirements of the C
13966 standard). This made problems when used from OpenSSH.
13970 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13971 dh->length and always used
13973 BN_rand_range(priv_key, dh->p).
13975 BN_rand_range() is not necessary for Diffie-Hellman, and this
13976 specific range makes Diffie-Hellman unnecessarily inefficient if
13977 dh->length (recommended exponent length) is much smaller than the
13978 length of dh->p. We could use BN_rand_range() if the order of
13979 the subgroup was stored in the DH structure, but we only have
13984 BN_rand(priv_key, l, ...)
13986 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13993 RSA_eay_public_encrypt
13994 RSA_eay_private_decrypt
13995 RSA_eay_private_encrypt (signing)
13996 RSA_eay_public_decrypt (signature verification)
13998 (default implementations for RSA_public_encrypt,
13999 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14000 always reject numbers >= n.
14004 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14005 to synchronize access to 'locking_thread'. This is necessary on
14006 systems where access to 'locking_thread' (an 'unsigned long'
14007 variable) is not atomic.
14011 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14012 *before* setting the 'crypto_lock_rand' flag. The previous code had
14013 a race condition if 0 is a valid thread ID.
14015 *Travis Vitek <vitek@roguewave.com>*
14017 * Add support for shared libraries under Irix.
14019 *Albert Chin-A-Young <china@thewrittenword.com>*
14021 * Add configuration option to build on Linux on both big-endian and
14022 little-endian MIPS.
14024 *Ralf Baechle <ralf@uni-koblenz.de>*
14026 * Add the possibility to create shared libraries on HP-UX.
14030 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14032 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14033 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14034 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14035 PRNG state recovery was possible based on the output of
14036 one PRNG request appropriately sized to gain knowledge on
14037 'md' followed by enough consecutive 1-byte PRNG requests
14038 to traverse all of 'state'.
14040 1. When updating 'md_local' (the current thread's copy of 'md')
14041 during PRNG output generation, hash all of the previous
14042 'md_local' value, not just the half used for PRNG output.
14044 2. Make the number of bytes from 'state' included into the hash
14045 independent from the number of PRNG bytes requested.
14047 The first measure alone would be sufficient to avoid
14048 Markku-Juhani's attack. (Actually it had never occurred
14049 to me that the half of 'md_local' used for chaining was the
14050 half from which PRNG output bytes were taken -- I had always
14051 assumed that the secret half would be used.) The second
14052 measure makes sure that additional data from 'state' is never
14053 mixed into 'md_local' in small portions; this heuristically
14054 further strengthens the PRNG.
14058 * Fix crypto/bn/asm/mips3.s.
14062 * When only the key is given to "enc", the IV is undefined. Print out
14063 an error message in this case.
14067 * Handle special case when X509_NAME is empty in X509 printing routines.
14071 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14072 positive and less than q.
14076 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14077 used: it isn't thread safe and the add_lock_callback should handle
14080 *Paul Rose <Paul.Rose@bridge.com>*
14082 * Verify that incoming data obeys the block size in
14083 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14089 *Ulf Möller, Bodo Möller*
14091 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14092 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14093 when fixing the server behaviour for backwards-compatible 'client
14094 hello' messages. (Note that the attack is impractical against
14095 SSL 3.0 and TLS 1.0 anyway because length and version checking
14096 means that the probability of guessing a valid ciphertext is
14097 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14100 Before 0.9.5, the countermeasure (hide the error by generating a
14101 random 'decryption result') did not work properly because
14102 ERR_clear_error() was missing, meaning that SSL_get_error() would
14103 detect the supposedly ignored error.
14105 Both problems are now fixed.
14109 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14110 (previously it was 1024).
14114 * Fix for compatibility mode trust settings: ignore trust settings
14115 unless some valid trust or reject settings are present.
14119 * Fix for blowfish EVP: its a variable length cipher.
14123 * Fix various bugs related to DSA S/MIME verification. Handle missing
14124 parameters in DSA public key structures and return an error in the
14125 DSA routines if parameters are absent.
14129 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14130 in the current directory if neither $RANDFILE nor $HOME was set.
14131 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14132 caused some confusion to Windows users who haven't defined $HOME.
14133 Thus RAND_file_name() is changed again: e_os.h can define a
14134 DEFAULT_HOME, which will be used if $HOME is not set.
14135 For Windows, we use "C:"; on other platforms, we still require
14136 environment variables.
14138 * Move 'if (!initialized) RAND_poll()' into regions protected by
14139 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14140 having multiple threads call RAND_poll() concurrently.
14144 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14145 combination of a flag and a thread ID variable.
14146 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14147 flag), *other* threads can enter ssleay_add_bytes without obeying
14148 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14149 that they do not hold after the first thread unsets add_do_not_lock).
14153 * Change bctest again: '-x' expressions are not available in all
14154 versions of 'test'.
14158 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14160 * Fix a couple of memory leaks in PKCS7_dataDecode()
14162 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14164 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14165 the default extension for executables, if any. Also, make the perl
14166 scripts that use symlink() to test if it really exists and use "cp"
14167 if it doesn't. All this made OpenSSL compilable and installable in
14172 * Fix for asn1_GetSequence() for indefinite length constructed data.
14173 If SEQUENCE is length is indefinite just set c->slen to the total
14174 amount of data available.
14176 *Steve Henson, reported by shige@FreeBSD.org*
14178 *This change does not apply to 0.9.7.*
14180 * Change bctest to avoid here-documents inside command substitution
14181 (workaround for FreeBSD /bin/sh bug).
14182 For compatibility with Ultrix, avoid shell functions (introduced
14183 in the bctest version that searches along $PATH).
14187 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14188 with des_encrypt() defined on some operating systems, like Solaris
14193 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14194 On the Importance of Eliminating Errors in Cryptographic
14195 Computations, J. Cryptology 14 (2001) 2, 101-119,
14196 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14200 * MIPS assembler BIGNUM division bug fix.
14204 * Disabled incorrect Alpha assembler code.
14208 * Fix PKCS#7 decode routines so they correctly update the length
14209 after reading an EOC for the EXPLICIT tag.
14213 *This change does not apply to 0.9.7.*
14215 * Fix bug in PKCS#12 key generation routines. This was triggered
14216 if a 3DES key was generated with a 0 initial byte. Include
14217 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14218 (but broken) behaviour.
14222 * Enhance bctest to search for a working bc along $PATH and print
14225 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14227 * Fix memory leaks in err.c: free err_data string if necessary;
14228 don't write to the wrong index in ERR_set_error_data.
14232 * Implement ssl23_peek (analogous to ssl23_read), which previously
14237 * Replace rdtsc with `_emit` statements for VC++ version 5.
14239 *Jeremy Cooper <jeremy@baymoo.org>*
14241 * Make it possible to reuse SSLv2 sessions.
14245 * In copy_email() check for >= 0 as a return value for
14246 X509_NAME_get_index_by_NID() since 0 is a valid index.
14248 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14250 * Avoid coredump with unsupported or invalid public keys by checking if
14251 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14252 PKCS7_verify() fails with non detached data.
14256 * Don't use getenv in library functions when run as setuid/setgid.
14257 New function OPENSSL_issetugid().
14261 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14262 due to incorrect handling of multi-threading:
14264 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14266 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14268 3. Count how many times MemCheck_off() has been called so that
14269 nested use can be treated correctly. This also avoids
14270 inband-signalling in the previous code (which relied on the
14271 assumption that thread ID 0 is impossible).
14275 * Add "-rand" option also to s_client and s_server.
14279 * Fix CPU detection on Irix 6.x.
14280 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14281 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14283 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14288 *This change does not apply to 0.9.7.*
14290 * Use the cached encoding of an X509_NAME structure rather than
14291 copying it. This is apparently the reason for the libsafe "errors"
14292 but the code is actually correct.
14296 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14297 Bleichenbacher's DSA attack.
14298 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14299 to be set and top=0 forces the highest bit to be set; top=-1 is new
14300 and leaves the highest bit random.
14302 *Ulf Moeller, Bodo Moeller*
14304 * In the `NCONF_...`-based implementations for `CONF_...` queries
14305 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14306 a temporary CONF structure with the data component set to NULL
14307 (which gives segmentation faults in lh_retrieve).
14308 Instead, use NULL for the CONF pointer in CONF_get_string and
14309 CONF_get_number (which may use environment variables) and directly
14310 return NULL from CONF_get_section.
14314 * Fix potential buffer overrun for EBCDIC.
14318 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14319 keyUsage if basicConstraints absent for a CA.
14323 * Make SMIME_write_PKCS7() write mail header values with a format that
14324 is more generally accepted (no spaces before the semicolon), since
14325 some programs can't parse those values properly otherwise. Also make
14326 sure BIO's that break lines after each write do not create invalid
14331 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14332 macros previously used would not encode an empty SEQUENCE OF
14333 and break the signature.
14337 *This change does not apply to 0.9.7.*
14339 * Zero the premaster secret after deriving the master secret in
14344 * Add some EVP_add_digest_alias registrations (as found in
14345 OpenSSL_add_all_digests()) to SSL_library_init()
14346 aka OpenSSL_add_ssl_algorithms(). This provides improved
14347 compatibility with peers using X.509 certificates
14348 with unconventional AlgorithmIdentifier OIDs.
14352 * Fix for Irix with NO_ASM.
14354 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14356 * ./config script fixes.
14358 *Ulf Moeller, Richard Levitte*
14360 * Fix 'openssl passwd -1'.
14364 * Change PKCS12_key_gen_asc() so it can cope with non null
14365 terminated strings whose length is passed in the passlen
14366 parameter, for example from PEM callbacks. This was done
14367 by adding an extra length parameter to asc2uni().
14369 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14371 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14372 call failed, free the DSA structure.
14376 * Fix to uni2asc() to cope with zero length Unicode strings.
14377 These are present in some PKCS#12 files.
14381 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14382 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14383 when writing a 32767 byte record.
14385 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14387 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14388 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14390 (RSA objects have a reference count access to which is protected
14391 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14392 so they are meant to be shared between threads.)
14393 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14394 "Reddie, Steven" <Steven.Reddie@ca.com>*
14396 * Fix a deadlock in CRYPTO_mem_leaks().
14400 * Use better test patterns in bntest.
14404 * rand_win.c fix for Borland C.
14408 * BN_rshift bugfix for n == 0.
14412 * Add a 'bctest' script that checks for some known 'bc' bugs
14413 so that 'make test' does not abort just because 'bc' is broken.
14417 * Store verify_result within SSL_SESSION also for client side to
14418 avoid potential security hole. (Re-used sessions on the client side
14419 always resulted in verify_result==X509_V_OK, not using the original
14420 result of the server certificate verification.)
14424 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14425 SSL3_RT_APPLICATION_DATA, return 0.
14426 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14431 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14432 releases, have been re-implemented by renaming the previous
14433 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14434 and ssl3_read_internal, respectively, and adding 'peek' parameters
14435 to them. The new ssl[23]_{read,peek} functions are calls to
14436 ssl[23]_read_internal with the 'peek' flag set appropriately.
14437 A 'peek' parameter has also been added to ssl3_read_bytes, which
14438 does the actual work for ssl3_read_internal.
14442 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14443 the method-specific "init()" handler. Also clean up ex_data after
14444 calling the method-specific "finish()" handler. Previously, this was
14445 happening the other way round.
14449 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14450 The previous value, 12, was not always sufficient for BN_mod_exp().
14454 * Make sure that shared libraries get the internal name engine with
14455 the full version number and not just 0. This should mark the
14456 shared libraries as not backward compatible. Of course, this should
14457 be changed again when we can guarantee backward binary compatibility.
14461 * Fix typo in get_cert_by_subject() in by_dir.c
14463 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14465 * Rework the system to generate shared libraries:
14467 - Make note of the expected extension for the shared libraries and
14468 if there is a need for symbolic links from for example libcrypto.so.0
14469 to libcrypto.so.0.9.7. There is extended info in Configure for
14472 - Make as few rebuilds of the shared libraries as possible.
14474 - Still avoid linking the OpenSSL programs with the shared libraries.
14476 - When installing, install the shared libraries separately from the
14481 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14483 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14484 and not in SSL_clear because the latter is also used by the
14485 accept/connect functions; previously, the settings made by
14486 SSL_set_read_ahead would be lost during the handshake.
14488 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14490 * Correct util/mkdef.pl to be selective about disabled algorithms.
14491 Previously, it would create entries for disabled algorithms no
14496 * Added several new manual pages for SSL_* function.
14500 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14502 * In ssl23_get_client_hello, generate an error message when faced
14503 with an initial SSL 3.0/TLS record that is too small to contain the
14504 first two bytes of the ClientHello message, i.e. client_version.
14505 (Note that this is a pathologic case that probably has never happened
14506 in real life.) The previous approach was to use the version number
14507 from the record header as a substitute; but our protocol choice
14508 should not depend on that one because it is not authenticated
14509 by the Finished messages.
14513 * More robust randomness gathering functions for Windows.
14515 *Jeffrey Altman <jaltman@columbia.edu>*
14517 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14518 not set then we don't setup the error code for issuer check errors
14519 to avoid possibly overwriting other errors which the callback does
14520 handle. If an application does set the flag then we assume it knows
14521 what it is doing and can handle the new informational codes
14526 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14527 a general "ANY" type, as such it should be able to decode anything
14528 including tagged types. However it didn't check the class so it would
14529 wrongly interpret tagged types in the same way as their universal
14530 counterpart and unknown types were just rejected. Changed so that the
14531 tagged and unknown types are handled in the same way as a SEQUENCE:
14532 that is the encoding is stored intact. There is also a new type
14533 "V_ASN1_OTHER" which is used when the class is not universal, in this
14534 case we have no idea what the actual type is so we just lump them all
14539 * On VMS, stdout may very well lead to a file that is written to
14540 in a record-oriented fashion. That means that every write() will
14541 write a separate record, which will be read separately by the
14542 programs trying to read from it. This can be very confusing.
14544 The solution is to put a BIO filter in the way that will buffer
14545 text until a linefeed is reached, and then write everything a
14546 line at a time, so every record written will be an actual line,
14547 not chunks of lines and not (usually doesn't happen, but I've
14548 seen it once) several lines in one record. BIO_f_linebuffer() is
14551 Currently, it's a VMS-only method, because that's where it has
14552 been tested well enough.
14556 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14557 it can return incorrect results.
14558 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14559 but it was in 0.9.6-beta[12].)
14563 * Disable the check for content being present when verifying detached
14564 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14565 include zero length content when signing messages.
14569 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14570 BIO_ctrl (for BIO pairs).
14574 * Add DSO method for VMS.
14578 * Bug fix: Montgomery multiplication could produce results with the
14583 * Add RPM specification openssl.spec and modify it to build three
14584 packages. The default package contains applications, application
14585 documentation and run-time libraries. The devel package contains
14586 include files, static libraries and function documentation. The
14587 doc package contains the contents of the doc directory. The original
14588 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14592 * Add a large number of documentation files for many SSL routines.
14594 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14596 * Add a configuration entry for Sony News 4.
14598 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14600 * Don't set the two most significant bits to one when generating a
14601 random number < q in the DSA library.
14605 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14606 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14607 the underlying transport is blocking) if a handshake took place.
14608 (The default behaviour is needed by applications such as s_client
14609 and s_server that use select() to determine when to use SSL_read;
14610 but for applications that know in advance when to expect data, it
14611 just makes things more complicated.)
14615 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14620 * Add a few more EBCDIC conditionals that make `req` and `x509`
14621 work better on such systems.
14623 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14625 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14626 Update PKCS12_parse() so it copies the friendlyName and the
14627 keyid to the certificates aux info.
14631 * Fix bug in PKCS7_verify() which caused an infinite loop
14632 if there was more than one signature.
14634 *Sven Uszpelkat <su@celocom.de>*
14636 * Major change in util/mkdef.pl to include extra information
14637 about each symbol, as well as presenting variables as well
14638 as functions. This change means that there's n more need
14639 to rebuild the .num files when some algorithms are excluded.
14643 * Allow the verify time to be set by an application,
14644 rather than always using the current time.
14648 * Phase 2 verify code reorganisation. The certificate
14649 verify code now looks up an issuer certificate by a
14650 number of criteria: subject name, authority key id
14651 and key usage. It also verifies self signed certificates
14652 by the same criteria. The main comparison function is
14653 X509_check_issued() which performs these checks.
14655 Lot of changes were necessary in order to support this
14656 without completely rewriting the lookup code.
14658 Authority and subject key identifier are now cached.
14660 The LHASH 'certs' is X509_STORE has now been replaced
14661 by a STACK_OF(X509_OBJECT). This is mainly because an
14662 LHASH can't store or retrieve multiple objects with
14663 the same hash value.
14665 As a result various functions (which were all internal
14666 use only) have changed to handle the new X509_STORE
14667 structure. This will break anything that messed round
14668 with X509_STORE internally.
14670 The functions X509_STORE_add_cert() now checks for an
14671 exact match, rather than just subject name.
14673 The X509_STORE API doesn't directly support the retrieval
14674 of multiple certificates matching a given criteria, however
14675 this can be worked round by performing a lookup first
14676 (which will fill the cache with candidate certificates)
14677 and then examining the cache for matches. This is probably
14678 the best we can do without throwing out X509_LOOKUP
14679 entirely (maybe later...).
14681 The X509_VERIFY_CTX structure has been enhanced considerably.
14683 All certificate lookup operations now go via a get_issuer()
14684 callback. Although this currently uses an X509_STORE it
14685 can be replaced by custom lookups. This is a simple way
14686 to bypass the X509_STORE hackery necessary to make this
14687 work and makes it possible to use more efficient techniques
14688 in future. A very simple version which uses a simple
14689 STACK for its trusted certificate store is also provided
14690 using X509_STORE_CTX_trusted_stack().
14692 The verify_cb() and verify() callbacks now have equivalents
14693 in the X509_STORE_CTX structure.
14695 X509_STORE_CTX also has a 'flags' field which can be used
14696 to customise the verify behaviour.
14700 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14701 excludes S/MIME capabilities.
14705 * When a certificate request is read in keep a copy of the
14706 original encoding of the signed data and use it when outputting
14707 again. Signatures then use the original encoding rather than
14708 a decoded, encoded version which may cause problems if the
14709 request is improperly encoded.
14713 * For consistency with other BIO_puts implementations, call
14714 buffer_write(b, ...) directly in buffer_puts instead of calling
14717 In BIO_puts, increment b->num_write as in BIO_write.
14719 *Peter.Sylvester@EdelWeb.fr*
14721 * Fix BN_mul_word for the case where the word is 0. (We have to use
14722 BN_zero, we may not return a BIGNUM with an array consisting of
14723 words set to zero.)
14727 * Avoid calling abort() from within the library when problems are
14728 detected, except if preprocessor symbols have been defined
14729 (such as REF_CHECK, BN_DEBUG etc.).
14733 * New openssl application 'rsautl'. This utility can be
14734 used for low level RSA operations. DER public key
14735 BIO/fp routines also added.
14739 * New Configure entry and patches for compiling on QNX 4.
14741 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14743 * A demo state-machine implementation was sponsored by
14744 Nuron (<http://www.nuron.com/>) and is now available in
14745 demos/state_machine.
14749 * New options added to the 'dgst' utility for signature
14750 generation and verification.
14754 * Unrecognized PKCS#7 content types are now handled via a
14755 catch all ASN1_TYPE structure. This allows unsupported
14756 types to be stored as a "blob" and an application can
14757 encode and decode it manually.
14761 * Fix various signed/unsigned issues to make a_strex.c
14762 compile under VC++.
14764 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14766 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14767 length if passed a buffer. ASN1_INTEGER_to_BN failed
14768 if passed a NULL BN and its argument was negative.
14770 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14772 * Modification to PKCS#7 encoding routines to output definite
14773 length encoding. Since currently the whole structures are in
14774 memory there's not real point in using indefinite length
14775 constructed encoding. However if OpenSSL is compiled with
14776 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14780 * Added BIO_vprintf() and BIO_vsnprintf().
14784 * Added more prefixes to parse for in the strings written
14785 through a logging bio, to cover all the levels that are available
14786 through syslog. The prefixes are now:
14788 PANIC, EMERG, EMR => LOG_EMERG
14789 ALERT, ALR => LOG_ALERT
14790 CRIT, CRI => LOG_CRIT
14791 ERROR, ERR => LOG_ERR
14792 WARNING, WARN, WAR => LOG_WARNING
14793 NOTICE, NOTE, NOT => LOG_NOTICE
14794 INFO, INF => LOG_INFO
14795 DEBUG, DBG => LOG_DEBUG
14797 and as before, if none of those prefixes are present at the
14798 beginning of the string, LOG_ERR is chosen.
14800 On Win32, the `LOG_*` levels are mapped according to this:
14802 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14803 LOG_WARNING => EVENTLOG_WARNING_TYPE
14804 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14808 * Made it possible to reconfigure with just the configuration
14809 argument "reconf" or "reconfigure". The command line arguments
14810 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14811 and are retrieved from there when reconfiguring.
14817 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14819 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14823 * The obj_dat.pl script was messing up the sorting of object
14824 names. The reason was that it compared the quoted version
14825 of strings as a result "OCSP" > "OCSP Signing" because
14826 " > SPACE. Changed script to store unquoted versions of
14827 names and add quotes on output. It was also omitting some
14828 names from the lookup table if they were given a default
14829 value (that is if SN is missing it is given the same
14830 value as LN and vice versa), these are now added on the
14831 grounds that if an object has a name we should be able to
14832 look it up. Finally added warning output when duplicate
14833 short or long names are found.
14837 * Changes needed for Tandem NSK.
14839 *Scott Uroff <scott@xypro.com>*
14841 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14842 RSA_padding_check_SSLv23(), special padding was never detected
14843 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14844 version rollback attacks was not effective.
14846 In s23_clnt.c, don't use special rollback-attack detection padding
14847 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14848 client; similarly, in s23_srvr.c, don't do the rollback check if
14849 SSL 2.0 is the only protocol enabled in the server.
14853 * Make it possible to get hexdumps of unprintable data with 'openssl
14854 asn1parse'. By implication, the functions ASN1_parse_dump() and
14855 BIO_dump_indent() are added.
14859 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14860 these print out strings and name structures based on various
14861 flags including RFC2253 support and proper handling of
14862 multibyte characters. Added options to the 'x509' utility
14863 to allow the various flags to be set.
14867 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14868 Also change the functions X509_cmp_current_time() and
14869 X509_gmtime_adj() work with an ASN1_TIME structure,
14870 this will enable certificates using GeneralizedTime in validity
14871 dates to be checked.
14875 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14876 negative public key encodings) on by default,
14877 NO_NEG_PUBKEY_BUG can be set to disable it.
14881 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14882 content octets. An i2c_ASN1_OBJECT is unnecessary because
14883 the encoding can be trivially obtained from the structure.
14887 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14888 not read locks (`CRYPTO_r_[un]lock`).
14892 * A first attempt at creating official support for shared
14893 libraries through configuration. I've kept it so the
14894 default is static libraries only, and the OpenSSL programs
14895 are always statically linked for now, but there are
14896 preparations for dynamic linking in place.
14897 This has been tested on Linux and Tru64.
14901 * Randomness polling function for Win9x, as described in:
14902 Peter Gutmann, Software Generation of Practically Strong
14907 * Fix so PRNG is seeded in req if using an already existing
14912 * New options to smime application. -inform and -outform
14913 allow alternative formats for the S/MIME message including
14914 PEM and DER. The -content option allows the content to be
14915 specified separately. This should allow things like Netscape
14916 form signing output easier to verify.
14920 * Fix the ASN1 encoding of tags using the 'long form'.
14924 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14925 STRING types. These convert content octets to and from the
14926 underlying type. The actual tag and length octets are
14927 already assumed to have been read in and checked. These
14928 are needed because all other string types have virtually
14929 identical handling apart from the tag. By having versions
14930 of the ASN1 functions that just operate on content octets
14931 IMPLICIT tagging can be handled properly. It also allows
14932 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14933 and ASN1_INTEGER are identical apart from the tag.
14937 * Change the handling of OID objects as follows:
14939 - New object identifiers are inserted in objects.txt, following
14940 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14941 - objects.pl is used to process obj_mac.num and create a new
14943 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14946 This is currently kind of a hack, and the perl code in objects.pl
14947 isn't very elegant, but it works as I intended. The simplest way
14948 to check that it worked correctly is to look in obj_dat.h and
14949 check the array nid_objs and make sure the objects haven't moved
14950 around (this is important!). Additions are OK, as well as
14951 consistent name changes.
14955 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14959 * Addition of the command line parameter '-rand file' to 'openssl req'.
14960 The given file adds to whatever has already been seeded into the
14961 random pool through the RANDFILE configuration file option or
14962 environment variable, or the default random state file.
14966 * mkstack.pl now sorts each macro group into lexical order.
14967 Previously the output order depended on the order the files
14968 appeared in the directory, resulting in needless rewriting
14973 * Patches to make OpenSSL compile under Win32 again. Mostly
14974 work arounds for the VC++ problem that it treats func() as
14975 func(void). Also stripped out the parts of mkdef.pl that
14976 added extra typesafe functions: these no longer exist.
14980 * Reorganisation of the stack code. The macros are now all
14981 collected in safestack.h . Each macro is defined in terms of
14982 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14983 DEBUG_SAFESTACK is now handled in terms of function casts,
14984 this has the advantage of retaining type safety without the
14985 use of additional functions. If DEBUG_SAFESTACK is not defined
14986 then the non typesafe macros are used instead. Also modified the
14987 mkstack.pl script to handle the new form. Needs testing to see
14988 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14989 the default if no major problems. Similar behaviour for ASN1_SET_OF
14990 and PKCS12_STACK_OF.
14994 * When some versions of IIS use the 'NET' form of private key the
14995 key derivation algorithm is different. Normally MD5(password) is
14996 used as a 128 bit RC4 key. In the modified case
14997 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14998 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14999 as the old Netscape_RSA functions except they have an additional
15000 'sgckey' parameter which uses the modified algorithm. Also added
15001 an -sgckey command line option to the rsa utility. Thanks to
15002 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15003 algorithm to openssl-dev.
15007 * The evp_local.h macros were using 'c.##kname' which resulted in
15008 invalid expansion on some systems (SCO 5.0.5 for example).
15009 Corrected to 'c.kname'.
15011 *Phillip Porch <root@theporch.com>*
15013 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15014 a STACK of email addresses from a certificate or request, these look
15015 in the subject name and the subject alternative name extensions and
15016 omit any duplicate addresses.
15020 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15021 This makes DSA verification about 2 % faster.
15025 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15026 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15027 plus overhead for 1024 bit moduli).
15028 This makes exponentiations about 0.5 % faster for 1024 bit
15029 exponents (as measured by "openssl speed rsa2048").
15033 * Rename memory handling macros to avoid conflicts with other
15035 Malloc => OPENSSL_malloc
15036 Malloc_locked => OPENSSL_malloc_locked
15037 Realloc => OPENSSL_realloc
15038 Free => OPENSSL_free
15042 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15043 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15047 * CygWin32 support.
15049 *John Jarvie <jjarvie@newsguy.com>*
15051 * The type-safe stack code has been rejigged. It is now only compiled
15052 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15053 by default all type-specific stack functions are "#define"d back to
15054 standard stack functions. This results in more streamlined output
15055 but retains the type-safety checking possibilities of the original
15060 * The STACK code has been cleaned up, and certain type declarations
15061 that didn't make a lot of sense have been brought in line. This has
15062 also involved a cleanup of sorts in safestack.h to more correctly
15063 map type-safe stack functions onto their plain stack counterparts.
15064 This work has also resulted in a variety of "const"ifications of
15065 lots of the code, especially `_cmp` operations which should normally
15066 be prototyped with "const" parameters anyway.
15070 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15071 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15072 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15073 where all of 'md' is used each time the PRNG is used, but 'state'
15074 is used only indexed by a cyclic counter. As entropy may not be
15075 well distributed from the beginning, 'md' is important as a
15076 chaining variable. However, the output function chains only half
15077 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15078 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15079 in all of 'state' being rewritten, with the new values depending
15080 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15084 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15085 the handshake is continued after ssl_verify_cert_chain();
15086 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15087 can lead to 'unexplainable' connection aborts later.
15089 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15091 * Major EVP API cipher revision.
15092 Add hooks for extra EVP features. This allows various cipher
15093 parameters to be set in the EVP interface. Support added for variable
15094 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15095 setting of RC2 and RC5 parameters.
15097 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15100 Remove lots of duplicated code from the EVP library. For example *every*
15101 cipher init() function handles the 'iv' in the same way according to the
15102 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15103 for CFB and OFB modes they zero ctx->num.
15105 New functionality allows removal of S/MIME code RC2 hack.
15107 Most of the routines have the same form and so can be declared in terms
15110 By shifting this to the top level EVP_CipherInit() it can be removed from
15111 all individual ciphers. If the cipher wants to handle IVs or keys
15112 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15115 Change lots of functions like EVP_EncryptUpdate() to now return a
15116 value: although software versions of the algorithms cannot fail
15117 any installed hardware versions can.
15121 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15122 this option is set, tolerate broken clients that send the negotiated
15123 protocol version number instead of the requested protocol version
15128 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15129 i.e. non-zero for export ciphersuites, zero otherwise.
15130 Previous versions had this flag inverted, inconsistent with
15131 rsa_tmp_cb (..._TMP_RSA_CB).
15133 *Bodo Moeller; problem reported by Amit Chopra*
15135 * Add missing DSA library text string. Work around for some IIS
15136 key files with invalid SEQUENCE encoding.
15140 * Add a document (doc/standards.txt) that list all kinds of standards
15141 and so on that are implemented in OpenSSL.
15145 * Enhance c_rehash script. Old version would mishandle certificates
15146 with the same subject name hash and wouldn't handle CRLs at all.
15147 Added -fingerprint option to crl utility, to support new c_rehash
15152 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15156 * Fix for SSL server purpose checking. Server checking was
15157 rejecting certificates which had extended key usage present
15158 but no ssl client purpose.
15160 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15162 * Make PKCS#12 code work with no password. The PKCS#12 spec
15163 is a little unclear about how a blank password is handled.
15164 Since the password in encoded as a BMPString with terminating
15165 double NULL a zero length password would end up as just the
15166 double NULL. However no password at all is different and is
15167 handled differently in the PKCS#12 key generation code. NS
15168 treats a blank password as zero length. MSIE treats it as no
15169 password on export: but it will try both on import. We now do
15170 the same: PKCS12_parse() tries zero length and no password if
15171 the password is set to "" or NULL (NULL is now a valid password:
15172 it wasn't before) as does the pkcs12 application.
15176 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15177 perror when PEM_read_bio_X509_REQ fails, the error message must
15178 be obtained from the error queue.
15182 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15183 it in ERR_remove_state if appropriate, and change ERR_get_state
15184 accordingly to avoid race conditions (this is necessary because
15185 thread_hash is no longer constant once set).
15189 * Bugfix for linux-elf makefile.one.
15193 * RSA_get_default_method() will now cause a default
15194 RSA_METHOD to be chosen if one doesn't exist already.
15195 Previously this was only set during a call to RSA_new()
15196 or RSA_new_method(NULL) meaning it was possible for
15197 RSA_get_default_method() to return NULL.
15201 * Added native name translation to the existing DSO code
15202 that will convert (if the flag to do so is set) filenames
15203 that are sufficiently small and have no path information
15204 into a canonical native form. Eg. "blah" converted to
15205 "libblah.so" or "blah.dll" etc.
15209 * New function ERR_error_string_n(e, buf, len) which is like
15210 ERR_error_string(e, buf), but writes at most 'len' bytes
15211 including the 0 terminator. For ERR_error_string_n, 'buf'
15214 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15216 * CONF library reworked to become more general. A new CONF
15217 configuration file reader "class" is implemented as well as a
15218 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15219 old `CONF_*` functions are still there, but are reimplemented to
15220 work in terms of the new functions. Also, a set of functions
15221 to handle the internal storage of the configuration data is
15222 provided to make it easier to write new configuration file
15223 reader "classes" (I can definitely see something reading a
15224 configuration file in XML format, for example), called `_CONF_*`,
15225 or "the configuration storage API"...
15227 The new configuration file reading functions are:
15229 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15230 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15232 NCONF_default, NCONF_WIN32
15234 NCONF_dump_fp, NCONF_dump_bio
15236 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15237 NCONF_new creates a new CONF object. This works in the same way
15238 as other interfaces in OpenSSL, like the BIO interface.
15239 `NCONF_dump_*` dump the internal storage of the configuration file,
15240 which is useful for debugging. All other functions take the same
15241 arguments as the old `CONF_*` functions with the exception of the
15242 first that must be a `CONF *` instead of a `LHASH *`.
15244 To make it easier to use the new classes with the old `CONF_*` functions,
15245 the function CONF_set_default_method is provided.
15249 * Add '-tls1' option to 'openssl ciphers', which was already
15250 mentioned in the documentation but had not been implemented.
15251 (This option is not yet really useful because even the additional
15252 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15256 * Initial DSO code added into libcrypto for letting OpenSSL (and
15257 OpenSSL-based applications) load shared libraries and bind to
15258 them in a portable way.
15260 *Geoff Thorpe, with contributions from Richard Levitte*
15262 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15264 * Make sure _lrotl and _lrotr are only used with MSVC.
15266 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15267 (the default implementation of RAND_status).
15269 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15270 to '-clrext' (= clear extensions), as intended and documented.
15271 *Bodo Moeller; inconsistency pointed out by Michael Attili
15272 <attili@amaxo.com>*
15274 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15275 was larger than the MD block size.
15277 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15279 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15280 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15281 using the passed key: if the passed key was a private key the result
15282 of X509_print(), for example, would be to print out all the private key
15287 * des_quad_cksum() byte order bug fix.
15288 *Ulf Möller, using the problem description in krb4-0.9.7, where
15289 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15291 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15294 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15296 * For easily testing in shell scripts whether some command
15297 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15298 returns with exit code 0 iff no command of the given name is available.
15299 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15300 the output goes to stdout and nothing is printed to stderr.
15301 Additional arguments are always ignored.
15303 Since for each cipher there is a command of the same name,
15304 the 'no-cipher' compilation switches can be tested this way.
15306 ('openssl no-XXX' is not able to detect pseudo-commands such
15307 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15311 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15315 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15316 is set; it will be thrown away anyway because each handshake creates
15318 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15319 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15320 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15321 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15325 * New s_client option -ign_eof: EOF at stdin is ignored, and
15326 'Q' and 'R' lose their special meanings (quit/renegotiate).
15327 This is part of what -quiet does; unlike -quiet, -ign_eof
15328 does not suppress any output.
15332 * Add compatibility options to the purpose and trust code. The
15333 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15334 accepts a certificate or CA, this was the previous behaviour,
15335 with all the associated security issues.
15337 X509_TRUST_COMPAT is the old trust behaviour: only and
15338 automatically trust self signed roots in certificate store. A
15339 new trust setting X509_TRUST_DEFAULT is used to specify that
15340 a purpose has no associated trust setting and it should instead
15341 use the value in the default purpose.
15345 * Fix the PKCS#8 DSA private key code so it decodes keys again
15346 and fix a memory leak.
15350 * In util/mkerr.pl (which implements 'make errors'), preserve
15351 reason strings from the previous version of the .c file, as
15352 the default to have only downcase letters (and digits) in
15353 automatically generated reasons codes is not always appropriate.
15357 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15358 using strerror. Previously, ERR_reason_error_string() returned
15359 library names as reason strings for SYSerr; but SYSerr is a special
15360 case where small numbers are errno values, not library numbers.
15364 * Add '-dsaparam' option to 'openssl dhparam' application. This
15365 converts DSA parameters into DH parameters. (When creating parameters,
15366 DSA_generate_parameters is used.)
15370 * Include 'length' (recommended exponent length) in C code generated
15371 by 'openssl dhparam -C'.
15375 * The second argument to set_label in perlasm was already being used
15376 so couldn't be used as a "file scope" flag. Moved to third argument
15381 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15382 instead of RAND_bytes for encryption IVs and salts.
15386 * Include RAND_status() into RAND_METHOD instead of implementing
15387 it only for md_rand.c Otherwise replacing the PRNG by calling
15388 RAND_set_rand_method would be impossible.
15392 * Don't let DSA_generate_key() enter an infinite loop if the random
15393 number generation fails.
15397 * New 'rand' application for creating pseudo-random output.
15401 * Added configuration support for Linux/IA64
15403 *Rolf Haberrecker <rolf@suse.de>*
15405 * Assembler module support for Mingw32.
15409 * Shared library support for HPUX (in shlib/).
15411 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15413 * Shared library support for Solaris gcc.
15415 *Lutz Behnke <behnke@trustcenter.de>*
15417 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15419 * PKCS7_encrypt() was adding text MIME headers twice because they
15420 were added manually and by SMIME_crlf_copy().
15424 * In bntest.c don't call BN_rand with zero bits argument.
15426 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15428 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15429 case was implemented. This caused BN_div_recp() to fail occasionally.
15433 * Add an optional second argument to the set_label() in the perl
15434 assembly language builder. If this argument exists and is set
15435 to 1 it signals that the assembler should use a symbol whose
15436 scope is the entire file, not just the current function. This
15437 is needed with MASM which uses the format label:: for this scope.
15439 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15441 * Change the ASN1 types so they are typedefs by default. Before
15442 almost all types were #define'd to ASN1_STRING which was causing
15443 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15448 * Change names of new functions to the new get1/get0 naming
15449 convention: After 'get1', the caller owns a reference count
15450 and has to call `..._free`; 'get0' returns a pointer to some
15451 data structure without incrementing reference counters.
15452 (Some of the existing 'get' functions increment a reference
15453 counter, some don't.)
15454 Similarly, 'set1' and 'add1' functions increase reference
15455 counters or duplicate objects.
15459 * Allow for the possibility of temp RSA key generation failure:
15460 the code used to assume it always worked and crashed on failure.
15464 * Fix potential buffer overrun problem in BIO_printf().
15465 *Ulf Möller, using public domain code by Patrick Powell; problem
15466 pointed out by David Sacerdote <das33@cornell.edu>*
15468 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15469 RAND_egd() and RAND_status(). In the command line application,
15470 the EGD socket can be specified like a seed file using RANDFILE
15475 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15476 Some CAs (e.g. Verisign) distribute certificates in this form.
15480 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15481 list to exclude them. This means that no special compilation option
15482 is needed to use anonymous DH: it just needs to be included in the
15487 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15488 EVP_MD_type. The old functionality is available in a new macro called
15489 EVP_MD_md(). Change code that uses it and update docs.
15493 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15494 where the `void *` argument is replaced by a function pointer argument.
15495 Previously `void *` was abused to point to functions, which works on
15496 many platforms, but is not correct. As these functions are usually
15497 called by macros defined in OpenSSL header files, most source code
15498 should work without changes.
15502 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15503 sections with information on -D... compiler switches used for
15504 compiling the library so that applications can see them. To enable
15505 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15506 must be defined. E.g.,
15507 #define OPENSSL_ALGORITHM_DEFINES
15508 #include <openssl/opensslconf.h>
15509 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15511 *Richard Levitte, Ulf and Bodo Möller*
15513 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15518 * Change the 'other' type in certificate aux info to a STACK_OF
15519 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15520 the required ASN1 format: arbitrary types determined by an OID.
15524 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15525 argument to 'req'. This is not because the function is newer or
15526 better than others it just uses the work 'NEW' in the certificate
15527 request header lines. Some software needs this.
15531 * Reorganise password command line arguments: now passwords can be
15532 obtained from various sources. Delete the PEM_cb function and make
15533 it the default behaviour: i.e. if the callback is NULL and the
15534 usrdata argument is not NULL interpret it as a null terminated pass
15535 phrase. If usrdata and the callback are NULL then the pass phrase
15536 is prompted for as usual.
15540 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15541 the support is automatically enabled. The resulting binaries will
15542 autodetect the card and use it if present.
15544 *Ben Laurie and Compaq Inc.*
15546 * Work around for Netscape hang bug. This sends certificate request
15547 and server done in one record. Since this is perfectly legal in the
15548 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15549 the bugs/SSLv3 entry for more info.
15553 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15557 * Add -rand argument to smime and pkcs12 applications and read/write
15562 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15566 * Add command line password options to the remaining applications.
15570 * Bug fix for BN_div_recp() for numerators with an even number of
15575 * More tests in bntest.c, and changed test_bn output.
15579 * ./config recognizes MacOS X now.
15583 * Bug fix for BN_div() when the first words of num and divisor are
15584 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15588 * Add support for various broken PKCS#8 formats, and command line
15589 options to produce them.
15593 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15594 get temporary BIGNUMs from a BN_CTX.
15598 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15603 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15604 include a #define from the old name to the new. The original intent
15605 was that statically linked binaries could for example just call
15606 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15607 link with digests. This never worked because SSLeay_add_all_digests()
15608 and SSLeay_add_all_ciphers() were in the same source file so calling
15609 one would link with the other. They are now in separate source files.
15613 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15617 * Use a less unusual form of the Miller-Rabin primality test (it used
15618 a binary algorithm for exponentiation integrated into the Miller-Rabin
15619 loop, our standard modexp algorithms are faster).
15623 * Support for the EBCDIC character set completed.
15625 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15627 * Source code cleanups: use const where appropriate, eliminate casts,
15628 use `void *` instead of `char *` in lhash.
15632 * Bugfix: ssl3_send_server_key_exchange was not restartable
15633 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15634 this the server could overwrite ephemeral keys that the client
15639 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15640 using 50 iterations of the Rabin-Miller test.
15642 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15643 iterations of the Rabin-Miller test as required by the appendix
15644 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15645 As BN_is_prime_fasttest includes trial division, DSA parameter
15646 generation becomes much faster.
15648 This implies a change for the callback functions in DSA_is_prime
15649 and DSA_generate_parameters: The callback function is called once
15650 for each positive witness in the Rabin-Miller test, not just
15651 occasionally in the inner loop; and the parameters to the
15652 callback function now provide an iteration count for the outer
15653 loop rather than for the current invocation of the inner loop.
15654 DSA_generate_parameters additionally can call the callback
15655 function with an 'iteration count' of -1, meaning that a
15656 candidate has passed the trial division test (when q is generated
15657 from an application-provided seed, trial division is skipped).
15661 * New function BN_is_prime_fasttest that optionally does trial
15662 division before starting the Rabin-Miller test and has
15663 an additional BN_CTX * argument (whereas BN_is_prime always
15664 has to allocate at least one BN_CTX).
15665 'callback(1, -1, cb_arg)' is called when a number has passed the
15666 trial division stage.
15670 * Fix for bug in CRL encoding. The validity dates weren't being handled
15675 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15679 * New function BN_pseudo_rand().
15683 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15684 bignum version of BN_from_montgomery() with the working code from
15685 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15690 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15691 made it impossible to use the same SSL_SESSION data structure in
15692 SSL2 clients in multiple threads.
15696 * The return value of RAND_load_file() no longer counts bytes obtained
15697 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15698 to seed the PRNG (previously an explicit byte count was required).
15700 *Ulf Möller, Bodo Möller*
15702 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15703 used `char *` instead of `void *` and had casts all over the place.
15707 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15711 * Retain source code compatibility for BN_prime_checks macro:
15712 BN_is_prime(..., BN_prime_checks, ...) now uses
15713 BN_prime_checks_for_size to determine the appropriate number of
15714 Rabin-Miller iterations.
15718 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15719 DH_CHECK_P_NOT_SAFE_PRIME.
15720 (Check if this is true? OpenPGP calls them "strong".)
15724 * Merge the functionality of "dh" and "gendh" programs into a new program
15725 "dhparam". The old programs are retained for now but will handle DH keys
15726 (instead of parameters) in future.
15730 * Make the ciphers, s_server and s_client programs check the return values
15731 when a new cipher list is set.
15735 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15736 ciphers. Before when the 56bit ciphers were enabled the sorting was
15739 The syntax for the cipher sorting has been extended to support sorting by
15740 cipher-strength (using the strength_bits hard coded in the tables).
15741 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15743 Fix a bug in the cipher-command parser: when supplying a cipher command
15744 string with an "undefined" symbol (neither command nor alphanumeric
15745 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15746 an error is flagged.
15748 Due to the strength-sorting extension, the code of the
15749 ssl_create_cipher_list() function was completely rearranged. I hope that
15750 the readability was also increased :-)
15752 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15754 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15755 for the first serial number and places 2 in the serial number file. This
15756 avoids problems when the root CA is created with serial number zero and
15757 the first user certificate has the same issuer name and serial number
15762 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15763 the new code. Add documentation for this stuff.
15767 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15768 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15769 structures and behave in an analogous way to the X509v3 functions:
15770 they shouldn't be called directly but wrapper functions should be used
15773 So we also now have some wrapper functions that call the X509at functions
15774 when passed certificate requests. (TO DO: similar things can be done with
15775 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15776 things. Some of these need some d2i or i2d and print functionality
15777 because they handle more complex structures.)
15781 * Add missing #ifndefs that caused missing symbols when building libssl
15782 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15783 NO_RSA in `ssl/s2*.c`.
15785 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15787 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15788 has a return value which indicates the quality of the random data
15789 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15790 error queue. New function RAND_pseudo_bytes() generates output that is
15791 guaranteed to be unique but not unpredictable. RAND_add is like
15792 RAND_seed, but takes an extra argument for an entropy estimate
15793 (RAND_seed always assumes full entropy).
15797 * Do more iterations of Rabin-Miller probable prime test (specifically,
15798 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15799 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15800 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15801 false-positive rate of at most 2^-80 for random input.
15805 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15809 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15810 in the 0.9.5 release), this returns the chain
15811 from an X509_CTX structure with a dup of the stack and all
15812 the X509 reference counts upped: so the stack will exist
15813 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15816 Also make SSL_SESSION_print() print out the verify return
15821 * Add manpage for the pkcs12 command. Also change the default
15822 behaviour so MAC iteration counts are used unless the new
15823 -nomaciter option is used. This improves file security and
15824 only older versions of MSIE (4.0 for example) need it.
15828 * Honor the no-xxx Configure options when creating .DEF files.
15832 * Add PKCS#10 attributes to field table: challengePassword,
15833 unstructuredName and unstructuredAddress. These are taken from
15834 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15835 international characters are used.
15837 More changes to X509_ATTRIBUTE code: allow the setting of types
15838 based on strings. Remove the 'loc' parameter when adding
15839 attributes because these will be a SET OF encoding which is sorted
15844 * Initial changes to the 'req' utility to allow request generation
15845 automation. This will allow an application to just generate a template
15846 file containing all the field values and have req construct the
15849 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15850 used all over the place including certificate requests and PKCS#7
15851 structures. They are currently handled manually where necessary with
15852 some primitive wrappers for PKCS#7. The new functions behave in a
15853 manner analogous to the X509 extension functions: they allow
15854 attributes to be looked up by NID and added.
15856 Later something similar to the X509V3 code would be desirable to
15857 automatically handle the encoding, decoding and printing of the
15858 more complex types. The string types like challengePassword can
15859 be handled by the string table functions.
15861 Also modified the multi byte string table handling. Now there is
15862 a 'global mask' which masks out certain types. The table itself
15863 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15864 is useful when for example there is only one permissible type
15865 (as in countryName) and using the mask might result in no valid
15870 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15871 SSL_get_peer_finished to allow applications to obtain the latest
15872 Finished messages sent to the peer or expected from the peer,
15873 respectively. (SSL_get_peer_finished is usually the Finished message
15874 actually received from the peer, otherwise the protocol will be aborted.)
15876 As the Finished message are message digests of the complete handshake
15877 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15878 be used for external authentication procedures when the authentication
15879 provided by SSL/TLS is not desired or is not enough.
15883 * Enhanced support for Alpha Linux is added. Now ./config checks if
15884 the host supports BWX extension and if Compaq C is present on the
15885 $PATH. Just exploiting of the BWX extension results in 20-30%
15886 performance kick for some algorithms, e.g. DES and RC4 to mention
15887 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15892 * Add support for MS "fast SGC". This is arguably a violation of the
15893 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15894 weak crypto and after checking the certificate is SGC a second one
15895 with strong crypto. MS SGC stops the first handshake after receiving
15896 the server certificate message and sends a second client hello. Since
15897 a server will typically do all the time consuming operations before
15898 expecting any further messages from the client (server key exchange
15899 is the most expensive) there is little difference between the two.
15901 To get OpenSSL to support MS SGC we have to permit a second client
15902 hello message after we have sent server done. In addition we have to
15903 reset the MAC if we do get this second client hello.
15907 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15908 if a DER encoded private key is RSA or DSA traditional format. Changed
15909 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15910 format DER encoded private key. Newer code should use PKCS#8 format which
15911 has the key type encoded in the ASN1 structure. Added DER private key
15912 support to pkcs8 application.
15916 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15917 ciphersuites has been selected (as required by the SSL 3/TLS 1
15918 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15919 is set, we interpret this as a request to violate the specification
15920 (the worst that can happen is a handshake failure, and 'correct'
15921 behaviour would result in a handshake failure anyway).
15925 * In SSL_CTX_add_session, take into account that there might be multiple
15926 SSL_SESSION structures with the same session ID (e.g. when two threads
15927 concurrently obtain them from an external cache).
15928 The internal cache can handle only one SSL_SESSION with a given ID,
15929 so if there's a conflict, we now throw out the old one to achieve
15934 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15935 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15936 some routines that use cipher OIDs: some ciphers do not have OIDs
15937 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15942 * Simplify the trust setting structure and code. Now we just have
15943 two sequences of OIDs for trusted and rejected settings. These will
15944 typically have values the same as the extended key usage extension
15945 and any application specific purposes.
15947 The trust checking code now has a default behaviour: it will just
15948 check for an object with the same NID as the passed id. Functions can
15949 be provided to override either the default behaviour or the behaviour
15950 for a given id. SSL client, server and email already have functions
15951 in place for compatibility: they check the NID and also return "trusted"
15952 if the certificate is self signed.
15956 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15957 traditional format into an EVP_PKEY structure.
15961 * Add a password callback function PEM_cb() which either prompts for
15962 a password if usr_data is NULL or otherwise assumes it is a null
15963 terminated password. Allow passwords to be passed on command line
15964 environment or config files in a few more utilities.
15968 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15969 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15970 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15971 Update documentation.
15975 * Support for ASN1 "NULL" type. This could be handled before by using
15976 ASN1_TYPE but there wasn't any function that would try to read a NULL
15977 and produce an error if it couldn't. For compatibility we also have
15978 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15979 don't allocate anything because they don't need to.
15983 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15986 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15988 * Rebuild of the memory allocation routines used by OpenSSL code and
15989 possibly others as well. The purpose is to make an interface that
15990 provide hooks so anyone can build a separate set of allocation and
15991 deallocation routines to be used by OpenSSL, for example memory
15992 pool implementations, or something else, which was previously hard
15993 since Malloc(), Realloc() and Free() were defined as macros having
15994 the values malloc, realloc and free, respectively (except for Win32
15995 compilations). The same is provided for memory debugging code.
15996 OpenSSL already comes with functionality to find memory leaks, but
15997 this gives people a chance to debug other memory problems.
15999 With these changes, a new set of functions and macros have appeared:
16001 CRYPTO_set_mem_debug_functions() [F]
16002 CRYPTO_get_mem_debug_functions() [F]
16003 CRYPTO_dbg_set_options() [F]
16004 CRYPTO_dbg_get_options() [F]
16005 CRYPTO_malloc_debug_init() [M]
16007 The memory debug functions are NULL by default, unless the library
16008 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16009 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16010 gives the standard debugging functions that come with OpenSSL) or
16011 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16012 provided by the library user) must be used. When the standard
16013 debugging functions are used, CRYPTO_dbg_set_options can be used to
16014 request additional information:
16015 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16016 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16018 Also, things like CRYPTO_set_mem_functions will always give the
16019 expected result (the new set of functions is used for allocation
16020 and deallocation) at all times, regardless of platform and compiler
16023 To finish it up, some functions that were never use in any other
16024 way than through macros have a new API and new semantic:
16026 CRYPTO_dbg_malloc()
16027 CRYPTO_dbg_realloc()
16030 All macros of value have retained their old syntax.
16032 *Richard Levitte and Bodo Moeller*
16034 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16035 ordering of SMIMECapabilities wasn't in "strength order" and there
16036 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16041 * Some ASN1 types with illegal zero length encoding (INTEGER,
16042 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16044 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16046 * Merge in my S/MIME library for OpenSSL. This provides a simple
16047 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16048 functionality to handle multipart/signed properly) and a utility
16049 called 'smime' to call all this stuff. This is based on code I
16050 originally wrote for Celo who have kindly allowed it to be
16051 included in OpenSSL.
16055 * Add variants des_set_key_checked and des_set_key_unchecked of
16056 des_set_key (aka des_key_sched). Global variable des_check_key
16057 decides which of these is called by des_set_key; this way
16058 des_check_key behaves as it always did, but applications and
16059 the library itself, which was buggy for des_check_key == 1,
16060 have a cleaner way to pick the version they need.
16064 * New function PKCS12_newpass() which changes the password of a
16069 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16070 dynamic mix. In both cases the ids can be used as an index into the
16071 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16072 functions so they accept a list of the field values and the
16073 application doesn't need to directly manipulate the X509_TRUST
16078 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16083 * Modify the way the V3 extension code looks up extensions. This now
16084 works in a similar way to the object code: we have some "standard"
16085 extensions in a static table which is searched with OBJ_bsearch()
16086 and the application can add dynamic ones if needed. The file
16087 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16088 updated whenever a new extension is added to the core code and kept
16089 in ext_nid order. There is a simple program 'tabtest.c' which checks
16090 this. New extensions are not added too often so this file can readily
16091 be maintained manually.
16093 There are two big advantages in doing things this way. The extensions
16094 can be looked up immediately and no longer need to be "added" using
16095 X509V3_add_standard_extensions(): this function now does nothing.
16096 Side note: I get *lots* of email saying the extension code doesn't
16097 work because people forget to call this function.
16098 Also no dynamic allocation is done unless new extensions are added:
16099 so if we don't add custom extensions there is no need to call
16100 X509V3_EXT_cleanup().
16104 * Modify enc utility's salting as follows: make salting the default. Add a
16105 magic header, so unsalted files fail gracefully instead of just decrypting
16106 to garbage. This is because not salting is a big security hole, so people
16107 should be discouraged from doing it.
16111 * Fixes and enhancements to the 'x509' utility. It allowed a message
16112 digest to be passed on the command line but it only used this
16113 parameter when signing a certificate. Modified so all relevant
16114 operations are affected by the digest parameter including the
16115 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16116 DSA key was used because it didn't fix the digest.
16120 * Initial certificate chain verify code. Currently tests the untrusted
16121 certificates for consistency with the verify purpose (which is set
16122 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16124 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16125 this is because it will reject chains with invalid extensions whereas
16126 every previous version of OpenSSL and SSLeay made no checks at all.
16128 Trust code: checks the root CA for the relevant trust settings. Trust
16129 settings have an initial value consistent with the verify purpose: e.g.
16130 if the verify purpose is for SSL client use it expects the CA to be
16131 trusted for SSL client use. However the default value can be changed to
16132 permit custom trust settings: one example of this would be to only trust
16133 certificates from a specific "secure" set of CAs.
16135 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16136 which should be used for version portability: especially since the
16137 verify structure is likely to change more often now.
16139 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16140 to set them. If not set then assume SSL clients will verify SSL servers
16143 Two new options to the verify program: -untrusted allows a set of
16144 untrusted certificates to be passed in and -purpose which sets the
16145 intended purpose of the certificate. If a purpose is set then the
16146 new chain verify code is used to check extension consistency.
16150 * Support for the authority information access extension.
16154 * Modify RSA and DSA PEM read routines to transparently handle
16155 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16156 public keys in a format compatible with certificate
16157 SubjectPublicKeyInfo structures. Unfortunately there were already
16158 functions called *_PublicKey_* which used various odd formats so
16159 these are retained for compatibility: however the DSA variants were
16160 never in a public release so they have been deleted. Changed dsa/rsa
16161 utilities to handle the new format: note no releases ever handled public
16162 keys so we should be OK.
16164 The primary motivation for this change is to avoid the same fiasco
16165 that dogs private keys: there are several incompatible private key
16166 formats some of which are standard and some OpenSSL specific and
16167 require various evil hacks to allow partial transparent handling and
16168 even then it doesn't work with DER formats. Given the option anything
16169 other than PKCS#8 should be dumped: but the other formats have to
16170 stay in the name of compatibility.
16172 With public keys and the benefit of hindsight one standard format
16173 is used which works with EVP_PKEY, RSA or DSA structures: though
16174 it clearly returns an error if you try to read the wrong kind of key.
16176 Added a -pubkey option to the 'x509' utility to output the public key.
16177 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16178 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16179 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16180 that do the same as the `EVP_PKEY_assign_*()` except they up the
16181 reference count of the added key (they don't "swallow" the
16186 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16187 CRLs would fail if the file contained no certificates or no CRLs:
16188 added a new function to read in both types and return the number
16189 read: this means that if none are read it will be an error. The
16190 DER versions of the certificate and CRL reader would always fail
16191 because it isn't possible to mix certificates and CRLs in DER format
16192 without choking one or the other routine. Changed this to just read
16193 a certificate: this is the best we can do. Also modified the code
16194 in apps/verify.c to take notice of return codes: it was previously
16195 attempting to read in certificates from NULL pointers and ignoring
16196 any errors: this is one reason why the cert and CRL reader seemed
16197 to work. It doesn't check return codes from the default certificate
16198 routines: these may well fail if the certificates aren't installed.
16202 * Code to support otherName option in GeneralName.
16206 * First update to verify code. Change the verify utility
16207 so it warns if it is passed a self signed certificate:
16208 for consistency with the normal behaviour. X509_verify
16209 has been modified to it will now verify a self signed
16210 certificate if *exactly* the same certificate appears
16211 in the store: it was previously impossible to trust a
16212 single self signed certificate. This means that:
16213 openssl verify ss.pem
16214 now gives a warning about a self signed certificate but
16215 openssl verify -CAfile ss.pem ss.pem
16220 * For servers, store verify_result in SSL_SESSION data structure
16221 (and add it to external session representation).
16222 This is needed when client certificate verifications fails,
16223 but an application-provided verification callback (set by
16224 SSL_CTX_set_cert_verify_callback) allows accepting the session
16225 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16226 but returns 1): When the session is reused, we have to set
16227 ssl->verify_result to the appropriate error code to avoid
16230 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16232 * Fix a bug in the new PKCS#7 code: it didn't consider the
16233 case in PKCS7_dataInit() where the signed PKCS7 structure
16234 didn't contain any existing data because it was being created.
16236 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16238 * Add a salt to the key derivation routines in enc.c. This
16239 forms the first 8 bytes of the encrypted file. Also add a
16240 -S option to allow a salt to be input on the command line.
16244 * New function X509_cmp(). Oddly enough there wasn't a function
16245 to compare two certificates. We do this by working out the SHA1
16246 hash and comparing that. X509_cmp() will be needed by the trust
16251 * SSL_get1_session() is like SSL_get_session(), but increments
16252 the reference count in the SSL_SESSION returned.
16254 *Geoff Thorpe <geoff@eu.c2.net>*
16256 * Fix for 'req': it was adding a null to request attributes.
16257 Also change the X509_LOOKUP and X509_INFO code to handle
16258 certificate auxiliary information.
16262 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16267 * Add the possibility to add extra information to the memory leak
16268 detecting output, to form tracebacks, showing from where each
16269 allocation was originated: CRYPTO_push_info("constant string") adds
16270 the string plus current file name and line number to a per-thread
16271 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16272 is like calling CYRPTO_pop_info() until the stack is empty.
16273 Also updated memory leak detection code to be multi-thread-safe.
16277 * Add options -text and -noout to pkcs7 utility and delete the
16278 encryption options which never did anything. Update docs.
16282 * Add options to some of the utilities to allow the pass phrase
16283 to be included on either the command line (not recommended on
16284 OSes like Unix) or read from the environment. Update the
16285 manpages and fix a few bugs.
16289 * Add a few manpages for some of the openssl commands.
16293 * Fix the -revoke option in ca. It was freeing up memory twice,
16294 leaking and not finding already revoked certificates.
16298 * Extensive changes to support certificate auxiliary information.
16299 This involves the use of X509_CERT_AUX structure and X509_AUX
16300 functions. An X509_AUX function such as PEM_read_X509_AUX()
16301 can still read in a certificate file in the usual way but it
16302 will also read in any additional "auxiliary information". By
16303 doing things this way a fair degree of compatibility can be
16304 retained: existing certificates can have this information added
16305 using the new 'x509' options.
16307 Current auxiliary information includes an "alias" and some trust
16308 settings. The trust settings will ultimately be used in enhanced
16309 certificate chain verification routines: currently a certificate
16310 can only be trusted if it is self signed and then it is trusted
16315 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16316 The problem was that one of the replacement routines had not been working
16317 since SSLeay releases. For now the offending routine has been replaced
16318 with non-optimised assembler. Even so, this now gives around 95%
16319 performance improvement for 1024 bit RSA signs.
16323 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16324 handling. Most clients have the effective key size in bits equal to
16325 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16326 A few however don't do this and instead use the size of the decrypted key
16327 to determine the RC2 key length and the AlgorithmIdentifier to determine
16328 the effective key length. In this case the effective key length can still
16329 be 40 bits but the key length can be 168 bits for example. This is fixed
16330 by manually forcing an RC2 key into the EVP_PKEY structure because the
16331 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16332 the key length and effective key length are equal.
16336 * Add a bunch of functions that should simplify the creation of
16337 X509_NAME structures. Now you should be able to do:
16338 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16339 and have it automatically work out the correct field type and fill in
16340 the structures. The more adventurous can try:
16341 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16342 and it will (hopefully) work out the correct multibyte encoding.
16346 * Change the 'req' utility to use the new field handling and multibyte
16347 copy routines. Before the DN field creation was handled in an ad hoc
16348 way in req, ca, and x509 which was rather broken and didn't support
16349 BMPStrings or UTF8Strings. Since some software doesn't implement
16350 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16351 using the dirstring_type option. See the new comment in the default
16352 openssl.cnf for more info.
16356 * Make crypto/rand/md_rand.c more robust:
16357 - Assure unique random numbers after fork().
16358 - Make sure that concurrent threads access the global counter and
16359 md serializably so that we never lose entropy in them
16360 or use exactly the same state in multiple threads.
16361 Access to the large state is not always serializable because
16362 the additional locking could be a performance killer, and
16363 md should be large enough anyway.
16367 * New file apps/app_rand.c with commonly needed functionality
16368 for handling the random seed file.
16370 Use the random seed file in some applications that previously did not:
16372 dsaparam -genkey (which also ignored its '-rand' option),
16375 x509 (when signing).
16376 Except on systems with /dev/urandom, it is crucial to have a random
16377 seed file at least for key creation, DSA signing, and for DH exchanges;
16378 for RSA signatures we could do without one.
16380 gendh and gendsa (unlike genrsa) used to read only the first byte
16381 of each file listed in the '-rand' option. The function as previously
16382 found in genrsa is now in app_rand.c and is used by all programs
16383 that support '-rand'.
16387 * In RAND_write_file, use mode 0600 for creating files;
16388 don't just chmod when it may be too late.
16392 * Report an error from X509_STORE_load_locations
16393 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16397 * New function ASN1_mbstring_copy() this copies a string in either
16398 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16399 into an ASN1_STRING type. A mask of permissible types is passed
16400 and it chooses the "minimal" type to use or an error if not type
16405 * Add function equivalents to the various macros in asn1.h. The old
16406 macros are retained with an `M_` prefix. Code inside the library can
16407 use the `M_` macros. External code (including the openssl utility)
16408 should *NOT* in order to be "shared library friendly".
16412 * Add various functions that can check a certificate's extensions
16413 to see if it usable for various purposes such as SSL client,
16414 server or S/MIME and CAs of these types. This is currently
16415 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16416 verification. Also added a -purpose flag to x509 utility to
16417 print out all the purposes.
16421 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16426 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16427 for, obtain and decode and extension and obtain its critical flag.
16428 This allows all the necessary extension code to be handled in a
16429 single function call.
16433 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16434 platforms. See crypto/rc4/rc4_enc.c for further details.
16438 * New -noout option to asn1parse. This causes no output to be produced
16439 its main use is when combined with -strparse and -out to extract data
16440 from a file (which may not be in ASN.1 format).
16444 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16445 when producing the local key id.
16447 *Richard Levitte <levitte@stacken.kth.se>*
16449 * New option -dhparam in s_server. This allows a DH parameter file to be
16450 stated explicitly. If it is not stated then it tries the first server
16451 certificate file. The previous behaviour hard coded the filename
16456 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16457 a public key to be input or output. For example:
16458 openssl rsa -in key.pem -pubout -out pubkey.pem
16459 Also added necessary DSA public key functions to handle this.
16463 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16464 in the message. This was handled by allowing
16465 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16467 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16469 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16470 to the end of the strings whereas this didn't. This would cause problems
16471 if strings read with d2i_ASN1_bytes() were later modified.
16473 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16475 * Fix for base64 decode bug. When a base64 bio reads only one line of
16476 data and it contains EOF it will end up returning an error. This is
16477 caused by input 46 bytes long. The cause is due to the way base64
16478 BIOs find the start of base64 encoded data. They do this by trying a
16479 trial decode on each line until they find one that works. When they
16480 do a flag is set and it starts again knowing it can pass all the
16481 data directly through the decoder. Unfortunately it doesn't reset
16482 the context it uses. This means that if EOF is reached an attempt
16483 is made to pass two EOFs through the context and this causes the
16484 resulting error. This can also cause other problems as well. As is
16485 usual with these problems it takes *ages* to find and the fix is
16486 trivial: move one line.
16488 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16490 * Ugly workaround to get s_client and s_server working under Windows. The
16491 old code wouldn't work because it needed to select() on sockets and the
16492 tty (for keypresses and to see if data could be written). Win32 only
16493 supports select() on sockets so we select() with a 1s timeout on the
16494 sockets and then see if any characters are waiting to be read, if none
16495 are present then we retry, we also assume we can always write data to
16496 the tty. This isn't nice because the code then blocks until we've
16497 received a complete line of data and it is effectively polling the
16498 keyboard at 1s intervals: however it's quite a bit better than not
16499 working at all :-) A dedicated Windows application might handle this
16500 with an event loop for example.
16504 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16505 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16506 will be called when RSA_sign() and RSA_verify() are used. This is useful
16507 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16508 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16509 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16510 This necessitated the support of an extra signature type NID_md5_sha1
16511 for SSL signatures and modifications to the SSL library to use it instead
16512 of calling RSA_public_decrypt() and RSA_private_encrypt().
16516 * Add new -verify -CAfile and -CApath options to the crl program, these
16517 will lookup a CRL issuers certificate and verify the signature in a
16518 similar way to the verify program. Tidy up the crl program so it
16519 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16520 less strict. It will now permit CRL extensions even if it is not
16521 a V2 CRL: this will allow it to tolerate some broken CRLs.
16525 * Initialize all non-automatic variables each time one of the openssl
16526 sub-programs is started (this is necessary as they may be started
16527 multiple times from the "OpenSSL>" prompt).
16529 *Lennart Bang, Bodo Moeller*
16531 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16532 removing all other RSA functionality (this is what NO_RSA does). This
16533 is so (for example) those in the US can disable those operations covered
16534 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16539 * Non-copying interface to BIO pairs.
16540 (still largely untested)
16544 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16545 ASCII string. This was handled independently in various places before.
16549 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16550 UTF8 strings a character at a time.
16554 * Use client_version from client hello to select the protocol
16555 (s23_srvr.c) and for RSA client key exchange verification
16556 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16560 * Add various utility functions to handle SPKACs, these were previously
16561 handled by poking round in the structure internals. Added new function
16562 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16563 print, verify and generate SPKACs. Based on an original idea from
16564 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16568 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16572 * Allow the config file extension section to be overwritten on the
16573 command line. Based on an original idea from Massimiliano Pala
16574 <madwolf@comune.modena.it>. The new option is called -extensions
16575 and can be applied to ca, req and x509. Also -reqexts to override
16576 the request extensions in req and -crlexts to override the crl extensions
16581 * Add new feature to the SPKAC handling in ca. Now you can include
16582 the same field multiple times by preceding it by "XXXX." for example:
16585 this is the same syntax as used in the req config file.
16589 * Allow certificate extensions to be added to certificate requests. These
16590 are specified in a 'req_extensions' option of the req section of the
16591 config file. They can be printed out with the -text option to req but
16592 are otherwise ignored at present.
16596 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16597 data read consists of only the final block it would not decrypted because
16598 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16599 A misplaced 'break' also meant the decrypted final block might not be
16600 copied until the next read.
16604 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16605 a few extra parameters to the DH structure: these will be useful if
16606 for example we want the value of 'q' or implement X9.42 DH.
16610 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16611 provides hooks that allow the default DSA functions or functions on a
16612 "per key" basis to be replaced. This allows hardware acceleration and
16613 hardware key storage to be handled without major modification to the
16614 library. Also added low level modexp hooks and CRYPTO_EX structure and
16615 associated functions.
16619 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16620 as "read only": it can't be written to and the buffer it points to will
16621 not be freed. Reading from a read only BIO is much more efficient than
16622 a normal memory BIO. This was added because there are several times when
16623 an area of memory needs to be read from a BIO. The previous method was
16624 to create a memory BIO and write the data to it, this results in two
16625 copies of the data and an O(n^2) reading algorithm. There is a new
16626 function BIO_new_mem_buf() which creates a read only memory BIO from
16627 an area of memory. Also modified the PKCS#7 routines to use read only
16632 * Bugfix: ssl23_get_client_hello did not work properly when called in
16633 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16634 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16635 but a retry condition occurred while trying to read the rest.
16639 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16640 NID_pkcs7_encrypted by default: this was wrong since this should almost
16641 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16642 the encrypted data type: this is a more sensible place to put it and it
16643 allows the PKCS#12 code to be tidied up that duplicated this
16648 * Changed obj_dat.pl script so it takes its input and output files on
16649 the command line. This should avoid shell escape redirection problems
16654 * Initial support for certificate extension requests, these are included
16655 in things like Xenroll certificate requests. Included functions to allow
16656 extensions to be obtained and added.
16660 * -crlf option to s_client and s_server for sending newlines as
16661 CRLF (as required by many protocols).
16665 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16667 * Install libRSAglue.a when OpenSSL is built with RSAref.
16669 *Ralf S. Engelschall*
16671 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16673 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16675 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16680 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16681 DH parameters/keys (q is lost during that conversion, but the resulting
16682 DH parameters contain its length).
16684 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16685 much faster than DH_generate_parameters (which creates parameters
16686 where `p = 2*q + 1`), and also the smaller q makes DH computations
16687 much more efficient (160-bit exponentiation instead of 1024-bit
16688 exponentiation); so this provides a convenient way to support DHE
16689 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16690 utter importance to use
16691 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16693 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16694 when such DH parameters are used, because otherwise small subgroup
16695 attacks may become possible!
16699 * Avoid memory leak in i2d_DHparams.
16703 * Allow the -k option to be used more than once in the enc program:
16704 this allows the same encrypted message to be read by multiple recipients.
16708 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16709 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16710 it will always use the numerical form of the OID, even if it has a short
16715 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16716 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16717 otherwise bn_mod_exp was called. In the case of hardware keys for example
16718 no private key components need be present and it might store extra data
16719 in the RSA structure, which cannot be accessed from bn_mod_exp.
16720 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16721 private key operations.
16725 * Added support for SPARC Linux.
16729 * pem_password_cb function type incompatibly changed from
16730 typedef int pem_password_cb(char *buf, int size, int rwflag);
16732 ....(char *buf, int size, int rwflag, void *userdata);
16733 so that applications can pass data to their callbacks:
16734 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16735 additional void * argument, which is just handed through whenever
16736 the password callback is called.
16738 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16740 New function SSL_CTX_set_default_passwd_cb_userdata.
16742 Compatibility note: As many C implementations push function arguments
16743 onto the stack in reverse order, the new library version is likely to
16744 interoperate with programs that have been compiled with the old
16745 pem_password_cb definition (PEM_whatever takes some data that
16746 happens to be on the stack as its last argument, and the callback
16747 just ignores this garbage); but there is no guarantee whatsoever that
16750 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16751 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16752 problems not only on Windows, but also on some Unix platforms.
16753 To avoid problematic command lines, these definitions are now in an
16754 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16755 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16759 * MIPS III/IV assembler module is reimplemented.
16763 * More DES library cleanups: remove references to srand/rand and
16764 delete an unused file.
16768 * Add support for the free Netwide assembler (NASM) under Win32,
16769 since not many people have MASM (ml) and it can be hard to obtain.
16770 This is currently experimental but it seems to work OK and pass all
16771 the tests. Check out INSTALL.W32 for info.
16775 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16776 without temporary keys kept an extra copy of the server key,
16777 and connections with temporary keys did not free everything in case
16782 * New function RSA_check_key and new openssl rsa option -check
16783 for verifying the consistency of RSA keys.
16785 *Ulf Moeller, Bodo Moeller*
16787 * Various changes to make Win32 compile work:
16788 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16789 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16790 comparison" warnings.
16791 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16795 * Add a debugging option to PKCS#5 v2 key generation function: when
16796 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16797 derived keys are printed to stderr.
16801 * Copy the flags in ASN1_STRING_dup().
16803 *Roman E. Pavlov <pre@mo.msk.ru>*
16805 * The x509 application mishandled signing requests containing DSA
16806 keys when the signing key was also DSA and the parameters didn't match.
16808 It was supposed to omit the parameters when they matched the signing key:
16809 the verifying software was then supposed to automatically use the CA's
16810 parameters if they were absent from the end user certificate.
16812 Omitting parameters is no longer recommended. The test was also
16813 the wrong way round! This was probably due to unusual behaviour in
16814 EVP_cmp_parameters() which returns 1 if the parameters match.
16815 This meant that parameters were omitted when they *didn't* match and
16816 the certificate was useless. Certificates signed with 'ca' didn't have
16819 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16821 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16822 The interface is as follows:
16823 Applications can use
16824 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16825 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16826 "off" is now the default.
16827 The library internally uses
16828 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16829 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16830 to disable memory-checking temporarily.
16832 Some inconsistent states that previously were possible (and were
16833 even the default) are now avoided.
16835 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16836 with each memory chunk allocated; this is occasionally more helpful
16837 than just having a counter.
16839 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16841 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16846 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16847 which largely parallels "options", but is for changing API behaviour,
16848 whereas "options" are about protocol behaviour.
16849 Initial "mode" flags are:
16851 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16852 a single record has been written.
16853 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16854 retries use the same buffer location.
16855 (But all of the contents must be
16860 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16863 * Fix problems with no-hmac etc.
16865 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16867 * New functions RSA_get_default_method(), RSA_set_method() and
16868 RSA_get_method(). These allows replacement of RSA_METHODs without having
16869 to mess around with the internals of an RSA structure.
16873 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16874 Also really enable memory leak checks in openssl.c and in some
16877 *Chad C. Mulligan, Bodo Moeller*
16879 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16880 up the length of negative integers. This has now been simplified to just
16881 store the length when it is first determined and use it later, rather
16882 than trying to keep track of where data is copied and updating it to
16884 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16886 * Add a new function PKCS7_signatureVerify. This allows the verification
16887 of a PKCS#7 signature but with the signing certificate passed to the
16888 function itself. This contrasts with PKCS7_dataVerify which assumes the
16889 certificate is present in the PKCS#7 structure. This isn't always the
16890 case: certificates can be omitted from a PKCS#7 structure and be
16891 distributed by "out of band" means (such as a certificate database).
16895 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16896 function prototypes in pem.h, also change util/mkdef.pl to add the
16897 necessary function names.
16901 * mk1mf.pl (used by Windows builds) did not properly read the
16902 options set by Configure in the top level Makefile, and Configure
16903 was not even able to write more than one option correctly.
16904 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16908 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16909 file to be loaded from a BIO or FILE pointer. The BIO version will
16910 for example allow memory BIOs to contain config info.
16914 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16915 Whoever hopes to achieve shared-library compatibility across versions
16916 must use this, not the compile-time macro.
16917 (Exercise 0.9.4: Which is the minimum library version required by
16919 Note: All this applies only to multi-threaded programs, others don't
16924 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16925 through a BIO pair triggered the default case, i.e.
16926 SSLerr(...,SSL_R_UNKNOWN_STATE).
16930 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16931 can use the SSL library even if none of the specific BIOs is
16936 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16937 for the encoded length.
16939 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16941 * Add initial documentation of the X509V3 functions.
16945 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16946 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16947 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16948 secure PKCS#8 private key format with a high iteration count.
16952 * Fix determination of Perl interpreter: A perl or perl5
16953 *directory* in $PATH was also accepted as the interpreter.
16955 *Ralf S. Engelschall*
16957 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16958 wrong with it but it was very old and did things like calling
16959 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16960 unusual formatting.
16964 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16965 to use the new extension code.
16969 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16970 with macros. This should make it easier to change their form, add extra
16971 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16976 * Add to configuration table a new entry that can specify an alternative
16977 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16978 according to Mark Crispin <MRC@Panda.COM>.
16982 * DES CBC did not update the IV. Weird.
16986 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16987 Changing the behaviour of the former might break existing programs --
16988 where IV updating is needed, des_ncbc_encrypt can be used.
16991 * When bntest is run from "make test" it drives bc to check its
16992 calculations, as well as internally checking them. If an internal check
16993 fails, it needs to cause bc to give a non-zero result or make test carries
16994 on without noticing the failure. Fixed.
16998 * DES library cleanups.
17002 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17003 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17004 ciphers. NOTE: although the key derivation function has been verified
17005 against some published test vectors it has not been extensively tested
17006 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17011 * Instead of "mkdir -p", which is not fully portable, use new
17012 Perl script "util/mkdir-p.pl".
17016 * Rewrite the way password based encryption (PBE) is handled. It used to
17017 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17018 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17019 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17020 the 'parameter' field of the AlgorithmIdentifier is passed to the
17021 underlying key generation function so it must do its own ASN1 parsing.
17022 This has also changed the EVP_PBE_CipherInit() function which now has a
17023 'parameter' argument instead of literal salt and iteration count values
17024 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17028 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17029 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17030 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17031 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17032 value was just used as a "magic string" and not used directly its
17033 value doesn't matter.
17037 * Introduce some semblance of const correctness to BN. Shame C doesn't
17042 * "linux-sparc64" configuration (ultrapenguin).
17044 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17045 "linux-sparc" configuration.
17047 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17049 * config now generates no-xxx options for missing ciphers.
17053 * Support the EBCDIC character set (work in progress).
17054 File ebcdic.c not yet included because it has a different license.
17056 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17058 * Support BS2000/OSD-POSIX.
17060 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17062 * Make callbacks for key generation use `void *` instead of `char *`.
17066 * Make S/MIME samples compile (not yet tested).
17070 * Additional typesafe stacks.
17074 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17078 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17080 * New configuration variant "sco5-gcc".
17082 * Updated some demos.
17084 *Sean O Riordain, Wade Scholine*
17086 * Add missing BIO_free at exit of pkcs12 application.
17090 * Fix memory leak in conf.c.
17094 * Updates for Win32 to assembler version of MD5.
17098 * Set #! path to perl in apps/der_chop to where we found it
17099 instead of using a fixed path.
17103 * SHA library changes for irix64-mips4-cc.
17107 * Improvements for VMS support.
17111 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17113 * Bignum library bug fix. IRIX 6 passes "make test" now!
17114 This also avoids the problems with SC4.2 and unpatched SC5.
17116 *Andy Polyakov <appro@fy.chalmers.se>*
17118 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17119 These are required because of the typesafe stack would otherwise break
17120 existing code. If old code used a structure member which used to be STACK
17121 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17122 sk_num or sk_value it would produce an error because the num, data members
17123 are not present in STACK_OF. Now it just produces a warning. sk_set
17124 replaces the old method of assigning a value to sk_value
17125 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17126 that does this will no longer work (and should use sk_set instead) but
17127 this could be regarded as a "questionable" behaviour anyway.
17131 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17132 correctly handle encrypted S/MIME data.
17136 * Change type of various DES function arguments from des_cblock
17137 (which means, in function argument declarations, pointer to char)
17138 to des_cblock * (meaning pointer to array with 8 char elements),
17139 which allows the compiler to do more typechecking; it was like
17140 that back in SSLeay, but with lots of ugly casts.
17142 Introduce new type const_des_cblock.
17146 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17147 problems: find RecipientInfo structure that matches recipient certificate
17148 and initialise the ASN1 structures properly based on passed cipher.
17152 * Belatedly make the BN tests actually check the results.
17156 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17157 to and from BNs: it was completely broken. New compilation option
17158 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17159 key elements as negative integers.
17163 * Reorganize and speed up MD5.
17165 *Andy Polyakov <appro@fy.chalmers.se>*
17169 *Richard Levitte <richard@levitte.org>*
17171 * New option -out to asn1parse to allow the parsed structure to be
17172 output to a file. This is most useful when combined with the -strparse
17173 option to examine the output of things like OCTET STRINGS.
17177 * Make SSL library a little more fool-proof by not requiring any longer
17178 that `SSL_set_{accept,connect}_state` be called before
17179 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17180 in many applications because usually everything *appeared* to work as
17181 intended anyway -- now it really works as intended).
17185 * Move openssl.cnf out of lib/.
17189 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17190 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17191 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17193 *Ralf S. Engelschall*
17195 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17196 handle PKCS#7 enveloped data properly.
17198 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17200 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17201 copying pointers. The cert_st handling is changed by this in
17202 various ways (and thus what used to be known as ctx->default_cert
17203 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17204 any longer when s->cert does not give us what we need).
17205 ssl_cert_instantiate becomes obsolete by this change.
17206 As soon as we've got the new code right (possibly it already is?),
17207 we have solved a couple of bugs of the earlier code where s->cert
17208 was used as if it could not have been shared with other SSL structures.
17210 Note that using the SSL API in certain dirty ways now will result
17211 in different behaviour than observed with earlier library versions:
17212 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17213 does not influence s as it used to.
17215 In order to clean up things more thoroughly, inside SSL_SESSION
17216 we don't use CERT any longer, but a new structure SESS_CERT
17217 that holds per-session data (if available); currently, this is
17218 the peer's certificate chain and, for clients, the server's certificate
17219 and temporary key. CERT holds only those values that can have
17220 meaningful defaults in an SSL_CTX.
17224 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17225 from the internal representation. Various PKCS#7 fixes: remove some
17226 evil casts and set the enc_dig_alg field properly based on the signing
17231 * Allow PKCS#12 password to be set from the command line or the
17232 environment. Let 'ca' get its config file name from the environment
17233 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17238 * Allow certificate policies extension to use an IA5STRING for the
17239 organization field. This is contrary to the PKIX definition but
17240 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17245 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17246 without disallowing inline assembler and the like for non-pedantic builds.
17250 * Support Borland C++ builder.
17252 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17258 * SHA-1 cleanups and performance enhancements.
17260 *Andy Polyakov <appro@fy.chalmers.se>*
17262 * Sparc v8plus assembler for the bignum library.
17264 *Andy Polyakov <appro@fy.chalmers.se>*
17266 * Accept any -xxx and +xxx compiler options in Configure.
17270 * Update HPUX configuration.
17274 * Add missing `sk_<type>_unshift()` function to safestack.h
17276 *Ralf S. Engelschall*
17278 * New function SSL_CTX_use_certificate_chain_file that sets the
17279 "extra_cert"s in addition to the certificate. (This makes sense
17280 only for "PEM" format files, as chains as a whole are not
17285 * Support verify_depth from the SSL API.
17286 x509_vfy.c had what can be considered an off-by-one-error:
17287 Its depth (which was not part of the external interface)
17288 was actually counting the number of certificates in a chain;
17289 now it really counts the depth.
17293 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17294 instead of X509err, which often resulted in confusing error
17295 messages since the error codes are not globally unique
17296 (e.g. an alleged error in ssl3_accept when a certificate
17297 didn't match the private key).
17299 * New function SSL_CTX_set_session_id_context that allows to set a default
17300 value (so that you don't need SSL_set_session_id_context for each
17301 connection using the SSL_CTX).
17305 * OAEP decoding bug fix.
17309 * Support INSTALL_PREFIX for package builders, as proposed by
17314 * New Configure options "threads" and "no-threads". For systems
17315 where the proper compiler options are known (currently Solaris
17316 and Linux), "threads" is the default.
17320 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17324 * Install various scripts to $(OPENSSLDIR)/misc, not to
17325 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17326 such as /usr/local/bin.
17330 * "make linux-shared" to build shared libraries.
17332 *Niels Poppe <niels@netbox.org>*
17334 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17338 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17339 extension adding in x509 utility.
17343 * Remove NOPROTO sections and error code comments.
17347 * Partial rewrite of the DEF file generator to now parse the ANSI
17352 * New Configure options --prefix=DIR and --openssldir=DIR.
17356 * Complete rewrite of the error code script(s). It is all now handled
17357 by one script at the top level which handles error code gathering,
17358 header rewriting and C source file generation. It should be much better
17359 than the old method: it now uses a modified version of Ulf's parser to
17360 read the ANSI prototypes in all header files (thus the old K&R definitions
17361 aren't needed for error creation any more) and do a better job of
17362 translating function codes into names. The old 'ASN1 error code embedded
17363 in a comment' is no longer necessary and it doesn't use .err files which
17364 have now been deleted. Also the error code call doesn't have to appear all
17365 on one line (which resulted in some large lines...).
17369 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17373 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17374 0 (which usually indicates a closed connection), but continue reading.
17378 * Fix some race conditions.
17382 * Add support for CRL distribution points extension. Add Certificate
17383 Policies and CRL distribution points documentation.
17387 * Move the autogenerated header file parts to crypto/opensslconf.h.
17391 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17392 8 of keying material. Merlin has also confirmed interop with this fix
17393 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17395 *Merlin Hughes <merlin@baltimore.ie>*
17397 * Fix lots of warnings.
17399 *Richard Levitte <levitte@stacken.kth.se>*
17401 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17402 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17404 *Richard Levitte <levitte@stacken.kth.se>*
17406 * Fix problems with sizeof(long) == 8.
17408 *Andy Polyakov <appro@fy.chalmers.se>*
17410 * Change functions to ANSI C.
17414 * Fix typos in error codes.
17416 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17418 * Remove defunct assembler files from Configure.
17422 * SPARC v8 assembler BIGNUM implementation.
17424 *Andy Polyakov <appro@fy.chalmers.se>*
17426 * Support for Certificate Policies extension: both print and set.
17427 Various additions to support the r2i method this uses.
17431 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17432 return a const string when you are expecting an allocated buffer.
17436 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17437 types DirectoryString and DisplayText.
17441 * Add code to allow r2i extensions to access the configuration database,
17442 add an LHASH database driver and add several ctx helper functions.
17446 * Fix an evil bug in bn_expand2() which caused various BN functions to
17447 fail when they extended the size of a BIGNUM.
17451 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17452 support typesafe stack.
17456 * Fix typo in SSL_[gs]et_options().
17458 *Nils Frostberg <nils@medcom.se>*
17460 * Delete various functions and files that belonged to the (now obsolete)
17461 old X509V3 handling code.
17465 * New Configure option "rsaref".
17469 * Don't auto-generate pem.h.
17473 * Introduce type-safe ASN.1 SETs.
17477 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17479 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17481 * Introduce type-safe STACKs. This will almost certainly break lots of code
17482 that links with OpenSSL (well at least cause lots of warnings), but fear
17483 not: the conversion is trivial, and it eliminates loads of evil casts. A
17484 few STACKed things have been converted already. Feel free to convert more.
17485 In the fullness of time, I'll do away with the STACK type altogether.
17489 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17490 specified in `<certfile>` by updating the entry in the index.txt file.
17491 This way one no longer has to edit the index.txt file manually for
17492 revoking a certificate. The -revoke option does the gory details now.
17494 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17496 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17497 `-text` option at all and this way the `-noout -text` combination was
17498 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17500 *Ralf S. Engelschall*
17502 * Make sure a corresponding plain text error message exists for the
17503 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17504 verify callback function determined that a certificate was revoked.
17506 *Ralf S. Engelschall*
17508 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17509 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17510 all available ciphers including rc5, which was forgotten until now.
17511 In order to let the testing shell script know which algorithms
17512 are available, a new (up to now undocumented) command
17513 `openssl list-cipher-commands` is used.
17517 * Bugfix: s_client occasionally would sleep in select() when
17518 it should have checked SSL_pending() first.
17522 * New functions DSA_do_sign and DSA_do_verify to provide access to
17523 the raw DSA values prior to ASN.1 encoding.
17527 * Tweaks to Configure
17529 *Niels Poppe <niels@netbox.org>*
17531 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17536 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17540 * New config option to avoid instructions that are illegal on the 80386.
17541 The default code is faster, but requires at least a 486.
17545 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17546 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17547 same as SSL2_VERSION anyway.
17551 * New "-showcerts" option for s_client.
17555 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17556 application. Various cleanups and fixes.
17560 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17561 modify error routines to work internally. Add error codes and PBE init
17562 to library startup routines.
17566 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17567 packing functions to asn1 and evp. Changed function names and error
17568 codes along the way.
17572 * PKCS12 integration: and so it begins... First of several patches to
17573 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17574 objects to objects.h
17578 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17579 and display support for Thawte strong extranet extension.
17583 * Add LinuxPPC support.
17585 *Jeff Dubrule <igor@pobox.org>*
17587 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17588 bn_div_words in alpha.s.
17590 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17592 * Make sure the RSA OAEP test is skipped under -DRSAref because
17593 OAEP isn't supported when OpenSSL is built with RSAref.
17595 *Ulf Moeller <ulf@fitug.de>*
17597 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17598 so they no longer are missing under -DNOPROTO.
17600 *Soren S. Jorvang <soren@t.dk>*
17602 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17604 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17605 doesn't work when the session is reused. Coming soon!
17609 * Fix a security hole, that allows sessions to be reused in the wrong
17610 context thus bypassing client cert protection! All software that uses
17611 client certs and session caches in multiple contexts NEEDS PATCHING to
17612 allow session reuse! A fuller solution is in the works.
17614 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17616 * Some more source tree cleanups (removed obsolete files
17617 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17618 permission on "config" script to be executable) and a fix for the INSTALL
17621 *Ulf Moeller <ulf@fitug.de>*
17623 * Remove some legacy and erroneous uses of malloc, free instead of
17626 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17628 * Make rsa_oaep_test return non-zero on error.
17630 *Ulf Moeller <ulf@fitug.de>*
17632 * Add support for native Solaris shared libraries. Configure
17633 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17634 if someone would make that last step automatic.
17636 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17638 * ctx_size was not built with the right compiler during "make links". Fixed.
17642 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17643 except NULL ciphers". This means the default cipher list will no longer
17644 enable NULL ciphers. They need to be specifically enabled e.g. with
17645 the string "DEFAULT:eNULL".
17649 * Fix to RSA private encryption routines: if p < q then it would
17650 occasionally produce an invalid result. This will only happen with
17651 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17655 * Be less restrictive and allow also `perl util/perlpath.pl
17656 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17657 because this way one can also use an interpreter named `perl5` (which is
17658 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17659 installed as `perl`).
17661 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17663 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17665 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17667 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17668 advapi32.lib to Win32 build and change the pem test comparison
17669 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17670 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17671 and crypto/des/ede_cbcm_enc.c.
17675 * DES quad checksum was broken on big-endian architectures. Fixed.
17679 * Comment out two functions in bio.h that aren't implemented. Fix up the
17680 Win32 test batch file so it (might) work again. The Win32 test batch file
17681 is horrible: I feel ill....
17685 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17686 in e_os.h. Audit of header files to check ANSI and non ANSI
17687 sections: 10 functions were absent from non ANSI section and not exported
17688 from Windows DLLs. Fixed up libeay.num for new functions.
17692 * Make `openssl version` output lines consistent.
17694 *Ralf S. Engelschall*
17696 * Fix Win32 symbol export lists for BIO functions: Added
17697 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17698 to ms/libeay{16,32}.def.
17700 *Ralf S. Engelschall*
17702 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17703 fine under Unix and passes some trivial tests I've now added. But the
17704 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17705 added to make sure no one expects that this stuff really works in the
17706 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17707 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17710 *Ralf S. Engelschall*
17712 * Fix the generation of two part addresses in perl.
17714 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17716 * Add config entry for Linux on MIPS.
17718 *John Tobey <jtobey@channel1.com>*
17720 * Make links whenever Configure is run, unless we are on Windoze.
17724 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17725 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17730 * Add a useful kludge to allow package maintainers to specify compiler and
17731 other platforms details on the command line without having to patch the
17732 Configure script every time: One now can use
17733 `perl Configure <id>:<details>`,
17734 i.e. platform ids are allowed to have details appended
17735 to them (separated by colons). This is treated as there would be a static
17736 pre-configured entry in Configure's %table under key `<id>` with value
17737 `<details>` and `perl Configure <id>` is called. So, when you want to
17738 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17739 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17740 now, which overrides the FreeBSD-elf entry on-the-fly.
17742 *Ralf S. Engelschall*
17744 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17748 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17749 on the `perl Configure ...` command line. This way one can compile
17750 OpenSSL libraries with Position Independent Code (PIC) which is needed
17751 for linking it into DSOs.
17753 *Ralf S. Engelschall*
17755 * Remarkably, export ciphers were totally broken and no-one had noticed!
17760 * Cleaned up the LICENSE document: The official contact for any license
17761 questions now is the OpenSSL core team under openssl-core@openssl.org.
17762 And add a paragraph about the dual-license situation to make sure people
17763 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17764 to the OpenSSL toolkit.
17766 *Ralf S. Engelschall*
17768 * General source tree makefile cleanups: Made `making xxx in yyy...`
17769 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17770 Additionally cleaned up the `make links` target: Remove unnecessary
17771 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17772 to speed processing and no longer clutter the display with confusing
17773 stuff. Instead only the actually done links are displayed.
17775 *Ralf S. Engelschall*
17777 * Permit null encryption ciphersuites, used for authentication only. It used
17778 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17779 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17784 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17785 signed attributes when verifying signatures (this would break them),
17786 the detached data encoding was wrong and public keys obtained using
17787 X509_get_pubkey() weren't freed.
17791 * Add text documentation for the BUFFER functions. Also added a work around
17792 to a Win95 console bug. This was triggered by the password read stuff: the
17793 last character typed gets carried over to the next fread(). If you were
17794 generating a new cert request using 'req' for example then the last
17795 character of the passphrase would be CR which would then enter the first
17800 * Added the new 'Includes OpenSSL Cryptography Software' button as
17801 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17802 button and can be used by applications based on OpenSSL to show the
17803 relationship to the OpenSSL project.
17805 *Ralf S. Engelschall*
17807 * Remove confusing variables in function signatures in files
17808 ssl/ssl_lib.c and ssl/ssl.h.
17810 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17812 * Don't install bss_file.c under PREFIX/include/
17814 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17816 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17817 functions that return function pointers and has support for NT specific
17818 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17819 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17820 unsigned to signed types: this was killing the Win32 compile.
17824 * Add new certificate file to stack functions,
17825 SSL_add_dir_cert_subjects_to_stack() and
17826 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17827 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17828 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17829 This means that Apache-SSL and similar packages don't have to mess around
17830 to add as many CAs as they want to the preferred list.
17834 * Experiment with doxygen documentation. Currently only partially applied to
17836 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17837 openssl.doxy as the configuration file.
17841 * Get rid of remaining C++-style comments which strict C compilers hate.
17843 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17845 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17846 compiled in by default: it has problems with large keys.
17850 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17851 DH private keys and/or callback functions which directly correspond to
17852 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17853 is needed for applications which have to configure certificates on a
17854 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17856 For the RSA certificate situation is makes no difference, but
17857 for the DSA certificate situation this fixes the "no shared cipher"
17858 problem where the OpenSSL cipher selection procedure failed because the
17859 temporary keys were not overtaken from the context and the API provided
17860 no way to reconfigure them.
17861 The new functions now let applications reconfigure the stuff and they
17862 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17863 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17864 non-public-API function ssl_cert_instantiate() is used as a helper
17865 function and also to reduce code redundancy inside ssl_rsa.c.
17867 *Ralf S. Engelschall*
17869 * Move s_server -dcert and -dkey options out of the undocumented feature
17870 area because they are useful for the DSA situation and should be
17871 recognized by the users.
17873 *Ralf S. Engelschall*
17875 * Fix the cipher decision scheme for export ciphers: the export bits are
17876 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17877 SSL_EXP_MASK. So, the original variable has to be used instead of the
17878 already masked variable.
17880 *Richard Levitte <levitte@stacken.kth.se>*
17882 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17884 *Richard Levitte <levitte@stacken.kth.se>*
17886 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17887 from `int` to `unsigned int` because it is a length and initialized by
17888 EVP_DigestFinal() which expects an `unsigned int *`.
17890 *Richard Levitte <levitte@stacken.kth.se>*
17892 * Don't hard-code path to Perl interpreter on shebang line of Configure
17893 script. Instead use the usual Shell->Perl transition trick.
17895 *Ralf S. Engelschall*
17897 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17898 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17899 -noout -modulus` as it's already the case for `openssl rsa -noout
17900 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17901 currently the public key is printed (a decision which was already done by
17902 `openssl dsa -modulus` in the past) which serves a similar purpose.
17903 Additionally the NO_RSA no longer completely removes the whole -modulus
17904 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17907 *Ralf S. Engelschall*
17909 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17910 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17912 *Arne Ansper <arne@ats.cyber.ee>*
17914 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17915 to be added. Now both 'req' and 'ca' can use new objects defined in the
17920 * Add cool BIO that does syslog (or event log on NT).
17922 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17924 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17925 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17926 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17927 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17931 * Add preliminary config info for new extension code.
17935 * Make RSA_NO_PADDING really use no padding.
17937 *Ulf Moeller <ulf@fitug.de>*
17939 * Generate errors when private/public key check is done.
17943 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17944 for some CRL extensions and new objects added.
17948 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17949 key usage extension and fuller support for authority key id.
17953 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17954 padding method for RSA, which is recommended for new applications in PKCS
17955 #1 v2.0 (RFC 2437, October 1998).
17956 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17957 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17958 against Bleichbacher's attack on RSA.
17959 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17962 * Updates to the new SSL compression code
17964 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17966 * Fix so that the version number in the master secret, when passed
17967 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17968 (because the server will not accept higher), that the version number
17969 is 0x03,0x01, not 0x03,0x00
17971 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17973 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17974 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17975 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17979 * Support for RAW extensions where an arbitrary extension can be
17980 created by including its DER encoding. See apps/openssl.cnf for
17985 * Make sure latest Perl versions don't interpret some generated C array
17986 code as Perl array code in the crypto/err/err_genc.pl script.
17988 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17990 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17991 not many people have the assembler. Various Win32 compilation fixes and
17992 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17993 build instructions.
17997 * Modify configure script 'Configure' to automatically create crypto/date.h
17998 file under Win32 and also build pem.h from pem.org. New script
17999 util/mkfiles.pl to create the MINFO file on environments that can't do a
18000 'make files': perl util/mkfiles.pl >MINFO should work.
18004 * Major rework of DES function declarations, in the pursuit of correctness
18005 and purity. As a result, many evil casts evaporated, and some weirdness,
18006 too. You may find this causes warnings in your code. Zapping your evil
18007 casts will probably fix them. Mostly.
18011 * Fix for a typo in asn1.h. Bug fix to object creation script
18012 obj_dat.pl. It considered a zero in an object definition to mean
18013 "end of object": none of the objects in objects.h have any zeros
18014 so it wasn't spotted.
18016 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18018 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18019 Masking (CBCM). In the absence of test vectors, the best I have been able
18020 to do is check that the decrypt undoes the encrypt, so far. Send me test
18021 vectors if you have them.
18025 * Correct calculation of key length for export ciphers (too much space was
18026 allocated for null ciphers). This has not been tested!
18030 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18031 message is now correct (it understands "crypto" and "ssl" on its
18032 command line). There is also now an "update" option. This will update
18033 the util/ssleay.num and util/libeay.num files with any new functions.
18035 perl util/mkdef.pl crypto ssl update
18036 it will update them.
18040 * Overhauled the Perl interface:
18041 - ported BN stuff to OpenSSL's different BN library
18042 - made the perl/ source tree CVS-aware
18043 - renamed the package from SSLeay to OpenSSL (the files still contain
18044 their history because I've copied them in the repository)
18045 - removed obsolete files (the test scripts will be replaced
18046 by better Test::Harness variants in the future)
18048 *Ralf S. Engelschall*
18050 * First cut for a very conservative source tree cleanup:
18051 1. merge various obsolete readme texts into doc/ssleay.txt
18052 where we collect the old documents and readme texts.
18053 2. remove the first part of files where I'm already sure that we no
18054 longer need them because of three reasons: either they are just temporary
18055 files which were left by Eric or they are preserved original files where
18056 I've verified that the diff is also available in the CVS via "cvs diff
18057 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18058 the crypto/md/ stuff).
18060 *Ralf S. Engelschall*
18062 * More extension code. Incomplete support for subject and issuer alt
18063 name, issuer and authority key id. Change the i2v function parameters
18064 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18065 what that's for :-) Fix to ASN1 macro which messed up
18066 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18070 * Preliminary support for ENUMERATED type. This is largely copied from the
18075 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18077 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18079 * Make sure `make rehash` target really finds the `openssl` program.
18081 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18083 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18084 like to hear about it if this slows down other processors.
18088 * Add CygWin32 platform information to Configure script.
18090 *Alan Batie <batie@aahz.jf.intel.com>*
18092 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18094 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18096 * New program nseq to manipulate netscape certificate sequences
18100 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18105 * Fixes to BN code. Previously the default was to define BN_RECURSION
18106 but the BN code had some problems that would cause failures when
18107 doing certificate verification and some other functions.
18109 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18111 * Add ASN1 and PEM code to support netscape certificate sequences.
18115 * Add ASN1 and PEM code to support netscape certificate sequences.
18119 * Add several PKIX and private extended key usage OIDs.
18123 * Modify the 'ca' program to handle the new extension code. Modify
18124 openssl.cnf for new extension format, add comments.
18128 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18129 and add a sample to openssl.cnf so req -x509 now adds appropriate
18134 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18135 error code, add initial support to X509_print() and x509 application.
18139 * Takes a deep breath and start adding X509 V3 extension support code. Add
18140 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18141 stuff is currently isolated and isn't even compiled yet.
18145 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18146 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18147 Removed the versions check from X509 routines when loading extensions:
18148 this allows certain broken certificates that don't set the version
18149 properly to be processed.
18153 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18154 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18155 can still be regenerated with "make depend".
18159 * Spelling mistake in C version of CAST-128.
18161 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18163 * Changes to the error generation code. The perl script err-code.pl
18164 now reads in the old error codes and retains the old numbers, only
18165 adding new ones if necessary. It also only changes the .err files if new
18166 codes are added. The makefiles have been modified to only insert errors
18167 when needed (to avoid needlessly modifying header files). This is done
18168 by only inserting errors if the .err file is newer than the auto generated
18169 C file. To rebuild all the error codes from scratch (the old behaviour)
18170 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18171 or delete all the .err files.
18175 * CAST-128 was incorrectly implemented for short keys. The C version has
18176 been fixed, but is untested. The assembler versions are also fixed, but
18177 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18178 to regenerate it if needed.
18179 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18180 Hagino <itojun@kame.net>*
18182 * File was opened incorrectly in randfile.c.
18184 *Ulf Möller <ulf@fitug.de>*
18186 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18187 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18188 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18189 al: it's just almost always a UTCTime. Note this patch adds new error
18190 codes so do a "make errors" if there are problems.
18194 * Correct Linux 1 recognition in config.
18196 *Ulf Möller <ulf@fitug.de>*
18198 * Remove pointless MD5 hash when using DSA keys in ca.
18200 *Anonymous <nobody@replay.com>*
18202 * Generate an error if given an empty string as a cert directory. Also
18203 generate an error if handed NULL (previously returned 0 to indicate an
18204 error, but didn't set one).
18206 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18208 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18212 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18213 parameters. This was causing a warning which killed off the Win32 compile.
18217 * Remove C++ style comments from crypto/bn/bn_local.h.
18219 *Neil Costigan <neil.costigan@celocom.com>*
18221 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18222 based on a text string, looking up short and long names and finally
18223 "dot" format. The "dot" format stuff didn't work. Added new function
18224 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18225 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18226 OID is not part of the table.
18230 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18231 X509_LOOKUP_by_alias().
18235 * Sort openssl functions by name.
18239 * Get the gendsa program working (hopefully) and add it to app list. Remove
18240 encryption from sample DSA keys (in case anyone is interested the password
18245 * Make *all* `*_free` functions accept a NULL pointer.
18247 *Frans Heymans <fheymans@isaserver.be>*
18249 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18252 *Anonymous <nobody@replay.com>*
18254 * s_server should send the CAfile as acceptable CAs, not its own cert.
18256 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18258 * Don't blow it for numeric -newkey arguments to apps/req.
18260 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18262 * Temp key "for export" tests were wrong in s3_srvr.c.
18264 *Anonymous <nobody@replay.com>*
18266 * Add prototype for temp key callback functions
18267 SSL_CTX_set_tmp_{rsa,dh}_callback().
18271 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18272 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18276 * X509_name_add_entry() freed the wrong thing after an error.
18278 *Arne Ansper <arne@ats.cyber.ee>*
18280 * rsa_eay.c would attempt to free a NULL context.
18282 *Arne Ansper <arne@ats.cyber.ee>*
18284 * BIO_s_socket() had a broken should_retry() on Windoze.
18286 *Arne Ansper <arne@ats.cyber.ee>*
18288 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18290 *Arne Ansper <arne@ats.cyber.ee>*
18292 * Make sure the already existing X509_STORE->depth variable is initialized
18293 in X509_STORE_new(), but document the fact that this variable is still
18294 unused in the certificate verification process.
18296 *Ralf S. Engelschall*
18298 * Fix the various library and apps files to free up pkeys obtained from
18299 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18303 * Fix reference counting in X509_PUBKEY_get(). This makes
18304 demos/maurice/example2.c work, amongst others, probably.
18306 *Steve Henson and Ben Laurie*
18308 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18309 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18310 are no longer created. This way we have a single and consistent command
18311 line interface `openssl <command>`, similar to `cvs <command>`.
18313 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18315 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18316 BIT STRING wrapper always have zero unused bits.
18320 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18324 * Make the top-level INSTALL documentation easier to understand.
18328 * Makefiles updated to exit if an error occurs in a sub-directory
18329 make (including if user presses ^C) [Paul Sutton]
18331 * Make Montgomery context stuff explicit in RSA data structure.
18335 * Fix build order of pem and err to allow for generated pem.h.
18339 * Fix renumbering bug in X509_NAME_delete_entry().
18343 * Enhanced the err-ins.pl script so it makes the error library number
18344 global and can add a library name. This is needed for external ASN1 and
18345 other error libraries.
18349 * Fixed sk_insert which never worked properly.
18353 * Fix ASN1 macros so they can handle indefinite length constructed
18354 EXPLICIT tags. Some non standard certificates use these: they can now
18359 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18360 into a single doc/ssleay.txt bundle. This way the information is still
18361 preserved but no longer messes up this directory. Now it's new room for
18362 the new set of documentation files.
18364 *Ralf S. Engelschall*
18366 * SETs were incorrectly DER encoded. This was a major pain, because they
18367 shared code with SEQUENCEs, which aren't coded the same. This means that
18368 almost everything to do with SETs or SEQUENCEs has either changed name or
18369 number of arguments.
18371 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18373 * Fix test data to work with the above.
18377 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18378 was already fixed by Eric for 0.9.1 it seems.
18380 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18382 * Autodetect FreeBSD3.
18386 * Fix various bugs in Configure. This affects the following platforms:
18390 unixware-2.0-pentium
18395 * Eliminate generated files from CVS. Reorder tests to regenerate files
18396 before they are needed.
18400 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18404 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18406 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18407 changed SSLeay to OpenSSL in version strings.
18409 *Ralf S. Engelschall*
18411 * Some fixups to the top-level documents.
18415 * Fixed the nasty bug where rsaref.h was not found under compile-time
18416 because the symlink to include/ was missing.
18418 *Ralf S. Engelschall*
18420 * Incorporated the popular no-RSA/DSA-only patches
18421 which allow to compile a RSA-free SSLeay.
18423 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18425 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18426 when "ssleay" is still not found.
18428 *Ralf S. Engelschall*
18430 * Added more platforms to Configure: Cray T3E, HPUX 11,
18432 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18434 * Updated the README file.
18436 *Ralf S. Engelschall*
18438 * Added various .cvsignore files in the CVS repository subdirs
18439 to make a "cvs update" really silent.
18441 *Ralf S. Engelschall*
18443 * Recompiled the error-definition header files and added
18444 missing symbols to the Win32 linker tables.
18446 *Ralf S. Engelschall*
18448 * Cleaned up the top-level documents;
18449 o new files: CHANGES and LICENSE
18450 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18451 o merged COPYRIGHT into LICENSE
18452 o removed obsolete TODO file
18453 o renamed MICROSOFT to INSTALL.W32
18455 *Ralf S. Engelschall*
18457 * Removed dummy files from the 0.9.1b source tree:
18458 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18459 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18460 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18461 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18462 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18464 *Ralf S. Engelschall*
18466 * Added various platform portability fixes.
18470 * The Genesis of the OpenSSL rpject:
18471 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18472 Young and Tim J. Hudson created while they were working for C2Net until
18475 *The OpenSSL Project*
18477 ### Changes between 0.9.0b and 0.9.1b [not released]
18479 * Updated a few CA certificates under certs/
18483 * Changed some BIGNUM api stuff.
18487 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18488 DGUX x86, Linux Alpha, etc.
18492 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18493 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18498 * Add -strparse option to asn1pars program which parses nested
18501 *Dr Stephen Henson <shenson@bigfoot.com>*
18503 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18507 * DSA fix for "ca" program.
18511 * Added "-genkey" option to "dsaparam" program.
18515 * Added RIPE MD160 (rmd160) message digest.
18519 * Added -a (all) option to "ssleay version" command.
18523 * Added PLATFORM define which is the id given to Configure.
18527 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18531 * Extended the ASN.1 parser routines.
18535 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18539 * Added a BN_CTX to the BN library.
18543 * Fixed the weak key values in DES library
18547 * Changed API in EVP library for cipher aliases.
18551 * Added support for RC2/64bit cipher.
18555 * Converted the lhash library to the crypto/mem.c functions.
18559 * Added more recognized ASN.1 object ids.
18563 * Added more RSA padding checks for SSL/TLS.
18567 * Added BIO proxy/filter functionality.
18571 * Added extra_certs to SSL_CTX which can be used
18572 send extra CA certificates to the client in the CA cert chain sending
18573 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18577 * Now Fortezza is denied in the authentication phase because
18578 this is key exchange mechanism is not supported by SSLeay at all.
18582 * Additional PKCS1 checks.
18586 * Support the string "TLSv1" for all TLS v1 ciphers.
18590 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18591 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18595 * Fixed a few memory leaks.
18599 * Fixed various code and comment typos.
18603 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18604 bytes sent in the client random.
18606 *Edward Bishop <ebishop@spyglass.com>*
18610 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18611 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18612 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18613 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18614 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18615 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18616 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18617 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18618 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18619 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18620 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18621 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18622 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18623 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18624 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18625 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18626 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18627 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18628 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18629 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18630 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18631 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18632 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18633 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18634 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18635 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18636 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18637 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18638 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18639 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18640 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18641 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18642 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18643 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18644 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18645 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18646 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18647 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18648 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18649 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18650 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18651 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18652 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18653 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18654 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18655 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18656 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18657 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18658 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18659 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18660 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18661 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18662 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18663 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18664 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18665 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18666 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18667 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18668 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18669 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18670 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18671 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18672 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18673 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18674 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18675 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18676 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18677 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18678 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18679 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18680 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18681 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18682 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18683 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18684 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18685 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18686 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18687 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18688 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18689 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18690 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18691 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18692 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18693 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18694 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18695 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18696 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18697 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18698 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18699 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18700 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18701 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18702 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18703 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18704 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18705 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18706 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18707 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18708 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18709 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18710 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18711 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18712 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18713 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18714 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18715 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18716 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18717 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18718 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18719 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18720 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18721 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18722 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18723 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18724 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18725 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18726 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18727 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18728 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18729 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18730 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18731 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18732 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18733 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18734 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18735 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18736 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18737 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18738 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18739 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18740 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18741 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18742 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18743 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18744 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18745 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18746 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18747 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18748 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18749 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18750 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18751 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18752 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18753 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18754 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18755 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18756 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18757 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18758 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18759 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18760 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18761 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18762 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18763 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18764 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18765 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18766 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18767 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18768 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18769 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18770 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655